├── .Rbuildignore
├── .github
├── .gitignore
└── workflows
│ └── check-bioc.yml
├── .gitignore
├── DESCRIPTION
├── LICENSE
├── NAMESPACE
├── NEWS.md
├── R
├── HMDB_mapper_vec.R
├── binarize_with_sign.R
├── check_COSMOS_inputs.R
├── check_gene_names.R
├── check_inputs_CARNIVAL.R
├── check_network_data_coverage.R
├── compress_same_children.R
├── cosmos_data.R
├── decoupleRnival.R
├── default_CARNIVAL_options.R
├── discretize_vector.R
├── display_node_neighboorhood.R
├── filter_input_nodes_not_in_pkn.R
├── filter_pkn_expressed_genes.R
├── filter_transcriptional_regulations.R
├── format_COSMOS_results.R
├── get_TF_activity_from_CARNIVAL.R
├── keep_downstream_neighbours.R
├── limit_string_vec.R
├── load_tf_regulon_dorothea.R
├── meta_network.R
├── prepare_metab_inputs.R
├── preprocess_COSMOS.R
├── preprocess_COSMOS_metabolism_to_signaling.R
├── preprocess_COSMOS_signaling_to_metabolism.R
├── process_CARNIVAL_results.R
├── runCARNIVAL_wrapper.R
├── run_COSMOS_metabolism_to_signaling.R
├── run_COSMOS_signaling_to_metabolism.R
├── support_ORA.R
├── support_decoupleR.R
├── support_pheatmap.R
├── sysdata.rda
├── toy_RNA.R
├── toy_metabolic_input.R
├── toy_network.R
└── toy_signaling_input.R
├── README.md
├── data
├── HMDB_mapper_vec.RData
├── meta_network.RData
├── toy_RNA.RData
├── toy_metabolic_input.RData
├── toy_network.RData
└── toy_signaling_input.RData
├── inst
└── CITATION
├── man
├── HMDB_mapper_vec.Rd
├── compress_same_children.Rd
├── cosmos_data.Rd
├── createLinearColors.Rd
├── decompress_moon_result.Rd
├── decompress_solution_network.Rd
├── decoupleRnival.Rd
├── default_CARNIVAL_options.Rd
├── display_node_neighboorhood.Rd
├── extract_nodes_for_ORA.Rd
├── figures
│ ├── graphical_abstract.png
│ ├── intro_data.png
│ ├── logo.png
│ └── summary.png
├── filter_incohrent_TF_target.Rd
├── format_COSMOS_res.Rd
├── format_LR_ressource.Rd
├── get_moon_scoring_network.Rd
├── gmt_to_dataframe.Rd
├── load_tf_regulon_dorothea.Rd
├── make_heatmap_color_palette.Rd
├── meta_network.Rd
├── meta_network_cleanup.Rd
├── moon.Rd
├── prepare_metab_inputs.Rd
├── preprocess_COSMOS_metabolism_to_signaling.Rd
├── preprocess_COSMOS_signaling_to_metabolism.Rd
├── print.cosmos_data.Rd
├── reduce_solution_network.Rd
├── run_COSMOS_metabolism_to_signaling.Rd
├── run_COSMOS_signaling_to_metabolism.Rd
├── toy_RNA.Rd
├── toy_metabolic_input.Rd
├── toy_network.Rd
├── toy_signaling_input.Rd
├── translate_column_HMDB.Rd
├── translate_res.Rd
└── wide_ulm_res.Rd
├── pkgdown
├── _pkgdown.yml
├── extra.css
└── favicon
│ ├── apple-touch-icon-120x120.png
│ ├── apple-touch-icon-152x152.png
│ ├── apple-touch-icon-180x180.png
│ ├── apple-touch-icon-60x60.png
│ ├── apple-touch-icon-76x76.png
│ ├── apple-touch-icon.png
│ ├── favicon-16x16.png
│ ├── favicon-32x32.png
│ └── favicon.ico
├── tests
├── testthat.R
└── testthat
│ ├── test-check_COSMOS_inputs.R
│ ├── test-cosmos_data.R
│ ├── test-load_tf_regulon_dorothea.R
│ ├── test-preprocess_COSMOS.R
│ ├── test-run_COSMOS_metabolism_to_signaling.R
│ └── test-run_COSMOS_signaling_to_metabolism.R
└── vignettes
├── M2Cf
└── figs
│ ├── Activity_estimations_for_factor_4_1.png
│ ├── Activity_estimations_for_factor_4_2.png
│ ├── Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_1.png
│ ├── Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_2.png
│ ├── Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_3.png
│ ├── Compare_MOFA_models_1.png
│ ├── Factor_weights_per_view_1.png
│ ├── Plot_Ligand_receptor_activity_estimates_1.png
│ ├── Plot_Metabolite_factor_weights_1.png
│ ├── Plot_RNA_weights_and_protein_weights_1.png
│ ├── Plot_RNA_weights_and_protein_weights_2.png
│ ├── Plot_TF_activity_estimates_1.png
│ ├── Preparing_MOFA_input_1.png
│ ├── Preparing_MOFA_input_2.png
│ ├── Preparing_MOFA_input_3.png
│ ├── Preparing_MOFA_input_4.png
│ ├── Preparing_MOFA_input_5.png
│ ├── Run_moon_rec_to_TFmetab_1.png
│ ├── Supp_figure_condition_prot_RNA_correlation_1.png
│ ├── Supp_figure_single_prot_RNA_correlation_1.png
│ ├── Total_variance_per_factor_1.png
│ ├── Variance_per_view_per_factor_1.png
│ ├── correlation_RNA
│ └── prot-1.png
│ └── run_moon_TF_to_lig_1.png
├── MOFA_to_COSMOS.Rmd
├── MOFA_to_COSMOS.md
├── NCI60_tutorial.Rmd
├── net_compr_MOON.md
├── net_compr_MOON_files
└── figure-markdown_strict
│ └── extract_subnetwork_from_scored_MOON_network_1.png
└── tutorial.Rmd
/.Rbuildignore:
--------------------------------------------------------------------------------
1 | ^.*\.Rproj$
2 | ^\.Rproj\.user$
3 | ^.git$
4 | ^pkgdown$
5 | cosmos_cache
6 | docs
7 | ^\.github$
8 | ^devel$
9 | ^doc$
10 | ^Meta$
11 |
--------------------------------------------------------------------------------
/.github/.gitignore:
--------------------------------------------------------------------------------
1 | *.html
2 |
--------------------------------------------------------------------------------
/.gitignore:
--------------------------------------------------------------------------------
1 | .Rproj.user
2 | *.Rproj
3 | .Rhistory
4 | .Rprofile
5 | vignettes/.Rhistory
6 | vignettes/.Rhistory
7 | vignettes/Tutorial.pdf
8 | vignettes/Tutorial.html
9 | .DS_Store
10 |
11 | docs/
12 | /doc/
13 | /Meta/
14 |
--------------------------------------------------------------------------------
/DESCRIPTION:
--------------------------------------------------------------------------------
1 | Package: cosmosR
2 | Type: Package
3 | Title: COSMOS (Causal Oriented Search of Multi-Omic Space)
4 | Version: 1.9.1
5 | Description: COSMOS (Causal Oriented Search of Multi-Omic Space) is a method
6 | that integrates phosphoproteomics, transcriptomics, and metabolomics data sets
7 | based on prior knowledge of signaling, metabolic, and gene regulatory
8 | networks. It estimated the activities of transcrption factors and kinases
9 | and finds a network-level causal reasoning. Thereby, COSMOS provides
10 | mechanistic hypotheses for experimental observations across mulit-omics datasets.
11 | Authors@R: c(
12 | person(given = "Aurélien",
13 | family = "Dugourd",
14 | role = c("aut"),
15 | email = "dugourdaurelien@gmail.com",
16 | comment = c(ORCID = "0000-0002-0714-028X")),
17 | person(given = "Attila",
18 | family = "Gabor",
19 | role = c("cre"),
20 | email = "attila.gabor@uni-heidelberg.de",
21 | comment = c(ORCID = "0000-0002-0776-1182")),
22 | person(given = "Katharina",
23 | family = "Zirngibl",
24 | role = c("aut"),
25 | email = "katharina.zirngibl@uni-heidelberg.de",
26 | comment = c(ORCID = "0000-0002-7518-0339")))
27 | URL: https://github.com/saezlab/COSMOSR
28 | BugReports: https://github.com/saezlab/COSMOSR/issues
29 | Depends: R (>= 4.1)
30 | License: GPL-3
31 | Encoding: UTF-8
32 | LazyData: false
33 | RoxygenNote: 7.2.3
34 | VignetteBuilder: knitr
35 | Imports:
36 | CARNIVAL,
37 | dorothea,
38 | dplyr,
39 | GSEABase,
40 | igraph,
41 | progress,
42 | purrr,
43 | rlang,
44 | stringr,
45 | utils,
46 | visNetwork,
47 | decoupleR
48 | Suggests:
49 | testthat,
50 | knitr,
51 | rmarkdown,
52 | htmltools,
53 | markdown,
54 | piano,
55 | ggplot2,
56 | stringi,
57 | reshape2
58 | biocViews: CellBiology, Pathways, Network, Proteomics, Metabolomics, Transcriptomics,
59 | GeneSignaling
60 |
--------------------------------------------------------------------------------
/NAMESPACE:
--------------------------------------------------------------------------------
1 | # Generated by roxygen2: do not edit by hand
2 |
3 | S3method(print,cosmos_data)
4 | export(compress_same_children)
5 | export(createLinearColors)
6 | export(decompress_moon_result)
7 | export(decompress_solution_network)
8 | export(decoupleRnival)
9 | export(default_CARNIVAL_options)
10 | export(display_node_neighboorhood)
11 | export(extract_nodes_for_ORA)
12 | export(filter_incohrent_TF_target)
13 | export(format_COSMOS_res)
14 | export(format_LR_ressource)
15 | export(get_moon_scoring_network)
16 | export(load_tf_regulon_dorothea)
17 | export(make_heatmap_color_palette)
18 | export(meta_network_cleanup)
19 | export(moon)
20 | export(prepare_metab_inputs)
21 | export(preprocess_COSMOS_metabolism_to_signaling)
22 | export(preprocess_COSMOS_signaling_to_metabolism)
23 | export(reduce_solution_network)
24 | export(run_COSMOS_metabolism_to_signaling)
25 | export(run_COSMOS_signaling_to_metabolism)
26 | export(translate_column_HMDB)
27 | export(translate_res)
28 | export(wide_ulm_res)
29 | import(decoupleR)
30 | import(dplyr)
31 | importFrom(dplyr,"%>%")
32 | importFrom(dplyr,as_tibble)
33 | importFrom(dplyr,bind_rows)
34 | importFrom(dplyr,filter)
35 | importFrom(dplyr,mutate)
36 | importFrom(dplyr,rename)
37 | importFrom(dplyr,select)
38 | importFrom(progress,progress_bar)
39 | importFrom(purrr,map)
40 | importFrom(rlang,.data)
41 | importFrom(stringr,str_extract)
42 | importFrom(utils,data)
43 |
--------------------------------------------------------------------------------
/NEWS.md:
--------------------------------------------------------------------------------
1 | ## Changes in version 0.99.2 (2020-05-12)
2 |
3 | ---
4 |
5 | - Submitted to bioRxiv
6 | - Release of github page
7 | - Submitted to Bioconductor
--------------------------------------------------------------------------------
/R/HMDB_mapper_vec.R:
--------------------------------------------------------------------------------
1 | #' Toy Input Transcription Data Set
2 | #'
3 | #' This exemplary transcription data are the specific deregulated gene expression of the 786-O cell line from the NCI60 dataset.
4 | #'
5 | #' @docType data
6 | #'
7 | #' @usage data(HMDB_mapper_vec)
8 | #'
9 | #' @format An object of class \dQuote{\code{character}} containing the marching between HMDB metabolite IDs and there coresponding metabolite names.
10 | #'
11 | #' @source
12 | #' \url{https://bioconductor.org/packages/release/data/annotation/html/metaboliteIDmapping.html}
13 | #'
14 | #'
15 | #' @examples
16 | #' data(HMDB_mapper_vec)
17 | #'
18 | #' @keywords datasets
19 | "HMDB_mapper_vec"
--------------------------------------------------------------------------------
/R/binarize_with_sign.R:
--------------------------------------------------------------------------------
1 | #' binarize_with_sign
2 | #'
3 | #' binarizes the data based on the threshold using the absolute value of the data.
4 | #' Gives a sign based on the sign of the input data.
5 | #' @param data numerical vector
6 | #' @param threshold threshold used to binarise the data
7 | #' @noRd
8 | binarize_with_sign <- function(data,threshold){
9 |
10 | out = sign(data) * as.numeric(abs(data) >= threshold)
11 | return(out)
12 | }
13 |
--------------------------------------------------------------------------------
/R/check_COSMOS_inputs.R:
--------------------------------------------------------------------------------
1 | #' check_COSMOS_inputs
2 | #'
3 | #' checks the format of the main data inputs. Checks only the data that was
4 | #' passed (not NULL).
5 | #' @param meta_network prior knowledge network
6 | #' @param tf_regulon transcription factor regulon
7 | #' @param signaling_data named numerical vector
8 | #' @param metabolic_data named numerical vector
9 | #' @param expression_data named numerical vector
10 | #' @noRd
11 | check_COSMOS_inputs <- function(meta_network = NULL,
12 | tf_regulon = NULL,
13 | signaling_data = NULL,
14 | metabolic_data = NULL,
15 | expression_data = NULL){
16 | # checking the data
17 | if(!is.null(signaling_data)) stopifnot(is.vector(signaling_data))
18 | if(!is.null(metabolic_data)) stopifnot(is.vector(metabolic_data))
19 | if(!is.null(expression_data)) stopifnot(is.vector(expression_data))
20 |
21 | # check duplicated names
22 | if(!is.null(signaling_data) & anyDuplicated(names(signaling_data))) {
23 | stop("duplicated gene names in signaling data detected. ")
24 | }
25 | if(!is.null(expression_data) & anyDuplicated(names(expression_data))) {
26 | stop("duplicated gene names in expression_data detected. ")
27 | }
28 | if(!is.null(metabolic_data) & anyDuplicated(names(metabolic_data))) {
29 | stop("duplicated gene names in metabolic_data detected. ")
30 | }
31 |
32 |
33 | # checking the networks
34 | if(!is.null(meta_network)){
35 | stopifnot(is.data.frame(meta_network))
36 | stopifnot(all(c("source","interaction","target" ) %in% names(meta_network)))
37 | stopifnot(ncol(meta_network)==3)
38 | }
39 |
40 | if(!is.null(tf_regulon)){
41 | stopifnot(is.data.frame(tf_regulon))
42 | stopifnot(all(c("tf","sign","target" ) %in% names(tf_regulon)))
43 | stopifnot(ncol(meta_network)==3)
44 | }
45 | return(TRUE)
46 | }
47 |
48 |
49 |
--------------------------------------------------------------------------------
/R/check_gene_names.R:
--------------------------------------------------------------------------------
1 | #' check_gene_names
2 | #'
3 | #' checks if gene names looks like an EntrezID with a prefix of X.
4 | #' @param ... takes named vectors
5 | #' @noRd
6 | check_gene_names <- function(...){
7 |
8 | res = lapply(list(...),function(x){
9 | if(any(duplicated(names(x)))){
10 | stop("Duplicated gene names detected. Gene names in inputs should be unique.")
11 | }
12 | TRUE
13 | }
14 | )
15 |
16 | return(unlist(res))
17 |
18 | }
19 |
--------------------------------------------------------------------------------
/R/check_inputs_CARNIVAL.R:
--------------------------------------------------------------------------------
1 | #' Check Inputs For CARNIVAL
2 | #'
3 | #' Checks the format of the main data inputs for CARNIVL. Checks the data format
4 | #' and coverage of nodes in the PKN and data. Ensures that all nodes in
5 | #' input_data and measured_data appear in the PKN.
6 | #'
7 | #' @param meta_network Prior knowledge network (PKN). \dQuote{\code{data.frame}}
8 | #' object with source, sign and target columns. By default COSMOS uses a PKN
9 | #' derived from Omnipath, STITCHdb and Recon3D. See details on the data
10 | #' \code{\link{meta_network}}.
11 | #'
12 | #' @param input_data Numerical vector, where names are input nodes in the PKN
13 | #' and values are from \{1, 0, -1\}.
14 | #'
15 | #' @param measured_data Numerical vector, where names are measured nodes in the
16 | #' PKN and values are continuous values. These values are compared to with
17 | #' the simulation.
18 | #'
19 | #' @seealso \code{\link{default_CARNIVAL_options}}
20 | #' @noRd
21 | check_inputs_for_CARNIVAL <- function(meta_network,
22 | input_data,
23 | measured_data){
24 | # checking the data
25 | stopifnot(is.vector(input_data))
26 | stopifnot(is.vector(measured_data))
27 |
28 | # checking the networks
29 |
30 | stopifnot(is.data.frame(meta_network))
31 | stopifnot(all(c("source","interaction","target" ) %in% names(meta_network)))
32 | stopifnot(ncol(meta_network)==3)
33 |
34 | # check inputs and measurements are in the network
35 | stopifnot(all(names(input_data) %in% c(meta_network$source, meta_network$target)))
36 | stopifnot(all(names(measured_data) %in% c(meta_network$source, meta_network$target)))
37 |
38 | return(TRUE)
39 | }
--------------------------------------------------------------------------------
/R/check_network_data_coverage.R:
--------------------------------------------------------------------------------
1 |
2 | #' checks coverage between data and prior knowledge
3 | #'
4 | #' checks the following
5 | #' - signaling data nodes are in PKN
6 | #' - metabolic data nodes are in PKN
7 | #' - expression data genes overlap with TF targets
8 | #' Stops if a check fails
9 | #'
10 | #' @param meta_network contains the PKN
11 | #' @param tf_regulon (optional) tf-target network with EntrezID
12 | #' @param signaling_data numerical vector, where names are signaling nodes
13 | #' in the PKN and values are from \{1,-1\}
14 | #' @param metabolic_data numerical vector, where names are metabolic nodes
15 | #' in the PKN and values are continuous values. These values are compared to
16 | #' with the the simulation [? range of values]
17 | #' @param expression_data (optional) numerical vector, where names are gene names
18 | #' and values are from \{1,-1\}
19 | #' @param verbose (default: TRUE) reports coverage
20 | #' @noRd
21 |
22 |
23 | # expand_metabolic_data format metabolic data to match meta_network nodes?
24 | # It will add "XMetab__" before the pubchem IDs and all the possible compartments
25 | # after it. (e.g, "1150" will become "Xmetab__1150___c____" and "Xmetab__1150___e____")
26 |
27 | check_network_data_coverage <- function(meta_network,
28 | tf_regulon = NULL,
29 | signaling_data,
30 | metabolic_data,
31 | expression_data = NULL,
32 | verbose = TRUE){
33 |
34 | # signaling should be in PKN
35 | signaling_nodes = names(signaling_data)
36 | missing_nodes <- signaling_nodes[!signaling_nodes %in% c(meta_network$source,
37 | meta_network$target)]
38 |
39 | if(length(missing_nodes)>0){
40 |
41 | message_missing_nodes <- paste("The following signaling nodes are not found in the PKN:",
42 | limit_string_vec(missing_nodes))
43 | stop(message_missing_nodes)
44 |
45 |
46 | }else{
47 | if(verbose) print(paste("COSMOS: all", length(signaling_nodes),
48 | "signaling nodes from data were found in the meta PKN"))
49 | }
50 |
51 | # metabolic nodes should be in PKN
52 | metabolic_nodes = names(metabolic_data)
53 | missing_nodes <- metabolic_nodes[!metabolic_nodes %in% c(meta_network$source,
54 | meta_network$target)]
55 | if(length(missing_nodes)>0){
56 |
57 | message_missing_nodes <- paste("The following signaling nodes are not found in the PKN:",
58 | limit_string_vec(missing_nodes))
59 | stop(message_missing_nodes)
60 |
61 |
62 |
63 |
64 | }else{
65 | if(verbose) print(paste("COSMOS: all",length(metabolic_nodes) ,
66 | "metabolic nodes from data were found in the meta PKN"))
67 |
68 | }
69 |
70 | # the expression data should overlap with the TF targets
71 | if(!is.null(tf_regulon) & !is.null(expression_data)){
72 | genes = names(expression_data)
73 | genes_as_tf_target <- sum(genes %in% tf_regulon$target)
74 |
75 | if(verbose) print(paste0("COSMOS: ", genes_as_tf_target,
76 | " of the ",length(genes),
77 | " genes in expression data were found as transcription factor target"))
78 | if(verbose) print(paste0("COSMOS: ", sum( unique(tf_regulon$target) %in% genes),
79 | " of the ",length(unique(tf_regulon$target)),
80 | " transcription factor targets were found in expression data"))
81 | if(genes_as_tf_target==0){
82 | print(genes, tf_regulon$target)
83 | stop("Expression data contains no gene that appear as transcription factor target.
84 | The expression_data must be a named vector using gene symboles.")
85 | }
86 | }
87 |
88 |
89 | }
90 |
91 |
92 |
93 |
--------------------------------------------------------------------------------
/R/compress_same_children.R:
--------------------------------------------------------------------------------
1 | #' Compress Network by Merging Nodes with Identical Children
2 | #'
3 | #' This function compresses a network by merging nodes that have the same children.
4 | #' The input network is represented as a data frame with three columns: source, target, and sign of interaction.
5 | #' The function returns a list containing the compressed network, node signatures, and duplicated signatures.
6 | #'
7 | #' @param df A data frame representing the network with three columns: source, target, and sign of interaction.
8 | #' @param sig_input A list of input node signatures to be considered for the merging process.
9 | #' @param metab_input A list of input metabolic signatures to be considered for the merging process.
10 | #'
11 | #' @return A list containing the following elements:
12 | #' \item{compressed_network}{A data frame representing the compressed network.}
13 | #' \item{node_signatures}{A list of signatures of nodes in the network after the merging process.}
14 | #' \item{duplicated_signatures}{A list of duplicated signatures in the network after the merging process.}
15 | #'
16 | #' @examples
17 | #' # Create a sample network
18 | #' df <- data.frame(source = c("A", "A", "B", "B"),
19 | #' target = c("C", "D", "C", "D"),
20 | #' sign_of_interaction = c(1, 1, 1, 1))
21 | #'
22 | #' # Define input node and metabolic signatures
23 | #' sig_input <- list()
24 | #' metab_input <- list()
25 | #'
26 | #' # Compress the network
27 | #' result <- compress_same_children(df, sig_input, metab_input)
28 | #' compressed_network <- result$compressed_network
29 | #'
30 | #' @export
31 | compress_same_children <- function(df, sig_input, metab_input)
32 | {
33 | nodes <- unique(c(df$source,df$target))
34 |
35 | parents <- nodes[which(nodes %in% df$source)]
36 |
37 | df_signature <- df
38 | df_signature[,2] <- paste(df_signature[,2],df_signature[,3],sep = "")
39 |
40 | node_signatures <- sapply(parents, function(parent,df_signature){
41 |
42 | return(paste("parent_of_",paste0(unlist(df_signature[which(df_signature[,1] == parent),2]), collapse = "_____"), sep = ""))
43 | },df_signature = df_signature, USE.NAMES = T, simplify = F)
44 |
45 | dubs <- node_signatures[duplicated(node_signatures) &
46 | !(names(node_signatures) %in% names(metab_input) |
47 | names(node_signatures) %in% names(sig_input))]
48 |
49 | duplicated_parents <- unlist(node_signatures[which(node_signatures %in% dubs)])
50 |
51 | df[,1] <- sapply(df[,1], function(node,duplicated_parents){
52 | if(node %in% names(duplicated_parents))
53 | {
54 | node <- duplicated_parents[node]
55 | }
56 | return(node)
57 | },duplicated_parents = duplicated_parents, simplify = T)
58 |
59 | df[,2] <- sapply(df[,2], function(node,duplicated_parents){
60 | if(node %in% names(duplicated_parents))
61 | {
62 | node <- duplicated_parents[node]
63 | }
64 | return(node)
65 | },duplicated_parents = duplicated_parents, simplify = T)
66 |
67 | df <- unique(df)
68 |
69 | return(list("compressed_network" = df, "node_signatures" = node_signatures, "duplicated_signatures" = duplicated_parents))
70 | }
71 |
72 | #' Decompress Solution Network
73 | #'
74 | #' This function decompresses a solution network by mapping node signatures back to their original identifiers.
75 | #' The input is a formatted solution network, a meta network, node signatures, and duplicated parents.
76 | #' The function returns a list containing the decompressed solution network and attribute table.
77 | #'
78 | #' @param formatted_res A list containing the solution network and attribute table.
79 | #' @param meta_network A data frame representing the meta network.
80 | #' @param node_signatures A list of node signatures.
81 | #' @param duplicated_parents A list of duplicated parents from the compression process.
82 | #'
83 | #' @return A list containing the following elements:
84 | #' \item{SIF}{A data frame representing the decompressed solution network.}
85 | #' \item{ATT}{A data frame containing the attributes of the decompressed solution network.}
86 | #'
87 | #' @examples
88 | #' # Create a sample formatted_res
89 | #' formatted_res <- list(
90 | #' SIF = data.frame(source = c("parent_of_D1", "D"),
91 | #' target = c("D", "F"),
92 | #' interaction = c(1, 1),
93 | #' Weight = c(1, 1)),
94 | #' ATT = data.frame(Nodes = c("parent_of_D1", "D", "F"),
95 | #' NodeType = c("","",""),
96 | #' ZeroAct = c(0,0,0),
97 | #' UpAct = c(1,1,1),
98 | #' DownAct = c(0,0,0),
99 | #' AvgAct = c(1,1,1),
100 | #' measured = c(0,0,0),
101 | #' Activity = c(1,1,1))
102 | #' )
103 | #'
104 | #' # Create a sample meta_network
105 | #' meta_network <- data.frame(source = c("A", "B", "D"),
106 | #' target = c("D", "D", "F"),
107 | #' interaction_type = c(1, 1, 1))
108 | #'
109 | #' # Define node_signatures and duplicated_parents
110 | #' node_signatures <- list("A" = "parent_of_D1","B" = "parent_of_D1","D" = "parent_F1")
111 | #' duplicated_parents <- c("A" = "parent_of_D1","B" = "parent_of_D1")
112 | #'
113 | #' # Decompress the solution network
114 | #' result <- decompress_solution_network(formatted_res, meta_network, node_signatures, duplicated_parents)
115 | #' decompressed_network <- result[[1]]
116 | #' attribute_table <- result[[2]]
117 | #'
118 | #' @export
119 | decompress_solution_network <- function(formatted_res, meta_network, node_signatures, duplicated_parents)
120 | {
121 | SIF <- formatted_res[[1]]
122 | ATT <- formatted_res[[2]]
123 |
124 | duplicated_parents_df <- data.frame(duplicated_parents)
125 | duplicated_parents_df$source_original <- row.names(duplicated_parents_df)
126 | names(duplicated_parents_df)[1] <- "Nodes"
127 |
128 |
129 | addons <- data.frame(names(node_signatures)[-which(names(node_signatures) %in% duplicated_parents_df$source_original)])
130 | names(addons)[1] <- "Nodes"
131 | addons$source_original <- addons$Nodes
132 |
133 | mapping_table <- as.data.frame(rbind(duplicated_parents_df,addons))
134 |
135 | data("HMDB_mapper_vec")
136 |
137 | mapping_table[, 1] <- sapply(mapping_table[, 1], function(x, HMDB_mapper_vec) {
138 | x <- gsub("Metab__", "", x)
139 | x <- gsub("^Gene", "Enzyme", x)
140 | suffixe <- stringr::str_extract(x, "_[a-z]$")
141 | x <- gsub("_[a-z]$", "", x)
142 | if (x %in% names(HMDB_mapper_vec)) {
143 | x <- HMDB_mapper_vec[x]
144 | x <- paste("Metab__", x, sep = "")
145 | }
146 | if (!is.na(suffixe)) {
147 | x <- paste(x, suffixe, sep = "")
148 | }
149 | return(x)
150 | }, HMDB_mapper_vec = HMDB_mapper_vec)
151 |
152 | mapping_table[, 2] <- sapply(mapping_table[, 2], function(x, HMDB_mapper_vec) {
153 | x <- gsub("Metab__", "", x)
154 | x <- gsub("^Gene", "Enzyme", x)
155 | suffixe <- stringr::str_extract(x, "_[a-z]$")
156 | x <- gsub("_[a-z]$", "", x)
157 | if (x %in% names(HMDB_mapper_vec)) {
158 | x <- HMDB_mapper_vec[x]
159 | x <- paste("Metab__", x, sep = "")
160 | }
161 | if (!is.na(suffixe)) {
162 | x <- paste(x, suffixe, sep = "")
163 | }
164 | return(x)
165 | }, HMDB_mapper_vec = HMDB_mapper_vec)
166 |
167 | # View(ATT[!(ATT$Nodes %in% mapping_table$Nodes),])
168 |
169 |
170 |
171 | ATT <- merge(ATT, mapping_table, by = "Nodes")
172 | ATT <- ATT[,c(9,2:8)]
173 | names(ATT)[1] <- "Nodes"
174 |
175 | SIF <- meta_network
176 |
177 | SIF[, 1] <- sapply(SIF[, 1], function(x, HMDB_mapper_vec) {
178 | x <- gsub("Metab__", "", x)
179 | x <- gsub("^Gene", "Enzyme", x)
180 | suffixe <- stringr::str_extract(x, "_[a-z]$")
181 | x <- gsub("_[a-z]$", "", x)
182 | if (x %in% names(HMDB_mapper_vec)) {
183 | x <- HMDB_mapper_vec[x]
184 | x <- paste("Metab__", x, sep = "")
185 | }
186 | if (!is.na(suffixe)) {
187 | x <- paste(x, suffixe, sep = "")
188 | }
189 | return(x)
190 | }, HMDB_mapper_vec = HMDB_mapper_vec)
191 |
192 | SIF[, 2] <- sapply(SIF[, 2], function(x, HMDB_mapper_vec) {
193 | x <- gsub("Metab__", "", x)
194 | x <- gsub("^Gene", "Enzyme", x)
195 | suffixe <- stringr::str_extract(x, "_[a-z]$")
196 | x <- gsub("_[a-z]$", "", x)
197 | if (x %in% names(HMDB_mapper_vec)) {
198 | x <- HMDB_mapper_vec[x]
199 | x <- paste("Metab__", x, sep = "")
200 | }
201 | if (!is.na(suffixe)) {
202 | x <- paste(x, suffixe, sep = "")
203 | }
204 | return(x)
205 | }, HMDB_mapper_vec = HMDB_mapper_vec)
206 |
207 | SIF <- SIF[SIF$source %in% ATT$Nodes &
208 | SIF$target %in% ATT$Nodes,]
209 |
210 | SIF$Weight <- apply(SIF, 1, function(x, ATT){
211 | source_act <- ATT[ATT$Nodes == x[1],"Activity"]
212 | target_act <- ATT[ATT$Nodes == x[2],"Activity"]
213 | coherence <- source_act * target_act
214 | weight <- ifelse(sign(coherence) == sign(as.numeric(x[3])), min(abs(c(source_act, target_act))), 0)
215 | return(weight)
216 | }, ATT = ATT)
217 |
218 | SIF <- SIF[which(SIF$Weight != 0),]
219 | ATT <- ATT[ATT$AvgAct != 0,]
220 |
221 | seeds <- SIF[!(SIF[,1] %in% SIF[,2]),1]
222 | bad_seeds <- seeds[!(seeds %in% ATT[ATT$NodeType == "P","Nodes"])]
223 |
224 | if(length(bad_seeds) > 0)
225 | {
226 | ATT <- ATT[!(ATT$Nodes %in% bad_seeds),]
227 | SIF <- SIF[SIF$source %in% ATT$Nodes &
228 | SIF$target %in% ATT$Nodes,]
229 | }
230 |
231 | formatted_res <- list(SIF,ATT)
232 |
233 | return(formatted_res)
234 | }
--------------------------------------------------------------------------------
/R/cosmos_data.R:
--------------------------------------------------------------------------------
1 |
2 | # Constructor ------------------------------------------------------------------
3 | # for internal use
4 | #' Create New Cosmos Data
5 | #'
6 | #' Constructor. Should not be exported. The object is not validated.
7 | #' @param meta_network Prior knowledge network (PKN). By default COSMOS use a
8 | #' PKN derived from Omnipath, STITCHdb and Recon3D. See details on the data
9 | #' \code{\link{meta_network}}.
10 | #' @param tf_regulon Collection of transcription factor - target interactions.
11 | #' A default collection from dorothea can be obtained by the
12 | #' \code{\link{load_tf_regulon_dorothea}} function.
13 | #' @param signaling_data Numerical vector, where names are signaling nodes
14 | #' in the PKN and values are from \{1, 0, -1\}. Continuous data will be
15 | #' discretized using the \code{\link{sign}} function.
16 | #' @param metabolic_data Numerical vector, where names are metabolic nodes
17 | #' in the PKN and values are continuous values that represents log2 fold change
18 | #' or t-values from a differential analysis. These values are compared to
19 | #' the simulation results (simulated nodes can take value -1, 0 or 1).
20 | #' @param expression_data Numerical vector that represents the results of a
21 | #' differential gene expression analysis. Names are gene names using EntrezID
22 | #' starting with an X and values are log fold change or t-values. Genes with
23 | #' NA values are considered none expressed and they will be removed from the
24 | #' TF-gene expression interactions.
25 | #' @noRd
26 | new_cosmos_data <- function(meta_network = data.frame(),
27 | tf_regulon = data.frame(),
28 | signaling_data = double(),
29 | metabolic_data = double(),
30 | expression_data = double()){
31 |
32 | stopifnot(is.data.frame(meta_network))
33 | stopifnot(is.data.frame(tf_regulon))
34 | stopifnot(is.double(signaling_data))
35 | stopifnot(is.double(metabolic_data))
36 | stopifnot(is.double(expression_data) | is.null(expression_data))
37 |
38 |
39 | structure(list(
40 | meta_network = meta_network,
41 | tf_regulon = tf_regulon,
42 | signaling_data = signaling_data,
43 | metabolic_data = metabolic_data,
44 | expression_data = expression_data,
45 | history = c()
46 | ), class = "cosmos_data")
47 |
48 | }
49 |
50 | # validator -------------------------------------------------------------------
51 | #' Validate cosmos data
52 | #'
53 | #' Checks the cosmos object.
54 | #'
55 | #' @param x \code{\link{cosmos_data}} object. Use the
56 | #' \code{\link{preprocess_COSMOS_metabolism_to_signaling}} function to create
57 | #' one.
58 | #' @noRd
59 | validate_cosmos_data <- function(x){
60 |
61 | if(!all(c("meta_network", "tf_regulon", "signaling_data", "metabolic_data") %in% names(x))){
62 | stop(
63 | "All fields should present in the cosmos object: meta_network, tf_regulon, signaling_data, metabolic_data",
64 | call. = FALSE
65 | )
66 | }
67 |
68 | meta_network = x$meta_network
69 | tf_regulon = x$tf_regulon
70 | signaling_data = x$signaling_data
71 | metabolic_data = x$metabolic_data
72 | expression_data = x$expression_data
73 |
74 |
75 | check_COSMOS_inputs(meta_network = meta_network,
76 | tf_regulon = tf_regulon,
77 | expression_data = expression_data,
78 | signaling_data = signaling_data,
79 | metabolic_data = metabolic_data)
80 |
81 | check_gene_names(signaling_data)
82 |
83 | # Check overlap among node names in the inputs
84 | check_network_data_coverage(meta_network = meta_network,
85 | signaling_data = signaling_data,
86 | metabolic_data = metabolic_data,
87 | verbose = FALSE)
88 |
89 |
90 | }
91 |
92 |
93 |
94 |
95 |
96 | # helper -----------------------------------------------------------------------
97 | #' Create Cosmos Data
98 | #'
99 | #' An S3 class that combines the required data into a comprehensive list. Use
100 | #' the \code{\link{preprocess_COSMOS_signaling_to_metabolism}} or
101 | #' \code{\link{preprocess_COSMOS_metabolism_to_signaling}} to create an instance.
102 | #'
103 | #' @param meta_network Prior knowledge network (PKN). By default COSMOS use a
104 | #' PKN derived from Omnipath, STITCHdb and Recon3D. See details on the data
105 | #' \code{\link{meta_network}}.
106 | #' @param tf_regulon Collection of transcription factor - target interactions.
107 | #' A default collection from dorothea can be obtained by the
108 | #' \code{\link{load_tf_regulon_dorothea}} function.
109 | #' @param signaling_data Numerical vector, where names are signaling nodes
110 | #' in the PKN and values are from \{1, 0, -1\}. Continuous data will be
111 | #' discretized using the \code{\link{sign}} function.
112 | #' @param metabolic_data Numerical vector, where names are metabolic nodes
113 | #' in the PKN and values are continuous values that represents log2 fold change
114 | #' or t-values from a differential analysis. These values are compared to
115 | #' the simulation results (simulated nodes can take value -1, 0 or 1).
116 | #' @param expression_data Numerical vector that represents the results of a
117 | #' differential gene expression analysis. Names are gene names using EntrezID
118 | #' starting with an X and values are log fold change or t-values. Genes with
119 | #' NA values are considered none expressed and they will be removed from the
120 | #' TF-gene expression interactions.
121 | #' @param verbose (default: TRUE) Reports details about the
122 | #' \code{\link{cosmos_data}} object.
123 | #' @return \code{cosmos data} class instance.
124 | #'
125 |
126 | cosmos_data <- function(meta_network,
127 | tf_regulon = NULL,
128 | signaling_data,
129 | metabolic_data,
130 | expression_data,
131 | verbose = TRUE){
132 |
133 | check_gene_names(signaling_data)
134 |
135 | # signaling should be in PKN
136 | signaling_nodes = names(signaling_data)
137 | missing_nodes <- signaling_nodes[!signaling_nodes %in% c(meta_network$source,
138 | meta_network$target)]
139 | if(length(missing_nodes)>0){
140 | message_missing_nodes <- paste("The following signaling nodes are not found in the PKN and will be removed from input:",
141 | limit_string_vec(missing_nodes))
142 | message("COSMOS: ", message_missing_nodes)
143 | signaling_data <- signaling_data[signaling_nodes != missing_nodes]
144 |
145 |
146 | }else{
147 | if(verbose) print(paste("COSMOS: all", length(signaling_nodes),
148 | "signaling nodes from data were found in the meta PKN"))
149 | }
150 |
151 | # metabolic nodes should be in PKN
152 | metabolic_nodes = names(metabolic_data)
153 | missing_nodes <- metabolic_nodes[!metabolic_nodes %in% c(meta_network$source,
154 | meta_network$target)]
155 | if(length(missing_nodes)>0){
156 |
157 | message_missing_nodes <- paste("The following metabolic nodes are not found in the PKN and will be removed from input:",
158 | limit_string_vec(missing_nodes))
159 | message("COSMOS: ", message_missing_nodes)
160 | # remove metabolic inputs not matching into the meta_network
161 | metabolic_data <- metabolic_data[metabolic_nodes != missing_nodes]
162 |
163 | }else{
164 | if(verbose) print(paste("COSMOS: all",length(metabolic_nodes) ,
165 | "metabolic nodes from data were found in the meta PKN"))
166 | }
167 |
168 | # the expression data should overlap with the TF targets
169 | if(!is.null(tf_regulon) & !is.null(expression_data)){
170 | genes = names(expression_data)
171 | genes_as_tf_target <- sum(genes %in% tf_regulon$target)
172 |
173 | if(verbose) print(paste0("COSMOS: ", genes_as_tf_target,
174 | " of the ",length(genes),
175 | " genes in expression data were found as transcription factor target"))
176 | if(verbose) print(paste0("COSMOS: ", sum( unique(tf_regulon$target) %in% genes),
177 | " of the ",length(unique(tf_regulon$target)),
178 | " transcription factor targets were found in expression data"))
179 | if(genes_as_tf_target==0){
180 | stop("Expression data contains no gene that appear as transcription factor target.
181 | The expression_data must be a named vector using EntrezIDs.")
182 | }
183 | }
184 |
185 | new_cosmos_data(meta_network = meta_network,
186 | tf_regulon = tf_regulon,
187 | signaling_data = signaling_data,
188 | metabolic_data = metabolic_data,
189 | expression_data = expression_data)
190 |
191 |
192 |
193 | }
194 |
195 |
196 | ### Generic functions ------------------------------------------------------------------
197 | #' Print Cosmos Data Summary
198 |
199 | #' Print a summary of cosmos data.
200 |
201 | #' @param x \code{\link{cosmos_data}} object. Use the
202 | #' \code{\link{preprocess_COSMOS_signaling_to_metabolism}} or
203 | #' \code{\link{preprocess_COSMOS_metabolism_to_signaling}} functions to
204 | #' create one.
205 | #' @param ... Further print arguments passed to or from other methods.
206 | #'
207 | #' @seealso \code{\link[base]{print}}, \code{\link{cosmos_data}}
208 | #'
209 | #'
210 | #' @return input (invisible)
211 | #' @export
212 | print.cosmos_data <- function(x, ...) {
213 | cat("cosmos_data contains the following data: \n")
214 | cat("Meta network: ", nrow(x$meta_network), "interactions and", length(unique(c(x$meta_network$source,x$meta_network$target))), "unique nodes\n")
215 | cat("TF regulon: ", nrow(x$tf_regulon), "interactions,", length(unique(c(x$tf_regulon$tf))),"TFs and", length(unique(c(x$tf_regulon$target))), "targets\n")
216 | cat("Signaling data: ", length(x$signaling_data), "measured nodes\n")
217 | cat("Metabolic data: ", length(x$metabolic_data), "measured nodes\n")
218 | cat("Expression data:", length(x$expression_data), "measured nodes\n")
219 | if(!is.null(x$history)) print(x$history)
220 |
221 | invisible(x)
222 | }
223 |
--------------------------------------------------------------------------------
/R/default_CARNIVAL_options.R:
--------------------------------------------------------------------------------
1 | #' Setting Default CARNIVAL Options
2 | #'
3 | #' Returns the default CARNIVAL options as a list. You can modify the elements
4 | #' of the list and then use it as an argument in \code{\link{run_COSMOS_metabolism_to_signaling}} or
5 | #' \code{\link{run_COSMOS_signaling_to_metabolism}}.
6 | #' If you choose CPLEX or CBC, you must modify then the solverPath field and point to
7 | #' the CPLEX/CBC executable (See Details).
8 | #'
9 | #' @details COSMOS is dependent on CARNIVAL for exhibiting the signalling pathway optimisation.
10 | #' CARNIVAL requires the interactive version of IBM Cplex, Gurobi or CBC-COIN solver
11 | #' as the network optimiser. The IBM ILOG Cplex is freely available through
12 | #' Academic Initiative \href{https://www.ibm.com/products/ilog-cplex-optimization-studio}{here}.
13 | #' Gurobi license is also free for academics, request a license following instructions
14 | #' \href{https://www.gurobi.com/downloads/end-user-license-agreement-academic/}{here}.
15 | #' The \href{https://projects.coin-or.org/Cbc}{CBC} solver is open source and freely
16 | #' available for any user, but has a significantly lower performance than CPLEX or
17 | #' Gurobi. Obtain CBC executable directly usable for cosmos
18 | #'\href{https://ampl.com/products/solvers/open-source/#cbc}{here}. Alternatively for
19 | #' small networks, users can rely on the freely available
20 | #' \href{https://cran.r-project.org/web/packages/lpSolve/index.html}{lpSolve R-package},
21 | #' which is automatically installed with the package.
22 | #'
23 | #' @param solver one of `cplex` (recommended, but require 3rd party tool), `cbc` (also require 3rd party tool) or `lpSolve` (only for small networks)
24 | #' @return returns a list with all possible options implemented in CARNIVAL.
25 | #' see the documentation on \code{\link[CARNIVAL]{runCARNIVAL}}.
26 | #' @examples
27 | #' # load and change default options:
28 | #' my_options = default_CARNIVAL_options(solver = "cplex")
29 | #'
30 | #' my_options$solverPath = "/Applications/CPLEX_Studio128/cplex/bin/x86-64_osx/cplex"
31 | #' my_options$threads = 2
32 | #' my_options$timelimit = 3600*15
33 | #' @export
34 | #'
35 | default_CARNIVAL_options = function(solver = NULL){
36 |
37 | if(is.null(solver)){
38 | stop("please call default_CARNIVAL_options(solver = 'cplex') with a
39 | specific solver argument. Valid solvers are: 'lpSolve','cplex' or 'cbc'.")
40 | }
41 | solver_options = c("lpSolve","cplex","cbc")
42 | solver <- match.arg(solver,choices = solver_options)
43 |
44 | if(solver == "lpSolve"){
45 | opts = CARNIVAL::defaultLpSolveCarnivalOptions()
46 | }else if(solver == "cplex"){
47 | opts = CARNIVAL::defaultCplexCarnivalOptions()
48 | }else if(solver == "cbc"){
49 | opts = CARNIVAL::defaultCbcSolveCarnivalOptions()
50 | }
51 | opts$keepLPFiles = FALSE
52 |
53 | return(opts)
54 | }
55 |
56 |
57 |
58 | #' Check CARNIVAL Options
59 | #'
60 | #' Checks the list of input options for CARNIVAL for completeness.
61 | #'
62 | #' @param opts An object of type \dQuote{\code{list}} defining the run parameters
63 | #' for CARNIVAL. Use the \code{\link{default_CARNIVAL_options}} function to
64 | #' create a list with default parameter settings. If cplex or cbc are chosen
65 | #' as the solver, the parameter solverPath needs to be supplied (not
66 | #' automatically added by \code{default_CARNIVAL_options()}).
67 | #'
68 | #' @seealso \code{\link{default_CARNIVAL_options}},
69 | #' \code{\link[CARNIVAL]{runCARNIVAL}}
70 | #' @noRd
71 | check_CARNIVAL_options <- function(opts){
72 |
73 |
74 | if(!is.list(opts)) stop("CARNIVAL options should be a list")
75 | req_names <- c("solver")
76 |
77 |
78 | if(!all(req_names %in% names(opts))){
79 | stop("CARNIVAL options should contain all options.
80 | Start by calling CARNIVAL::defaultCbcCarnivalOptions() for CPLEX or
81 | CARNIVAL::defaultCplexCarnivalOptions() for CBC solver, or
82 | CARNIVAL::defaultLpSolveCarnivalOptions() for LpSolve and replace entries. ")
83 | }
84 |
85 |
86 | if(!(opts$solver %in% c("cplex","cbc","lpSolve"))) stop("current version supports only the CPLEX or cbc solver. lpSolve is also available for test runs.")
87 | if(opts$solver == "cbc") print("We recommend the users to use CPLEX if possible.")
88 | if(opts$solver == "lpSolve") print("lpSolve does not scale well with large PKNs. This solver is mainly for testing purposes. To run COSMSO, we recommend using cplex, or cbc solvers.")
89 | if(is.null(opts$solverPath) & opts$solver != "lpSolve") stop("path to CPLEX or cbc solver must be provided")
90 |
91 | return('List of CARNIVAL options is complete.')
92 |
93 |
94 | }
95 |
--------------------------------------------------------------------------------
/R/discretize_vector.R:
--------------------------------------------------------------------------------
1 | #' discretize input vector
2 | #'
3 | #' convert a vector of continuous values to discrete {-1, 0, 1}
4 | #'
5 | #' @param input_vec numerical value
6 | #' @noRd
7 |
8 |
9 | discretize_input <- function(input_vec){
10 |
11 |
12 | if(!all(input_vec %in% c(-1,0,1))){
13 | message("Input nodes should have values from {-1, 0, 1}. We discretize your input with sign().")
14 | res = sign(input_vec)
15 | }else{
16 | res = input_vec
17 | }
18 |
19 | return(res)
20 | }
--------------------------------------------------------------------------------
/R/display_node_neighboorhood.R:
--------------------------------------------------------------------------------
1 | #' display_node_neighboorhood
2 | #'
3 | #' display input and measurements within n steps of a given set of nodes
4 | #' @param central_node character or character vector; node ID(s) around which a
5 | #' network will be branched out untill meansurments and input are reached
6 | #' @param sif df; COSMOS network solution in sif format like the first list
7 | #' element returned by the format_cosmos_res function
8 | #' @param att df; attributes of the nodes of the COMSOS network solution like
9 | #' the second list element returned by the format_cosmos_res function
10 | #' @param n numeric; maximum number of steps in the network to look for inputs
11 | #' and measurments
12 | #' @return a visnetwork object
13 | #' @importFrom dplyr %>%
14 | #' @examples
15 | #' CARNIVAL_options <- cosmosR::default_CARNIVAL_options("lpSolve")
16 | #' data(toy_network)
17 | #' data(toy_signaling_input)
18 | #' data(toy_metabolic_input)
19 | #' data(toy_RNA)
20 | #' test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
21 | #' signaling_data = toy_signaling_input,
22 | #' metabolic_data = toy_metabolic_input,
23 | #' diff_expression_data = toy_RNA,
24 | #' maximum_network_depth = 15,
25 | #' remove_unexpressed_nodes = TRUE,
26 | #' CARNIVAL_options = CARNIVAL_options
27 | #' )
28 | #' test_result_for <- run_COSMOS_signaling_to_metabolism(data = test_for,
29 | #' CARNIVAL_options = CARNIVAL_options)
30 | #' test_result_for <- format_COSMOS_res(test_result_for)
31 | #' network_plot <- display_node_neighboorhood(central_node = 'MYC',
32 | #' sif = test_result_for[[1]],
33 | #' att = test_result_for[[2]],
34 | #' n = 7)
35 | #' network_plot
36 | #' @export
37 | display_node_neighboorhood <- function(central_node,sif, att, n = 100)
38 | {
39 | # require(igraph)
40 | full_sif <- sif
41 | if(sum(names(full_sif) %in% c("Node1", "Node2", "Sign", "Weight")) != 4)
42 | {
43 | print('sif should be a dataframe with 4 column named "Node1", "Node2", "Sign" and "Weight"')
44 | return('Error:bad sif')
45 | }
46 | full_sif <- full_sif[,c("Node1", "Node2", "Sign", "Weight")]
47 | full_att <- att
48 |
49 | ig_net <- igraph::graph_from_data_frame(full_sif)
50 |
51 | ig_net <- igraph::make_ego_graph(ig_net, nodes = central_node, order = n, mode = "all")
52 |
53 | to_keep <- unlist(sapply(ig_net,function(x){igraph::V(x)$name}))
54 |
55 | full_sif <- full_sif[full_sif$Node1 %in% to_keep & full_sif$Node2 %in% to_keep,]
56 | full_att <- full_att[full_att$Nodes %in% to_keep,]
57 |
58 | ig_net <- igraph::graph_from_data_frame(full_sif)
59 |
60 | center_node <- sapply(central_node, function(x, ig_net, full_att) {
61 | names(unlist(igraph::get.shortest.paths(ig_net, from = x, to = full_att[full_att$measured == 1,1])$vpath))
62 | }, ig_net= ig_net, full_att = full_att, simplify = FALSE, USE.NAMES = TRUE)
63 |
64 | center_node <- unlist(center_node)
65 |
66 | center_node_out <- full_sif[full_sif$Node1 %in% center_node & full_sif$Node2 %in% center_node,]
67 |
68 | center_node <- sapply(central_node, function(x, ig_net, full_att) {
69 | names(unlist(igraph::get.shortest.paths(ig_net, from = x, to = full_att[full_att$measured == 1,1], mode = "in")$vpath))
70 | }, ig_net= ig_net, full_att = full_att, simplify = FALSE, USE.NAMES = TRUE)
71 |
72 | center_node <- unlist(center_node)
73 |
74 | center_node_in <- full_sif[full_sif$Node1 %in% center_node & full_sif$Node2 %in% center_node,]
75 |
76 | center_node_net <- as.data.frame(rbind(center_node_in,center_node_out))
77 |
78 | nodes <- full_att[full_att$Nodes %in% center_node_net$Node1 | full_att$Nodes %in% center_node_net$Node2,]
79 | edges <- center_node_net
80 |
81 | names(edges) <- c("from","to","sign","weigth")
82 | edges$color <- ifelse(edges$sign == 1, "grey","grey")
83 |
84 | # edges$arrows <- "to"
85 |
86 | edges$arrows.to.type <- ifelse(edges$sign == 1, "arrow","circle")
87 | edges$enabled <- TRUE
88 | edges$scaleFactor <- 1
89 |
90 | edges <- unique(edges)
91 |
92 | names(nodes)[1] <- "id"
93 | nodes$label <- nodes$id
94 | nodes$color <- ifelse(nodes$Activity > 0, "green","red")
95 | nodes <- nodes[!duplicated(nodes$id),]
96 | nodes$shape <- "dot"
97 | nodes[nodes$type == "metab_enzyme","shape"] <- "square"
98 | nodes[nodes$type == "protein","shape"] <- "square"
99 | nodes[nodes$type == "Kinase","shape"] <- "triangle"
100 | nodes[nodes$type == "TF","shape"] <- "diamond"
101 | nodes <- nodes[order(nodes$id),]
102 | nodes$shadow <- (nodes$measured == 1)
103 |
104 | return(visNetwork::visNetwork(nodes, edges, width = 1600, height = 1600) %>%
105 | visNetwork::visOptions(highlightNearest = TRUE,
106 | nodesIdSelection = list(enabled = TRUE,
107 | style = 'width: 200px; height: 26px;
108 | background: #f8f8f8;
109 | color: darkblue;
110 | border:none;
111 | outline:none;')))
112 |
113 | }
--------------------------------------------------------------------------------
/R/filter_input_nodes_not_in_pkn.R:
--------------------------------------------------------------------------------
1 |
2 | filter_input_nodes_not_in_pkn <- function(data,pkn){
3 |
4 | new_data = data[names(data) %in% c(pkn$source,pkn$target)]
5 |
6 | if(length(data) != length(new_data)){
7 | removed_nodes <- names(data)[!names(data) %in% names(new_data)]
8 |
9 | print(paste("COSMOS:",length(removed_nodes),
10 | "input/measured nodes are not in PKN any more:",
11 | limit_string_vec(removed_nodes)))
12 | }
13 | return(new_data)
14 | }
15 |
--------------------------------------------------------------------------------
/R/filter_pkn_expressed_genes.R:
--------------------------------------------------------------------------------
1 | #' filter_pkn_expressed_genes
2 | #'
3 | #' filters the non-expressed genes from the prior knowledge network
4 | #'
5 | #' @param expressed_genes_entrez Gene_symboles of expressed genes (mut be same ID as used in meta_PKN)
6 | #' @param meta_pkn COSMOS prior knowledge network
7 | #' @noRd
8 | filter_pkn_expressed_genes <- function(expressed_genes_entrez,meta_pkn){
9 |
10 | print(paste("COSMOS: removing unexpressed nodes from PKN..."))
11 |
12 | is_expressed <- function(x)
13 | {
14 | if(!grepl("Metab",x) & !grepl("orphanReac",x))
15 | {
16 | if(x %in% expressed_genes_entrez)
17 | {
18 | return(x)
19 | } else
20 | {
21 | if(grepl("Gene[0-9]+__[A-Z0-9_]+$",x))
22 | {
23 | genes <- gsub("Gene[0-9]+__","",x)
24 | genes <- strsplit(genes,"_")[[1]]
25 | if(sum(genes %in% expressed_genes_entrez) != length(genes))
26 | {
27 | return(NA)
28 | } else
29 | {
30 | return(x)
31 | }
32 | } else
33 | {
34 | if(grepl("Gene[0-9]+__[^_][a-z]",x))
35 | {
36 | print(x)
37 | return(x)
38 | } else
39 | {
40 | if(grepl("Gene[0-9]+__[A-Z0-9_]+reverse",x))
41 | {
42 | genes <- gsub("Gene[0-9]+__","",x)
43 | genes <- gsub("_reverse","",genes)
44 | genes <- strsplit(genes,"_")[[1]]
45 | if(sum(genes %in% expressed_genes_entrez) != length(genes))
46 | {
47 | return(NA)
48 | } else
49 | {
50 | return(x)
51 | }
52 | } else
53 | {
54 | return(NA)
55 | }
56 | }
57 | }
58 | }
59 | } else
60 | {
61 | return(x)
62 | }
63 | }
64 |
65 | # is_expressed("XGene3004__124975_91227")
66 |
67 | meta_pkn$source <- sapply(meta_pkn$source,is_expressed)
68 | n_removed = sum(!stats::complete.cases(meta_pkn))
69 | meta_pkn <- meta_pkn[stats::complete.cases(meta_pkn),]
70 | meta_pkn$target <- sapply(meta_pkn$target,is_expressed)
71 | meta_pkn <- meta_pkn[stats::complete.cases(meta_pkn),]
72 | n_removed = n_removed + sum(!stats::complete.cases(meta_pkn))
73 | print(paste("COSMOS:", n_removed, "interactions removed"))
74 | return(meta_pkn)
75 | }
76 |
77 | #' filter_pkn_expressed_genes_fast
78 | #'
79 | #' filters the non-expressed genes from the prior knowledge network
80 | #' USING A FASTER OPTIMISED ALGORITHM
81 | #'
82 | #' @param expressed_genes_entrez Gene_symboles of expressed genes (mut be same ID as used in meta_PKN)
83 | #' @param meta_pkn COSMOS prior knowledge network
84 | #' @noRd
85 | filter_pkn_expressed_genes_fast <- function(expressed_genes_entrez, meta_pkn) {
86 | message("COSMOS: removing unexpressed nodes from PKN...")
87 |
88 | # Create an environment to act as a hash table for fast membership checks.
89 | expressed_env <- new.env(hash = TRUE, parent = emptyenv())
90 | for (gene in expressed_genes_entrez) {
91 | expressed_env[[gene]] <- TRUE
92 | }
93 |
94 | vectorized_is_expressed <- function(x, env) {
95 | out <- rep(NA_character_, length(x))
96 |
97 | # 1. Nodes that contain "Metab" or "orphanReac" are kept as is.
98 | keep_idx <- grepl("Metab|orphanReac", x, perl = TRUE)
99 | out[keep_idx] <- x[keep_idx]
100 |
101 | # 2. Direct membership check: if the node exists in the environment.
102 | not_checked <- which(is.na(out))
103 | if (length(not_checked) > 0) {
104 | direct <- sapply(x[not_checked], function(s) exists(s, envir = env, inherits = FALSE))
105 | out[not_checked[direct]] <- x[not_checked[direct]]
106 | }
107 |
108 | # 3. For nodes that follow the pattern "Gene...":
109 | not_checked <- which(is.na(out))
110 | if (length(not_checked) > 0) {
111 | # Pattern for concatenated genes (e.g. "Gene1234__ID1_ID2")
112 | pattern_concat <- grepl("^Gene[0-9]+__[A-Z0-9_]+$", x[not_checked], perl = TRUE)
113 | if (any(pattern_concat)) {
114 | idx <- not_checked[pattern_concat]
115 | out[idx] <- sapply(x[idx], function(s) {
116 | genes_str <- sub("^Gene[0-9]+__", "", s)
117 | genes <- unlist(strsplit(genes_str, "_"))
118 | if (all(sapply(genes, exists, envir = env, inherits = FALSE))) s else NA_character_
119 | })
120 | }
121 |
122 | # Pattern for reverse-tagged genes (e.g. "Gene1234__ID1_ID2_reverse")
123 | not_checked <- which(is.na(out))
124 | pattern_reverse <- grepl("^Gene[0-9]+__[A-Z0-9_]+_reverse$", x[not_checked], perl = TRUE)
125 | if (any(pattern_reverse)) {
126 | idx <- not_checked[pattern_reverse]
127 | out[idx] <- sapply(x[idx], function(s) {
128 | genes_str <- sub("^Gene[0-9]+__", "", s)
129 | genes_str <- sub("_reverse$", "", genes_str)
130 | genes <- unlist(strsplit(genes_str, "_"))
131 | if (all(sapply(genes, exists, envir = env, inherits = FALSE))) s else NA_character_
132 | })
133 | }
134 |
135 | # Special pattern with a lower-case letter in the tag:
136 | not_checked <- which(is.na(out))
137 | pattern_lower <- grepl("^Gene[0-9]+__[^_][a-z]", x[not_checked], perl = TRUE)
138 | if (any(pattern_lower)) {
139 | idx <- not_checked[pattern_lower]
140 | # Optionally, print messages for these cases.
141 | for (s in x[idx]) message("Note: ", s)
142 | out[idx] <- x[idx]
143 | }
144 | }
145 |
146 | return(out)
147 | }
148 |
149 | meta_pkn$source <- vectorized_is_expressed(meta_pkn$source, expressed_env)
150 | meta_pkn$target <- vectorized_is_expressed(meta_pkn$target, expressed_env)
151 |
152 | initial_n <- nrow(meta_pkn)
153 | meta_pkn <- meta_pkn[stats::complete.cases(meta_pkn), ]
154 | removed <- initial_n - nrow(meta_pkn)
155 | message("COSMOS: ", removed, " interactions removed")
156 |
157 | return(meta_pkn)
158 | }
159 |
160 |
--------------------------------------------------------------------------------
/R/filter_transcriptional_regulations.R:
--------------------------------------------------------------------------------
1 | #' Filter Based On Transcriptional Regulations
2 | #'
3 | #' Filters interactions between TFs and genes in the PKN by the following points:
4 | #' - finds the TF target genes that change strongly (threshold)
5 | #' - finds TFs from the inputs (known TF activity)
6 | #' - remove interactions if the target gene expression doesn't change (regulation
7 | #' should not go through this node )
8 | #' - remove interactions if the TF activity is not in agreement with the change
9 | #' in the target gene expression (other factors influence these genes)
10 | #'
11 | #' @param network Prior knowledge network (PKN). By default COSMOS use a PKN
12 | #' derived from Omnipath, STITCHdb and Recon3D. See details on the data
13 | #' \code{\link{meta_network}}.
14 | #' @param gene_expression_binarized Named vector of {-1,0,1} difnining which
15 | #' genes changed.
16 | #' @param signaling_data Named vector containing known activity of signaling
17 | #' nodes (kinases, TFs).
18 | #' @param tf_regulon Collection of transcription factor - target interactions.
19 | #' A default collection from dorothea can be obtained by the
20 | #' \code{\link{load_tf_regulon_dorothea}} function.
21 | #' @return A filtered version of the network.
22 | #' @importFrom rlang .data
23 | #' @importFrom dplyr %>%
24 | #' @noRd
25 | filter_transcriptional_regulations <- function(network,
26 | gene_expression_binarized,
27 | signaling_data,
28 | tf_regulon)
29 | {
30 | check_COSMOS_inputs(meta_network = network,
31 | tf_regulon = tf_regulon,
32 | signaling_data = signaling_data,
33 | expression_data = gene_expression_binarized)
34 |
35 | # map the TF values and gene expression levels on the PKN
36 |
37 | gene_exp_df = data.frame(gene =names(gene_expression_binarized),
38 | target_sign = gene_expression_binarized)
39 |
40 | signaling_df = data.frame(TF =names(signaling_data),
41 | TF_sign = signaling_data) %>%
42 | dplyr::filter(.data$TF %in% tf_regulon$tf)
43 |
44 | annotated_network <- network %>%
45 | # add the gene expression data: (non-genes will be filled with NA)
46 | dplyr::left_join(gene_exp_df, by=c(target="gene")) %>%
47 | # add the TF data (non-TFs will be filled with NAs)
48 | dplyr::left_join( signaling_df, by=c(source="TF")) %>%
49 | dplyr::mutate(source_is_TF = .data$source %in% tf_regulon$tf)
50 |
51 | # find interactions where TF regulates a gene, but target
52 | # - does not change
53 | # - target gene is not measured
54 | # - target changes inconsistently
55 | annotated_network <- annotated_network %>%
56 | # genes didn't change
57 | dplyr::mutate(target_gene_unchanged = .data$target_sign==0) %>%
58 | # gene didnt change and the source is a TF (it is a transcriptional regulation)
59 | dplyr::mutate(TF_target_unchanged = .data$source_is_TF &
60 | (.data$target_gene_unchanged | is.na(.data$target_gene_unchanged))) %>%
61 | # gene changed, but not consistently with TF activity
62 | dplyr::mutate(inconsistent_TF_gene_sign =
63 | sign(.data$TF_sign) != interaction * sign(.data$target_sign) )
64 |
65 | # TODO: option to return these interactions
66 | removed_interactions <- annotated_network %>%
67 | dplyr::filter(.data$TF_target_unchanged | .data$inconsistent_TF_gene_sign)
68 |
69 | print(paste("COSMOS: ", nrow(removed_interactions),
70 | "interactions are removed from the PKN based on",
71 | "consistency check between TF activity and gene expression"))
72 |
73 |
74 | kept_interactions <- annotated_network %>%
75 | dplyr::filter(!.data$TF_target_unchanged | is.na(.data$TF_target_unchanged)) %>%
76 | dplyr::filter(!.data$inconsistent_TF_gene_sign | is.na(.data$inconsistent_TF_gene_sign))
77 |
78 | out_pkn <- kept_interactions %>% dplyr::select(.data$source,.data$interaction,.data$target)
79 | return(out_pkn)
80 | }
81 |
82 |
83 |
--------------------------------------------------------------------------------
/R/format_COSMOS_results.R:
--------------------------------------------------------------------------------
1 | #' format_COSMOS_res
2 | #'
3 | #' formats the network with readable names
4 | #'
5 | #' @param cosmos_res results of COSMOS run
6 | #' @param metab_mapping a named vector with HMDB Ids as names and desired metabolite names as values.
7 | #' @return list with network and attribute tables.
8 | #' @importFrom stringr str_extract
9 | #' @export
10 | format_COSMOS_res <- function(cosmos_res,
11 | metab_mapping = NULL)
12 | {
13 |
14 | if(is.null(metab_mapping))
15 | {
16 | data("HMDB_mapper_vec",package = "cosmosR",envir = environment())
17 | }
18 | SIF <- as.data.frame(cosmos_res$weightedSIF)
19 | ATT <- as.data.frame(cosmos_res$nodesAttributes)
20 | colnames(ATT)[1] <- "Nodes"
21 | ATT$measured <- ifelse(ATT$NodeType %in% c("M","T","S","P"),1,0)
22 | ATT$Activity <- ATT$AvgAct
23 |
24 | for(i in c(1,3))
25 | {
26 | SIF[,i] <- sapply(SIF[,i], function(x, HMDB_mapper_vec){
27 | x <- gsub("Metab__","",x)
28 | x <- gsub("^Gene","Enzyme",x)
29 | suffixe <- stringr::str_extract(x,"_[a-z]$")
30 | x <- gsub("_[a-z]$","",x)
31 | if(x %in% names(HMDB_mapper_vec))
32 | {
33 | x <- HMDB_mapper_vec[x]
34 | x <- paste("Metab__",paste(x,suffixe,sep = ""),sep = "")
35 | }
36 | return(x)
37 | },HMDB_mapper_vec = HMDB_mapper_vec)
38 | }
39 |
40 | ATT[,1] <- sapply(ATT[,1], function(x, HMDB_mapper_vec){
41 | x <- gsub("Metab__","",x)
42 | x <- gsub("^Gene","Enzyme",x)
43 | suffixe <- stringr::str_extract(x,"_[a-z]$")
44 | x <- gsub("_[a-z]$","",x)
45 | if(x %in% names(HMDB_mapper_vec))
46 | {
47 | x <- HMDB_mapper_vec[x]
48 | x <- paste("Metab__",x,sep = "")
49 | }
50 | if(!is.na(suffixe))
51 | {
52 | x <- paste(x,suffixe,sep = "")
53 | }
54 | return(x)
55 | },HMDB_mapper_vec = HMDB_mapper_vec)
56 |
57 | return(list(SIF,ATT))
58 | }
--------------------------------------------------------------------------------
/R/get_TF_activity_from_CARNIVAL.R:
--------------------------------------------------------------------------------
1 | #' get_TF_activity_from_CARNIVAL
2 | #'
3 | #' screens the CARNIVAL results and obtains the activity of transcription factors
4 | #'
5 | #' @param carnival_result list obtained from runCARNIVAL
6 | #' @param TFs character vector of transcription factors using EntrezIDs
7 | #' @return named numerical vector with TF activity found in CARNIVAL results.
8 | #' @importFrom rlang .data
9 | #' @importFrom dplyr %>% filter select as_tibble
10 | #' @noRd
11 | get_TF_activity_from_CARNIVAL <- function(carnival_result, TFs)
12 | {
13 |
14 | if(!validate_CARNIVAL_results(carnival_result)) warning("we failed to validate CARNIVAL results.")
15 |
16 | estimated_activity <- dplyr::as_tibble(carnival_result$nodesAttributes) %>%
17 | dplyr::select(.data$Node,.data$AvgAct) %>%
18 | dplyr::filter(.data$AvgAct!=0) %>%
19 | dplyr::filter(.data$Node %in% TFs)
20 |
21 | activity = as.numeric(estimated_activity$AvgAct)/100
22 | names(activity) = estimated_activity$Node
23 |
24 | return(activity)
25 | }
26 |
27 | #' Validate CARNIVAL Results
28 | #'
29 | #' Implement here basic checks to see if we got back from CARNIVAL what we
30 | #' expected. Subject to change if CARNIVAL results changes.
31 | #'
32 | #' @param CR CARNIVAL result object
33 | #' @noRd
34 | validate_CARNIVAL_results <- function(CR){
35 |
36 | list_names = c("weightedSIF","nodesAttributes","sifAll","attributesAll" ) %in% names(CR)
37 | if(!all(list_names)) return(FALSE)
38 |
39 | if(is.null(CR$weightedSIF)) return(FALSE)
40 | if(is.null(CR$nodesAttributes)) return(FALSE)
41 | if(is.null(CR$sifAll)) return(FALSE)
42 | if(is.null(CR$attributesAll)) return(FALSE)
43 |
44 | return(TRUE)
45 | }
46 |
--------------------------------------------------------------------------------
/R/keep_downstream_neighbours.R:
--------------------------------------------------------------------------------
1 | #' keep controllable neighbors
2 | #'
3 | #' keeps the nodes in network that are no more then n_steps away from the starting nodes.
4 | #' @param network network in 3 column data.frame format
5 | #' (source, interaction, target)
6 | #' @param n_steps largest distance t consider
7 | #' @param input_nodes names of the input nodes in the network
8 | #' @noRd
9 | keep_controllable_neighbours <- function(network, n_steps, input_nodes)
10 | {
11 | stopifnot(all(c("source","target","interaction") %in% colnames(network)))
12 | stopifnot(all(input_nodes %in% c(network$source,network$target)))
13 |
14 | print(paste("COSMOS: removing nodes that are not reachable from inputs within",n_steps,"steps"))
15 | meta_g <- igraph::graph_from_data_frame(network[,c("source",'target',"interaction")],directed = TRUE)
16 | dn_nbours <- igraph::ego(graph = meta_g, order = n_steps,nodes = input_nodes, mode = "out")
17 |
18 | sub_nodes <- c(unique(names(unlist(dn_nbours))), input_nodes)
19 |
20 | to_keep = network$source %in% sub_nodes & network$target %in% sub_nodes
21 |
22 | print(paste("COSMOS:", sum(!to_keep), "from ", length(to_keep),
23 | "interactions are removed from the PKN"))
24 |
25 | network <- network[to_keep,]
26 |
27 |
28 | return(network)
29 | }
30 |
31 | #' keep observable neighbors
32 | #'
33 | #' keeps the nodes in network that are no more then n_steps upstreams from the
34 | #' measured nodes
35 | #' @param network network in 3 column data.frame format
36 | #' (source, interaction, target)
37 | #' @param n_steps largest distance t consider
38 | #' @param observed_nodes names of the measured nodes in the network
39 | #' @noRd
40 | keep_observable_neighbours <- function(network, n_steps, observed_nodes)
41 | {
42 | stopifnot(all(c("source","target","interaction") %in% colnames(network)))
43 | stopifnot(all(observed_nodes %in% c(network$source,network$target)))
44 |
45 | print(paste("COSMOS: removing nodes that are not observable by measurements within",n_steps,"steps"))
46 | meta_g <- igraph::graph_from_data_frame(network[,c("source",'target',"interaction")],directed = TRUE)
47 | up_nbours <- igraph::ego(graph = meta_g, order = n_steps, nodes = observed_nodes, mode = "in")
48 |
49 | sub_nodes <- c(unique(names(unlist(up_nbours))), observed_nodes)
50 |
51 | to_keep = network$source %in% sub_nodes & network$target %in% sub_nodes
52 |
53 | print(paste("COSMOS:", sum(!to_keep), "from ", length(to_keep),
54 | "interactions are removed from the PKN"))
55 |
56 | network <- network[to_keep,]
57 |
58 |
59 | return(network)
60 | }
--------------------------------------------------------------------------------
/R/limit_string_vec.R:
--------------------------------------------------------------------------------
1 | #' Limit String Vec
2 | #'
3 | #' Writes the first 6 elements and the number of other elements in a string.
4 | #'
5 | #' @param str_vec A string vector to be transformed.
6 | #'
7 | #' @noRd
8 | limit_string_vec <- function(str_vec){
9 | paste(paste(utils::head(str_vec),collapse = ", "), "and",
10 | length(str_vec) - length(utils::head(str_vec)), "more.")
11 | }
12 |
--------------------------------------------------------------------------------
/R/load_tf_regulon_dorothea.R:
--------------------------------------------------------------------------------
1 | #' load transcription factor regulon
2 | #'
3 | #' load the transcription factors from \code{DOROTHEA} package and converts
4 | #' gene symbols to EntrezID using org.Hs.eg.db
5 | #'
6 | #' @param confidence strong vector (by default: c("A","B","C")). Subset of \{A, B,
7 | #' C, D, E\}. See the `dorothea` for the meaning of confidence levels.
8 | #' package for further details.
9 | #' @return returns a PKN of a form of a data table. Each row is an interaction.
10 | #' Columns names are:
11 | #'
12 | #' - `tf` transcription factor
13 | #' - `confidence` class of confidence
14 | #' - `target` target gene
15 | #' - `sign` indicates if interaction is up (1) or down-regulation (-1).
16 | #'
17 | #' @importFrom dplyr %>%
18 | #' @export
19 | #' @examples
20 | #' load_tf_regulon_dorothea()
21 | load_tf_regulon_dorothea <- function(confidence = c("A","B","C")){
22 | . <- NULL
23 |
24 | # load regulon from dorothea:
25 | regulon = dorothea::dorothea_hs
26 | regulon <- regulon %>% dplyr::rename(sign = "mor")
27 |
28 | conf = confidence
29 | regulon <- regulon %>% dplyr::filter(.data$confidence %in% conf)
30 | regulon <- regulon[,-which(names(regulon) == "confidence")]
31 |
32 | return(regulon)
33 | }
34 |
35 |
36 |
37 |
--------------------------------------------------------------------------------
/R/meta_network.R:
--------------------------------------------------------------------------------
1 | #' Meta Prior Knowledge Network
2 | #'
3 | #' Comprehensive Prior Knowledge Network (PKN), which combines signaling and
4 | #' metabolic interaction networks. The network was constructed using the Recon3D
5 | #' and STITCH metabolic networks as well as the signaling network from
6 | #' OmniPath.
7 | #'
8 | #' @docType data
9 | #'
10 | #' @usage data(meta_network)
11 | #'
12 | #' @format An object of class \dQuote{\code{tibble}} with 117065 rows
13 | #' (interactions) and three variables:
14 | #' \describe{
15 | #' \item{\code{source}}{Source node, either metabolite or protein}
16 | #' \item{\code{interaction}}{Type of interaction, 1 = Activation, -1 = Inhibition}
17 | #' \item{\code{target}}{Target node, either metabolite or protein}
18 | #' A detailed description of the identifier formatting can be found under
19 | #' \url{https://metapkn.omnipathdb.org/00__README.txt}.
20 | #' }
21 | #'
22 | #' @source The network is available in Omnipath:
23 | #' \url{https://metapkn.omnipathdb.org/metapkn__20200122.txt}, the scripts
24 | #' used for the build of the network are available under
25 | #' \url{https://github.com/saezlab/Meta_PKN}.
26 | #'
27 | #' @references {
28 | #' Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
29 | #' Systems Biology}. \bold{17}, e9730.
30 | #' }
31 | #'
32 | #' @examples
33 | #' data(meta_network)
34 | #'
35 | #' @keywords datasets
36 | "meta_network"
--------------------------------------------------------------------------------
/R/prepare_metab_inputs.R:
--------------------------------------------------------------------------------
1 | #' add metabolic compartment and metab__ prefix to metabolite IDs
2 | #'
3 | #' This function adds metabolic compartments to the metabolic identifiers provided by the user.
4 | #'
5 | #' @param metab_input a named vector with matebolic statistics as inputs and metabolite identifiers as names
6 | #' @param compartment_codes a character vector, the desired compartment codes to be added. Possible values are "r", "c", "e", "x", "m", "l", "n" and "g"
7 | #' @export
8 | #' @return a named vector with the compartment code and prefixed added to the names
9 | prepare_metab_inputs <- function(metab_input, compartment_codes)
10 | {
11 | comps <- c("r", "c", "e", "x", "m", "l", "n", "g")
12 |
13 | if(length(compartment_codes[which(!(compartment_codes %in% comps))]) > 0)
14 | {
15 | ignored <- compartment_codes[which(!(compartment_codes %in% comps))]
16 | print("The following compartment codes are not found in the PKN and will be ignored")
17 | print(ignored)
18 | }
19 |
20 | compartment_codes <- compartment_codes[which(compartment_codes %in% comps)]
21 |
22 | if(length(compartment_codes) == 0)
23 | {
24 | print("There are no valid compartment left. No compartiment codes will be added.")
25 | names(metab_input) <- paste("Metab__",names(metab_input),sep = "")
26 | return(metab_input)
27 | } else
28 | {
29 | print("Adding compartment codes.")
30 | metab_input_list <- lapply(compartment_codes, function(compartment_code, curr_metab_input){
31 | names(curr_metab_input) <- paste(names(curr_metab_input),compartment_code, sep = "_")
32 | names(curr_metab_input) <- paste("Metab__",names(curr_metab_input),sep = "")
33 | return(curr_metab_input)
34 | }, curr_metab_input = metab_input)
35 | metab_input <- unlist(metab_input_list)
36 | return(metab_input)
37 | }
38 | }
39 |
40 |
--------------------------------------------------------------------------------
/R/preprocess_COSMOS_metabolism_to_signaling.R:
--------------------------------------------------------------------------------
1 | #' Preprocess COSMOS Inputs For Metabolism to Signaling
2 | #'
3 | #' Runs checks on the input data and simplifies the prior knowledge network.
4 | #' Simplification includes the removal of (1) nodes that are not reachable from
5 | #' signaling nodes and (2) interactions between transcription factors and target
6 | #' genes if the target gene does not respond or the response is contradictory
7 | #' with the change in the transcription factor activity.
8 | #' Optionally, further TF activities are estimated via network optimization via
9 | #' CARNIVAL and the interactions between TF and genes are filtered again.
10 | #'
11 | #' @param meta_network prior knowledge network (PKN). A PKN released with COSMOS
12 | #' and derived from Omnipath, STITCHdb and Recon3D can be used. See details on
13 | #' the data \code{\link{meta_network}}.
14 | #' @param tf_regulon collection of transcription factor - target interactions.
15 | #' A default collection from dorothea can be obtained by the
16 | #' \code{\link{load_tf_regulon_dorothea}} function.
17 | #' @param signaling_data numerical vector, where names are signaling nodes
18 | #' in the PKN and values are from \{1, 0, -1\}. Continuous data will be
19 | #' discretized using the \code{\link{sign}} function.
20 | #' @param metabolic_data numerical vector, where names are metabolic nodes
21 | #' in the PKN and values are continuous values that represents log2 fold change
22 | #' or t-values from a differential analysis. These values are compared to
23 | #' the simulation results (simulated nodes can take value -1, 0 or 1)
24 | #' @param diff_expression_data (optional) numerical vector that represents the
25 | #' results of a differential gene expression analysis. Names are gene
26 | #' names using gene symbole and values are log fold change or
27 | #' t-values. We use the \dQuote{\code{diff_exp_threshold}} parameter to decide
28 | #' which genes changed significantly. Genes with NA values are considered none
29 | #' expressed and they will be removed from the TF-gene expression interactions.
30 | #' @param diff_exp_threshold threshold parameter (default 1) used to binarize
31 | #' the values of \dQuote{\code{diff_expression_data}}.
32 | #' @param maximum_network_depth integer > 0 (default: 8). Nodes that are further
33 | #' than \dQuote{\code{maximum_network_depth}} steps from the signaling nodes on
34 | #' the directed graph of the PKN are considered non-reachable and are removed.
35 | #' @param remove_unexpressed_nodes if TRUE (default) removes nodes from the PKN
36 | #' that are not expressed, see input \dQuote{\code{expressed_genes}}.
37 | #' @param expressed_genes character vector. Names of nodes that are expressed. By
38 | #' default we consider all the nodes that appear in \code{diff_expression_data} with
39 | #' a numeric value (i.e. nodes with NA are removed)
40 | #' @param filter_tf_gene_interaction_by_optimization (default:TRUE), if TRUE then runs
41 | #' a network optimization that estimates TF activity not included in the inputs
42 | #' and checks the consistency between the estimated activity and change in gene
43 | #' expression. Removes interactions where TF and gene expression are inconsistent
44 | #' @param CARNIVAL_options list that controls the options of CARNIVAL. See details
45 | #' in \code{\link{default_CARNIVAL_options}}.
46 | #' @importFrom utils data
47 | #' @export
48 | #' @return cosmos_data object with the following fields:
49 | #' \describe{
50 | #' \item{\code{meta_network}}{Filtered PKN}
51 | #' \item{\code{tf_regulon}}{TF - target regulatory network}
52 | #' \item{\code{signaling_data_bin}}{Binarised signaling data}
53 | #' \item{\code{metabolic_data}}{Metabolomics data}
54 | #' \item{\code{diff_expression_data_bin}}{Binarized gene expression data}
55 | #' \item{\code{optimized_network}}{Initial optimized network if
56 | #' \code{filter_tf_gene_interaction_by_optimization is TRUE}}
57 | #' }
58 | #' @seealso \code{\link{meta_network}} for meta PKN,
59 | #' \code{\link{load_tf_regulon_dorothea}} for tf regulon,
60 | #' \code{\link[CARNIVAL]{runCARNIVAL}}.
61 | #'
62 | #' @examples
63 | #' data(toy_network)
64 | #' data(toy_signaling_input)
65 | #' data(toy_metabolic_input)
66 | #' data(toy_RNA)
67 | #' test_back <- preprocess_COSMOS_metabolism_to_signaling(
68 | #' meta_network = toy_network,
69 | #' signaling_data = toy_signaling_input,
70 | #' metabolic_data = toy_metabolic_input,
71 | #' diff_expression_data = toy_RNA,
72 | #' maximum_network_depth = 15,
73 | #' remove_unexpressed_nodes = TRUE,
74 | #' CARNIVAL_options = default_CARNIVAL_options("lpSolve")
75 | #' )
76 | preprocess_COSMOS_metabolism_to_signaling <- function(meta_network = meta_network,
77 | tf_regulon = load_tf_regulon_dorothea(),
78 | signaling_data,
79 | metabolic_data,
80 | diff_expression_data = NULL,
81 | diff_exp_threshold = 1,
82 | maximum_network_depth = 8,
83 | expressed_genes = NULL,
84 | remove_unexpressed_nodes = TRUE,
85 | filter_tf_gene_interaction_by_optimization = TRUE,
86 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")){
87 |
88 | if(!is.null(diff_expression_data))
89 | {
90 | expressed_genes <- names(diff_expression_data)[!is.na(diff_expression_data)]
91 | } else
92 | {
93 | expressed_genes <- NULL
94 | remove_unexpressed_nodes <- FALSE
95 | }
96 | out_data <- preprocess_COSMOS_core(meta_network = meta_network,
97 | tf_regulon = tf_regulon,
98 | signaling_data=signaling_data,
99 | metabolic_data=metabolic_data,
100 | input_layer = "metabolic_data",
101 | output_layer = "signaling_data",
102 | diff_expression_data=diff_expression_data,
103 | diff_exp_threshold = diff_exp_threshold,
104 | maximum_network_depth = maximum_network_depth,
105 | expressed_genes = expressed_genes,
106 | remove_unexpressed_nodes = remove_unexpressed_nodes,
107 | filter_tf_gene_interaction_by_optimization = filter_tf_gene_interaction_by_optimization,
108 | CARNIVAL_options = CARNIVAL_options)
109 |
110 | out_data$history = c(out_data$history,"created by preprocess_COSMOS_metabolism_to_signaling\n")
111 |
112 | return(out_data)
113 | }
114 |
--------------------------------------------------------------------------------
/R/preprocess_COSMOS_signaling_to_metabolism.R:
--------------------------------------------------------------------------------
1 | #' Preprocess COSMOS Inputs For Signaling to Metabolism
2 | #'
3 | #' Runs checks on the input data and simplifies the prior knowledge network.
4 | #' Simplification includes the removal of (1) nodes that are not reachable from
5 | #' signaling nodes and (2) interactions between transcription factors and target
6 | #' genes if the target gene does not respond or the response is contradictory
7 | #' with the change in the transcription factor activity.
8 | #' Optionally, further TF activities are estimated via network optimization via
9 | #' CARNIVAL and the interactions between TF and genes are filtered again.
10 | #'
11 | #' @param meta_network prior knowledge network (PKN). A PKN released with COSMOS
12 | #' and derived from Omnipath, STITCHdb and Recon3D can be used. See details on
13 | #' the data \code{\link{meta_network}}.
14 | #' @param tf_regulon collection of transcription factor - target interactions.
15 | #' A default collection from dorothea can be obtained by the
16 | #' \code{\link{load_tf_regulon_dorothea}} function.
17 | #' @param signaling_data numerical vector, where names are signaling nodes
18 | #' in the PKN and values are from \{1, 0, -1\}. Continuous data will be
19 | #' discretized using the \code{\link{sign}} function.
20 | #' @param metabolic_data numerical vector, where names are metabolic nodes
21 | #' in the PKN and values are continuous values that represents log2 fold change
22 | #' or t-values from a differential analysis. These values are compared to
23 | #' the simulation results (simulated nodes can take value -1, 0 or 1)
24 | #' @param diff_expression_data (optional) numerical vector that represents the
25 | #' results of a differential gene expression analysis. Names are gene
26 | #' names using gene symbole and values are log fold change or
27 | #' t-values. We use the \dQuote{\code{diff_exp_threshold}} parameter to decide
28 | #' which genes changed significantly. Genes with NA values are considered none
29 | #' expressed and they will be removed from the TF-gene expression interactions.
30 | #' @param diff_exp_threshold threshold parameter (default 1) used to binarize
31 | #' the values of \dQuote{\code{diff_expression_data}}.
32 | #' @param maximum_network_depth integer > 0 (default: 8). Nodes that are further
33 | #' than \dQuote{\code{maximum_network_depth}} steps from the signaling nodes on
34 | #' the directed graph of the PKN are considered non-reachable and are removed.
35 | #' @param remove_unexpressed_nodes if TRUE (default) removes nodes from the PKN
36 | #' that are not expressed, see input \dQuote{\code{expressed_genes}}.
37 | #' @param expressed_genes character vector. Names of nodes that are expressed. By
38 | #' default we consider all the nodes that appear in \code{diff_expression_data} with
39 | #' a numeric value (i.e. nodes with NA are removed)
40 | #' @param filter_tf_gene_interaction_by_optimization (default:TRUE), if TRUE then runs
41 | #' a network optimization that estimates TF activity not included in the inputs
42 | #' and checks the consistency between the estimated activity and change in gene
43 | #' expression. Removes interactions where TF and gene expression are inconsistent
44 | #' @param CARNIVAL_options list that controls the options of CARNIVAL. See details
45 | #' in \code{\link{default_CARNIVAL_options}}.
46 | #' @importFrom utils data
47 | #' @export
48 | #' @return cosmos_data object with the following fields:
49 | #' \describe{
50 | #' \item{\code{meta_network}}{Filtered PKN}
51 | #' \item{\code{tf_regulon}}{TF - target regulatory network}
52 | #' \item{\code{signaling_data_bin}}{Binarised signaling data}
53 | #' \item{\code{metabolic_data}}{Metabolomics data}
54 | #' \item{\code{diff_expression_data_bin}}{Binarized gene expression data}
55 | #' \item{\code{optimized_network}}{Initial optimized network if
56 | #' \code{filter_tf_gene_interaction_by_optimization is TRUE}}
57 | #' }
58 | #' @seealso \code{\link{meta_network}} for meta PKN,
59 | #' \code{\link{load_tf_regulon_dorothea}} for tf regulon,
60 | #' \code{\link[CARNIVAL]{runCARNIVAL}}.
61 | #'
62 | #' @examples
63 | #' data(toy_network)
64 | #' data(toy_signaling_input)
65 | #' data(toy_metabolic_input)
66 | #' data(toy_RNA)
67 | #' test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
68 | #' signaling_data = toy_signaling_input,
69 | #' metabolic_data = toy_metabolic_input,
70 | #' diff_expression_data = toy_RNA,
71 | #' maximum_network_depth = 15,
72 | #' remove_unexpressed_nodes = TRUE,
73 | #' CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
74 |
75 | preprocess_COSMOS_signaling_to_metabolism <- function(meta_network = meta_network,
76 | tf_regulon = load_tf_regulon_dorothea(),
77 | signaling_data,
78 | metabolic_data,
79 | diff_expression_data = NULL,
80 | diff_exp_threshold = 1,
81 | maximum_network_depth = 8,
82 | expressed_genes = NULL,
83 | remove_unexpressed_nodes = TRUE,
84 | filter_tf_gene_interaction_by_optimization = TRUE,
85 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")){
86 |
87 | if(!is.null(diff_expression_data))
88 | {
89 | expressed_genes <- names(diff_expression_data)[!is.na(diff_expression_data)]
90 | } else
91 | {
92 | expressed_genes <- NULL
93 | remove_unexpressed_nodes <- FALSE
94 | }
95 |
96 | out_data <- preprocess_COSMOS_core(meta_network = meta_network,
97 | tf_regulon = tf_regulon,
98 | signaling_data=signaling_data,
99 | metabolic_data=metabolic_data,
100 | input_layer = "signaling_data",
101 | output_layer = "metabolic_data",
102 | diff_expression_data=diff_expression_data,
103 | diff_exp_threshold = diff_exp_threshold,
104 | maximum_network_depth = maximum_network_depth,
105 | expressed_genes = expressed_genes,
106 | remove_unexpressed_nodes = remove_unexpressed_nodes,
107 | filter_tf_gene_interaction_by_optimization = filter_tf_gene_interaction_by_optimization,
108 | CARNIVAL_options = CARNIVAL_options)
109 |
110 | out_data$history = c(out_data$history,"created by preprocess_COSMOS_signaling_to_metaboism\n")
111 |
112 | return(out_data)
113 | }
114 |
--------------------------------------------------------------------------------
/R/process_CARNIVAL_results.R:
--------------------------------------------------------------------------------
1 |
2 | #' process_CARNIVAL_results
3 | #'
4 | #' formats the raw CARNIVAL results to a more appealing list of networks.
5 | #' @param CARNIVAL_results list of matrices received from
6 | #' CRANIVAL::run_CARNIVAL()
7 | #' @return list with the following elements:
8 | #' \describe{
9 | #' \item{\code{aggregated_network}}{The averaged networks found by
10 | #' optimization in a format of a Simple Interaction network, i.e. each row
11 | #' codes an edge}
12 | #' \item{\code{N_networks}}{Number of solutions found by the
13 | #' optimization}
14 | #' \item{\code{aggregated_network_node_attributes}}{Estimated node properties}
15 | #' \item{\code{individual_networks}}{List of optimial networks found}
16 | #' \item{\code{individual_networks_node_attributes}}{Node activity in each
17 | #' network}
18 | #' }
19 | #' @importFrom rlang .data
20 | #' @importFrom dplyr %>% mutate rename as_tibble
21 | #' @importFrom purrr map
22 | #' @noRd
23 | process_CARNIVAL_results <- function(CARNIVAL_results){
24 |
25 | network_output = list()
26 |
27 | network_output$weightedSIF <- dplyr::as_tibble(CARNIVAL_results$weightedSIF) %>%
28 | dplyr::rename(Node1 = "Node1",
29 | Node2 = "Node2",
30 | Sign = "Sign",
31 | Weight = "Weight") %>%
32 | dplyr::mutate(Sign = as.numeric(.data$Sign),
33 | Weight = as.numeric(.data$Weight)/100)
34 |
35 | network_output$N_networks = length(CARNIVAL_results$sifAll)
36 |
37 | network_output$nodesAttributes =
38 | dplyr::as_tibble(CARNIVAL_results$nodesAttributes) %>%
39 | dplyr::mutate(ZeroAct = as.numeric(.data$ZeroAct)/100,
40 | UpAct = as.numeric(.data$UpAct)/100,
41 | DownAct = as.numeric(.data$DownAct)/100,
42 | AvgAct = as.numeric(.data$AvgAct)/100,
43 | )
44 |
45 | network_output$individual_networks =
46 | purrr::map(CARNIVAL_results$sifAll,function(Net){
47 | dplyr::as_tibble(Net) %>%
48 | dplyr::rename(source = "Node1",
49 | target = "Node2",
50 | interaction = "Sign") %>%
51 | dplyr::mutate(interaction = as.numeric(.data$interaction))
52 |
53 | } )
54 |
55 |
56 | network_output$individual_networks_node_attributes =
57 | purrr::map(CARNIVAL_results$attributesAll,function(Net){
58 | dplyr::as_tibble(Net) %>%
59 | dplyr::rename(node = "Nodes",
60 | activity = "Activity") %>%
61 | dplyr::mutate(activity = as.numeric(.data$activity))
62 |
63 | } )
64 |
65 | return(network_output)
66 |
67 | }
--------------------------------------------------------------------------------
/R/runCARNIVAL_wrapper.R:
--------------------------------------------------------------------------------
1 | #' Run CARNIVAL Wrapper
2 | #'
3 | #' Checks and formats the COSMOS data to CARNIVAL inputs and runs CARNIVAL.
4 | #'
5 | #' @param network Prior knowledge network (PKN). \dQuote{\code{data.frame}}
6 | #' object with source, sign and target columns. By default COSMOS uses a PKN
7 | #' derived from Omnipath, STITCHdb and Recon3D. See details on the data
8 | #' \code{\link{meta_network}}.
9 | #'
10 | #' @param input_data Numerical vector, where names are input nodes in the PKN
11 | #' and values are from \{1, 0, -1\}.
12 | #'
13 | #' @param measured_data Numerical vector, where names are measured nodes in the
14 | #' PKN and values are continuous values. These values are compared to with
15 | #' the simulation.
16 | #'
17 | #' @param options An object of type \dQuote{\code{list}} defining the run
18 | #' parameters for CARNIVAL. Use the \code{\link{default_CARNIVAL_options}}
19 | #' function to create a list with default parameter settings. If cplex or cbc
20 | #' are chosen as the solver, the parameter solverPath needs to be supplied
21 | #' (not automatically added by \code{default_CARNIVAL_options()}).
22 | #' @noRd
23 | runCARNIVAL_wrapper <- function(network,
24 | input_data,
25 | measured_data,
26 | options
27 | ){
28 |
29 | check_CARNIVAL_options(options)
30 | check_inputs_for_CARNIVAL(meta_network = network,
31 | input_data = input_data,
32 | measured_data = measured_data)
33 |
34 |
35 | CARNIVAL_Result <- CARNIVAL::runVanillaCarnival(perturbations = input_data,
36 | measurements = measured_data,
37 | priorKnowledgeNetwork = network,
38 | carnivalOptions = options)
39 |
40 |
41 | if(!validate_CARNIVAL_results(CARNIVAL_Result)) warning("We failed to validate CARNIVAL results.")
42 |
43 | return(CARNIVAL_Result)
44 | }
--------------------------------------------------------------------------------
/R/run_COSMOS_metabolism_to_signaling.R:
--------------------------------------------------------------------------------
1 | #' run COSMOS metabolism to signaling
2 | #'
3 | #' Runs COSMOS from metabolism to signaling. This function uses CARNIVAL to find
4 | #' a subset of the prior knowledge network based on optimization that (1)
5 | #' includes the most measured and input nodes and (2) which is in agreement with
6 | #' the data. Use \code{\link{preprocess_COSMOS_metabolism_to_signaling}} to
7 | #' prepare the the inputs, measurements and the prior knowledge network.
8 | #'
9 | #' @param data \code{\link{cosmos_data}} object. Use the
10 | #' \code{\link{preprocess_COSMOS_metabolism_to_signaling}} function to create
11 | #' an instance.
12 | #' @param CARNIVAL_options List that controls the options of CARNIVAL. See details
13 | #' in \code{\link{default_CARNIVAL_options}}.
14 | #' @export
15 | #' @return List with the following elements:
16 | #' \describe{
17 | #' \item{\code{weightedSIF}}{The averaged networks found by
18 | #' optimization in a format of a Simple Interaction network, i.e. each row
19 | #' codes an edge}
20 | #' \item{\code{N_networks}}{Number of solutions found by the
21 | #' optimization}
22 | #' \item{\code{nodesAttributes}}{Estimated node properties}
23 | #' \item{\code{individual_networks}}{List of optimial networks found}
24 | #' \item{\code{individual_networks_node_attributes}}{Node activity in each
25 | #' network}
26 | #' }
27 | #' @seealso \code{\link{preprocess_COSMOS_metabolism_to_signaling}},
28 | #' \code{\link[CARNIVAL]{runCARNIVAL}}, \code{\link{cosmos_data}}
29 | #' @examples
30 | #' data(toy_network)
31 | #' data(toy_signaling_input)
32 | #' data(toy_metabolic_input)
33 | #' data(toy_RNA)
34 | #' test_back <- preprocess_COSMOS_metabolism_to_signaling(meta_network = toy_network,
35 | #' signaling_data = toy_signaling_input,
36 | #' metabolic_data = toy_metabolic_input,
37 | #' diff_expression_data = toy_RNA,
38 | #' maximum_network_depth = 15,
39 | #' remove_unexpressed_nodes = TRUE,
40 | #' CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
41 | #'
42 | #' test_result_back <- run_COSMOS_metabolism_to_signaling(data = test_back,
43 | #' CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
44 | run_COSMOS_metabolism_to_signaling <- function(data,
45 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")){
46 |
47 | ## Checking COSMOS input format
48 | validate_cosmos_data_metabolism_to_signaling(data)
49 |
50 | check_CARNIVAL_options(CARNIVAL_options)
51 |
52 | disc_metabolic_data <- discretize_input(data$metabolic_data)
53 |
54 | CARNIVAL_results = runCARNIVAL_wrapper(network = data$meta_network,
55 | input_data = disc_metabolic_data,
56 | measured_data = data$signaling_data,
57 | options = CARNIVAL_options)
58 |
59 | network_results <- process_CARNIVAL_results(CARNIVAL_results)
60 |
61 | }
62 |
63 |
64 | validate_cosmos_data_metabolism_to_signaling <- function(data){
65 |
66 | validate_cosmos_data(data)
67 |
68 | if(!all(c("signaling_data_bin") %in% names(data)))
69 | stop("missing inputs detected. Input data should be obtained by running preprocess_cosmos_metabolism_to_signaling.")
70 |
71 | }
72 |
--------------------------------------------------------------------------------
/R/run_COSMOS_signaling_to_metabolism.R:
--------------------------------------------------------------------------------
1 | #' run COSMOS signaling to metabolism
2 | #'
3 | #' Runs COSMOS from signaling to metabolism. This function uses CARNIVAL to find
4 | #' a subset of the prior knowledge network based on optimisation that (1)
5 | #' includes the most measured and input nodes and (2) which is in agreement with
6 | #' the data. Use \code{\link{preprocess_COSMOS_signaling_to_metabolism}} to
7 | #' prepare inputs, measurements and prior knowledge network.
8 | #'
9 | #' @param data \code{\link{cosmos_data}} object. Use the
10 | #' \code{\link{preprocess_COSMOS_signaling_to_metabolism}} function to create
11 | #' an instance.
12 | #' @param CARNIVAL_options List that controls the options of CARNIVAL. See
13 | #' details in \code{\link{default_CARNIVAL_options}}.
14 | #' @export
15 | #' @return List with the following elements:
16 | #' \describe{
17 | #' \item{\code{weightedSIF}}{The averaged networks found by
18 | #' optimization in a format of a Simple Interaction network, i.e. each row
19 | #' codes an edge}
20 | #' \item{\code{N_networks}}{Number of solutions found by the
21 | #' optimization}
22 | #' \item{\code{nodesAttributes}}{Estimated node properties}
23 | #' \item{\code{individual_networks}}{List of optimial networks found}
24 | #' \item{\code{individual_networks_node_attributes}}{Node activity in each
25 | #' network}
26 | #' }
27 | #' @seealso \code{\link{preprocess_COSMOS_metabolism_to_signaling}},
28 | #' \code{\link[CARNIVAL]{runCARNIVAL}}, \code{\link{cosmos_data}}
29 | #' @examples
30 | #' data(toy_network)
31 | #' data(toy_signaling_input)
32 | #' data(toy_metabolic_input)
33 | #' data(toy_RNA)
34 | #' test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
35 | #' signaling_data = toy_signaling_input,
36 | #' metabolic_data = toy_metabolic_input,
37 | #' diff_expression_data = toy_RNA,
38 | #' maximum_network_depth = 15,
39 | #' remove_unexpressed_nodes = TRUE,
40 | #' CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
41 | #'
42 | #' test_result_for <- run_COSMOS_signaling_to_metabolism(data = test_for,
43 | #' CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
44 | run_COSMOS_signaling_to_metabolism <- function(data,
45 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")){
46 |
47 | ## Checking COSMOS input format
48 | validate_cosmos_data_signaling_to_metabolism(data)
49 |
50 | check_CARNIVAL_options(CARNIVAL_options)
51 |
52 | disc_signaling_data <- discretize_input(data$signaling_data)
53 |
54 | CARNIVAL_results = runCARNIVAL_wrapper(network = data$meta_network,
55 | input_data = disc_signaling_data,
56 | measured_data = data$metabolic_data,
57 | options = CARNIVAL_options)
58 |
59 | network_results <- process_CARNIVAL_results(CARNIVAL_results)
60 |
61 | }
62 |
63 |
64 | validate_cosmos_data_signaling_to_metabolism <- function(data){
65 |
66 | validate_cosmos_data(data)
67 |
68 | if(!all(c("signaling_data_bin") %in% names(data)))
69 | stop("missing inputs detected. Input data should be obtained by running preprocess_cosmos_signaling_to_metabolism.")
70 |
71 | }
72 |
73 |
--------------------------------------------------------------------------------
/R/support_ORA.R:
--------------------------------------------------------------------------------
1 | # Copyright (C) 2021 Caroline Lohoff, Aurelien Dugourd
2 |
3 | # based on "support_enrichment.R" from
4 | # https://github.com/saezlab/transcriptutorial/blob/master/scripts/06_analysis_CARNIVAL_results.md
5 | # Copyright (C) 2020 Aurelien Dugourd, Alberto Valdeolivas, Rosa Hernansaiz-Ballesteros
6 | # Contact : aurelien.dugourd@bioquant.uni-heidelberg.de
7 |
8 | # This program is free software: you can redistribute it and/or modify
9 | # it under the terms of the GNU General Public License as published by
10 | # the Free Software Foundation, either version 3 of the License, or
11 | # (at your option) any later version.
12 |
13 | # This program is distributed in the hope that it will be useful,
14 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
15 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 | # GNU General Public License for more details.
17 |
18 | # You should have received a copy of the GNU General Public License
19 | # along with this program. If not, see .
20 |
21 |
22 | #' Convert gmt file to data frame
23 | #'
24 | #' This function is designed to convert a gmt file (gene set file from MSigDB)
25 | #' into a two column data frame where the first column corresponds to omic
26 | #' features (genes) and the second column to associated terms (pathway the gene
27 | #' belongs to). One gene can belong to several pathways.
28 | #'
29 | #' @param gmtfile a full path name of the gmt file to be converted
30 | #' @return a two column data frame where the first column corresponds to omic
31 | #' features and the second column to associated terms (pathways).
32 | #' @importFrom dplyr bind_rows
33 | #' @importFrom progress progress_bar
34 |
35 | gmt_to_dataframe <- function(gmtfile)
36 | {
37 | genesets = GSEABase::getGmt(con = gmtfile)
38 | genesets = unlist(genesets)
39 |
40 | pb <- progress::progress_bar$new(total=length(genesets),
41 | format = "[:bar] :percent eta: :eta")
42 |
43 | gene_to_term <- lapply(genesets,function(geneset){
44 | pb$tick()
45 | temp <- GSEABase::geneIds(geneset)
46 | temp2 <- GSEABase::setName(geneset)
47 | temp3 <- as.data.frame(cbind(temp,rep(temp2,length(temp))))
48 |
49 | }) %>%
50 | dplyr::bind_rows()
51 |
52 | names(gene_to_term) <- c("gene","term")
53 | return(gene_to_term[stats::complete.cases(gene_to_term),])
54 | }
55 |
56 |
57 | #' Extract COSMOS nodes for ORA analysis
58 | #'
59 | #' Function to extract the nodes that appear in the COSMOS output network and
60 | #' the background genes (all genes present in the prior knowledge network)
61 | #'
62 | #' @param sif df; COSMOS network solution in sif format like the first list
63 | #' element returned by the format_cosmos_res function
64 | #' @param att df; attributes of the nodes of the COMSOS network solution like
65 | #' the second list element returned by the format_cosmos_res function
66 | #' @return List with 2 objects: the success and the background genes
67 | #' @export
68 | #' @examples
69 | #' CARNIVAL_options <- cosmosR::default_CARNIVAL_options("lpSolve")
70 | #' data(toy_network)
71 | #' data(toy_signaling_input)
72 | #' data(toy_metabolic_input)
73 | #' data(toy_RNA)
74 | #' test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
75 | #' signaling_data = toy_signaling_input,
76 | #' metabolic_data = toy_metabolic_input,
77 | #' diff_expression_data = toy_RNA,
78 | #' maximum_network_depth = 15,
79 | #' remove_unexpressed_nodes = TRUE,
80 | #' CARNIVAL_options = CARNIVAL_options
81 | #' )
82 | #' test_result_for <- run_COSMOS_signaling_to_metabolism(data = test_for,
83 | #' CARNIVAL_options = CARNIVAL_options)
84 | #' test_result_for <- format_COSMOS_res(test_result_for)
85 | #' extreacted_nodes <- extract_nodes_for_ORA(
86 | #' sif = test_result_for[[1]],
87 | #' att = test_result_for[[2]]
88 | #' )
89 |
90 | extract_nodes_for_ORA <- function(sif, att){
91 |
92 | # Extract all nodes from COSMOS's solution network
93 | CosmosNetwork <-
94 | as.data.frame(sif, stringsAsFactors = FALSE)
95 | colnames(CosmosNetwork) <- c("source", "sign", "target", "weight")
96 |
97 | # Remove all metabolites
98 | CosmosNetwork <- CosmosNetwork[!grepl("Metab", CosmosNetwork$source), ]
99 | CosmosNetwork <- CosmosNetwork[!grepl("Metab", CosmosNetwork$target), ]
100 |
101 | ## We define the set of unique nodes (genes) in the solution network
102 | sucesses <- unique(c(gsub("_.*","",CosmosNetwork$source),
103 | gsub("_.*","",CosmosNetwork$target)))
104 |
105 | # Extract all unique genes from PKN as background nodes
106 | Cosmos_attributes <- as.data.frame(att,
107 | stringsAsFactors = FALSE)
108 | Cosmos_attributes <- Cosmos_attributes[!grepl("Metab", Cosmos_attributes$Nodes), ]
109 | bg <- unique(gsub("_.*","",Cosmos_attributes$Nodes))
110 |
111 | return(list(sucesses = sucesses, bg = bg))
112 | }
113 |
--------------------------------------------------------------------------------
/R/support_decoupleR.R:
--------------------------------------------------------------------------------
1 | #' Format Ligand-Receptor Resource
2 | #'
3 | #' This function formats a ligand-receptor resource by creating a gene set
4 | #' with source-target pairs, converting it to a long format, and adding
5 | #' default values for 'mor' and 'likelihood'.
6 | #'
7 | #' @param ligrec_ressource A data frame representing the ligand-receptor resource with columns for source and target gene symbols.
8 | #'
9 | #' @return A data frame containing the formatted ligand-receptor gene set with columns:
10 | #' \item{gene}{The gene symbol from the ligand-receptor pairs.}
11 | #' \item{set}{The set identifier combining source and target gene symbols.}
12 | #' \item{mor}{Default value set to 1 for all entries.}
13 | #' \item{likelihood}{Default value set to 1 for all entries.}
14 | #'
15 | #' @examples
16 | #' # Create a sample ligand-receptor resource
17 | #' ligrec_ressource <- data.frame(source_genesymbol = c("L1", "L2"),
18 | #' target_genesymbol = c("R1", "R2"))
19 | #'
20 | #' # Format the ligand-receptor resource
21 | #' formatted_geneset <- format_LR_ressource(ligrec_ressource)
22 | #'
23 | #' @export
24 | format_LR_ressource <- function(ligrec_ressource)
25 | {
26 | ligrec_geneset <- ligrec_ressource[,c("source_genesymbol","target_genesymbol")]
27 | ligrec_geneset$set <- paste(ligrec_geneset$source_genesymbol, ligrec_geneset$target_genesymbol, sep = "___")
28 | ligrec_geneset <- reshape2::melt(ligrec_geneset, id.vars = "set")[,c(3,1)]
29 | names(ligrec_geneset)[1] <- "gene"
30 | ligrec_geneset$mor <- 1
31 | ligrec_geneset$likelihood <- 1
32 | ligrec_geneset <- distinct(ligrec_geneset)
33 |
34 | return(ligrec_geneset)
35 | }
36 |
37 | #' Convert ULM Results to Wide Format
38 | #'
39 | #' This function converts the results from a ULM analysis to a wide format data frame.
40 | #' The input is a data frame with columns for source, condition, and score. The output
41 | #' is a data frame where each row represents a source and each column represents a condition,
42 | #' with the corresponding scores as values.
43 | #'
44 | #' @param ulm_result A data frame representing the ULM results with columns: source, condition, and score.
45 | #'
46 | #' @return A data frame in wide format where each row is a source and each column is a condition.
47 | #'
48 | #' @examples
49 | #' # Create a sample ULM result
50 | #' ulm_result <- data.frame(source = c("A", "A", "B", "B"),
51 | #' condition = c("cond1", "cond2", "cond1", "cond2"),
52 | #' score = c(0.5, 0.8, 0.3, 0.7))
53 | #'
54 | #' # Convert to wide format
55 | #' wide_ulm_result <- wide_ulm_res(ulm_result)
56 | #'
57 | #' @export
58 | wide_ulm_res <- function(ulm_result)
59 | {
60 | ulm_result_df <- reshape2::dcast(ulm_result, formula = source~condition, value.var = "score")
61 | row.names(ulm_result_df) <- ulm_result_df$source
62 | ulm_result_df <- ulm_result_df[,-1]
63 |
64 | return(ulm_result_df)
65 | }
66 |
67 | #' Translate Column Using HMDB Mapper
68 | #'
69 | #' This function translates the values in a column using a provided Human Metabolome Database (HMDB) mapper vector.
70 | #' It modifies the input values by replacing certain prefixes and suffixes according to specific rules.
71 | #'
72 | #' @param my_column A vector of values to be translated.
73 | #' @param HMDB_mapper_vec A named vector where the names are the original identifiers and the values are the corresponding HMDB identifiers.
74 | #'
75 | #' @return A vector with the translated values.
76 | #'
77 | #' @examples
78 | #' # Create a sample column and HMDB mapper vector
79 | #' my_column <- c("Metab__1234_a", "Gene5678_b", "Metab__91011_c")
80 | #' HMDB_mapper_vec <- c("1234" = "HMDB00001", "5678" = "HMDB00002", "91011" = "HMDB00003")
81 | #'
82 | #' # Translate the column
83 | #' translated_column <- translate_column_HMDB(my_column, HMDB_mapper_vec)
84 | #'
85 | #' @export
86 | translate_column_HMDB <- function(my_column, HMDB_mapper_vec)
87 | {
88 | return(sapply(my_column, function(x, HMDB_mapper_vec) {
89 | x <- gsub("Metab__", "", x)
90 | x <- gsub("^Gene", "Enzyme", x)
91 | suffixe <- stringr::str_extract(x, "_[a-z]$")
92 | x <- gsub("_[a-z]$", "", x)
93 | if (x %in% names(HMDB_mapper_vec)) {
94 | x <- HMDB_mapper_vec[x]
95 | x <- paste("Metab__", x, sep = "")
96 | }
97 | if (!is.na(suffixe)) {
98 | x <- paste(x, suffixe, sep = "")
99 | }
100 | return(x)
101 | }, HMDB_mapper_vec = HMDB_mapper_vec))
102 | }
--------------------------------------------------------------------------------
/R/support_pheatmap.R:
--------------------------------------------------------------------------------
1 | #' Create Heatmap Color Palette
2 | #'
3 | #' This function generates a color palette suitable for heatmaps based on the values in a matrix. It uses the `createLinearColors` function to generate separate color gradients for positive and negative values.
4 | #'
5 | #' @param my_matrix A numeric matrix for which the heatmap color palette is to be generated.
6 | #'
7 | #' @return A character vector of colors representing the heatmap color palette based on the input matrix values.
8 | #'
9 | #' @examples
10 | #' # Create a sample matrix
11 | #' my_matrix <- matrix(c(-3, -1, 0, 1, 3), nrow = 1)
12 | #'
13 | #' # Generate heatmap color palette
14 | #' heatmap_palette <- make_heatmap_color_palette(my_matrix)
15 | #'
16 | #' @export
17 | make_heatmap_color_palette <- function(my_matrix)
18 | {
19 | t <- as.vector(t(my_matrix))
20 | palette1 <- createLinearColors(t[t < 0],withZero = F , maximum = abs(min(t,na.rm = T)) * 10)
21 | palette2 <- createLinearColors(t[t > 0],withZero = F , maximum = abs(max(t,na.rm = T)) * 10)
22 | palette <- c(palette1, palette2)
23 | }
24 |
25 | #' Create Linear Colors Based on Numeric Input
26 | #'
27 | #' This function generates a gradient of colors based on the provided numeric values. The colors can be adjusted to include zero and are configurable with a specified maximum and custom color palette.
28 | #'
29 | #' @param numbers A numeric vector for which the color gradient is to be generated.
30 | #' @param withZero A logical value indicating whether zero should be included in the color gradient. Default is TRUE.
31 | #' @param maximum An integer specifying the maximum number of colors to be generated in the gradient. Default is 100.
32 | #' @param my_colors A character vector of length three specifying the colors to be used in the gradient. Default is c("royalblue3", "white", "red").
33 | #'
34 | #' @return A character vector of colors representing the gradient based on the input numeric values.
35 | #'
36 | #' @examples
37 | #' # Generate colors for a set of numbers including zero
38 | #' numbers <- c(-50, -20, 0, 20, 50)
39 | #' colors <- createLinearColors(numbers, withZero = TRUE, maximum = 100)
40 | #'
41 | #' @export
42 | createLinearColors <- function(numbers, withZero = T, maximum = 100, my_colors = c("royalblue3","white","red"))
43 | {
44 | if (min(numbers, na.rm = T) > 0)
45 | {
46 | if(withZero)
47 | {
48 | numbers <- c(0,numbers)
49 | }
50 | myPalette <- colorRampPalette(my_colors[c(2,3)])
51 | myColors <- myPalette(maximum)
52 | }
53 | else
54 | {
55 | if (max(numbers, na.rm = T) < 0)
56 | {
57 | if(withZero)
58 | {
59 | numbers <- c(0,numbers)
60 | }
61 | myPalette <- colorRampPalette(my_colors[c(1,2)])
62 | myColors <- myPalette(maximum)
63 | }
64 | else
65 | {
66 | myPalette_pos <- colorRampPalette(c("white","red"))
67 | myPalette_neg <- colorRampPalette(c("royalblue3","white"))
68 | npos <- length(numbers[numbers >= 0]) + 1
69 | nneg <- length(numbers[numbers <= 0]) + 1
70 |
71 | myColors_pos <- myPalette_pos(npos)
72 | myColors_neg <- myPalette_neg(nneg)
73 |
74 | #print(myColors_neg)
75 | #print(myColors_pos)
76 |
77 | myColors <- c(myColors_neg[-(nneg)], myColors_pos[-1])
78 | }
79 | }
80 | return(myColors)
81 | }
82 |
--------------------------------------------------------------------------------
/R/sysdata.rda:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/R/sysdata.rda
--------------------------------------------------------------------------------
/R/toy_RNA.R:
--------------------------------------------------------------------------------
1 | #' Toy Input Transcription Data Set
2 | #'
3 | #' This exemplary transcription data are the specific deregulated gene expression of the 786-O cell line from the NCI60 dataset.
4 | #'
5 | #' @docType data
6 | #'
7 | #' @usage data(toy_RNA)
8 | #'
9 | #' @format An object of class \dQuote{\code{numeric}} containing the t-values of
10 | #' 9300 genes, which are named with gene symboles matching the toy network.
11 | #'
12 | #' @source
13 | #' \url{https://github.com/saezlab/COSMOS_MSB/blob/main/data/RNA_ttop_tumorvshealthy.csv}
14 | #'
15 | #' @references {
16 | #' Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
17 | #' Systems Biology}. \bold{17}, e9730.
18 | #' }
19 | #'
20 | #' @examples
21 | #' data(toy_RNA)
22 | #'
23 | #' @keywords datasets
24 | "toy_RNA"
--------------------------------------------------------------------------------
/R/toy_metabolic_input.R:
--------------------------------------------------------------------------------
1 | #' Toy Metabolic Input Data
2 | #'
3 | #' This metabolic data are a subset from the metabolic measurements of the 786-O cell line from the NCI60 dataset.
4 | #'
5 | #' @docType data
6 | #'
7 | #' @usage data(toy_metabolic_input)
8 | #'
9 | #' @format An object of class \dQuote{\code{numeric}} containing the t-values of
10 | #' 2 metabolites, which are named with metabolite HMDB Ids matching the
11 | #' toy network.
12 | #'
13 | #' @source Subset of:
14 | #' \url{https://github.com/saezlab/COSMOS_MSB/blob/main/data/metab_input_COSMOS.csv}
15 | #'
16 | #' @references {
17 | #' Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
18 | #' Systems Biology}. \bold{17}, e9730.
19 | #' }
20 | #'
21 | #' @examples
22 | #' data(toy_metabolic_input)
23 | #'
24 | #' @keywords datasets
25 | "toy_metabolic_input"
--------------------------------------------------------------------------------
/R/toy_network.R:
--------------------------------------------------------------------------------
1 | #' Toy Input Network
2 | #'
3 | #' This signaling network is the reduced COSMOS network solution obtained in the
4 | #' cosmos test on 786-O NCI60 data. Here, this network solution is reused as an
5 | #' exemplary input prior knowledge network (PKN).
6 | #'
7 | #' @docType data
8 | #'
9 | #' @usage data(toy_network)
10 | #'
11 | #' @format An object of class \dQuote{\code{data.frame}} with 19 rows
12 | #' (interactions) and three variables:
13 | #' \describe{
14 | #' \item{\code{source}}{Source node, either metabolite or protein}
15 | #' \item{\code{interaction}}{Type of interaction, 1 = Activation, -1 = Inhibition}
16 | #' \item{\code{target}}{Target node, either metabolite or protein}
17 | #' A detailed description of the identifier formatting can be found under
18 | #' \url{https://metapkn.omnipathdb.org/00__README.txt}.
19 | #' }
20 | #'
21 | #' @source The network data are available on github:
22 | #' \url{https://github.com/saezlab/COSMOS_MSB/tree/main/results/COSMOS_result/COSMOS_res_session.RData}.
23 | #' The toy_network is the combined network of the COSMOS network solutions
24 | #' CARNIVAL_Result2 and CARNIVAL_Result_rerun subsequently reduced to 19
25 | #' exemplary nodes.
26 | #'
27 | #' @references {
28 | #' Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
29 | #' Systems Biology}. \bold{17}, e9730.
30 | #' }
31 | #'
32 | #' @examples
33 | #' data(toy_network)
34 | #'
35 | #' @keywords datasets
36 | "toy_network"
--------------------------------------------------------------------------------
/R/toy_signaling_input.R:
--------------------------------------------------------------------------------
1 | #' Toy Signaling Input
2 | #'
3 | #' This signaling data are a subset of the footprint-based signaling activity
4 | #' estimates of transcription factors of the 786-O cell line from the NCI60 dataset.
5 | #'
6 | #' @docType data
7 | #'
8 | #' @usage data(toy_signaling_input)
9 | #'
10 | #' @format An object of class \dQuote{\code{data.frame}} containing the normalised
11 | #' enrichment scores (NES) of 2 signaling proteins, which are named with their
12 | #' respective gene Entrez ID matching the toy network.
13 | #'
14 | #' @source Subset of:
15 | #' \url{https://github.com/saezlab/COSMOS_MSB/blob/main/data/signaling_input_COSMOS.csv}
16 | #'
17 | #' @references {
18 | #' Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
19 | #' Systems Biology}. \bold{17}, e9730.
20 | #' }
21 | #'
22 | #' @examples
23 | #' data(toy_signaling_input)
24 | #'
25 | #' @keywords datasets
26 | "toy_signaling_input"
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # cosmosR 
2 |
3 |
4 | [](https://github.com/saezlab/cosmosr/actions)
5 |
6 |
7 | ## Overview
8 |
9 | COSMOS (Causal Oriented Search of Multi-Omic Space) is a method that integrates phosphoproteomics, transcriptomics, and metabolomics data sets. COSMOS leverages extensive prior knowledge of signaling pathways, metabolic networks, and gene regulation with computational methods to estimate activities of transcription factors and kinases as well as network-level causal reasoning. This pipeline can provide mechanistic explanations for experimental observations across multiple omic data sets.
10 |
11 |
12 | 
13 |
14 | COSMOS finds coherent subnetwork causally connecting as many deregulated TFs, kinases/phosphatases and metabolites as possible. The subnetwork is extracted from a novel integrated PKN (available [here](http://metapkn.omnipathdb.org/)) spanning signaling, transcriptional regulation and metabolism. Transcription factors activities are inferred from gene expression with [decoupleR](https://saezlab.github.io/decoupleR/). Kinase activities are inferred from phosphoproteomic with a kinase/substrate network of [Omnipath](http://omnipathdb.org/), a meta resource of protein-protein. The scripts to generate the current network can be found here: https://github.com/saezlab/meta_PKN_BIGG.
15 |
16 |
17 | You can also use COSMOS if you don't have metabolomic data, to connect TF activities (from transcriptomic) with kinase activities (from phosphoproteomic) for exmaple !
18 |
19 | 
20 |
21 |
22 | ## Installation
23 |
24 | R >= 4.1 is required
25 | ```r
26 | # install from bioconductor
27 | if (!requireNamespace("BiocManager", quietly = TRUE))
28 | install.packages("BiocManager")
29 |
30 | BiocManager::install("cosmosR")
31 |
32 | # We advise to instal from github to get the latest version of the tool.
33 | if (!requireNamespace("devtools", quietly = TRUE))
34 | install.packages("devtools")
35 |
36 | devtools::install_github("saezlab/cosmosR")
37 | ```
38 |
39 | If you don't have R 4.1, you can also clone the github repository on your machine, create a new R project with R studio from the cosmosR folder, change the R version to your own R version in the DESCRIPTION file and then install it with devtools:install()
40 |
41 | But 4.1 is advised in any case.
42 |
43 | ## tutorial to use MOFA and COSMOS
44 |
45 | [Here](https://github.com/saezlab/Factor_COSMOS/) you can find an extensive tutorial showing how to use MOFA and COSMOS with the NCI60 dataset. This is an extensive tutorial, if you wish to get a quicker plug and play introduction to COSMOS, see below.
46 |
47 | !!! THIS is were you can find the input data and the pre-processing scripts that corespond to the featured vignette !!!
48 |
49 | ## Tutorial (NCI60 playground)
50 |
51 | We made a repository that contains pre-processed inputs and an example script to use cosmos with the NCI60 RNA+metabolomic datasets.
52 | You can find the repository [here](https://github.com/saezlab/COSMOS_basic).
53 |
54 | !!! THIS is were you can find the input data and the pre-processing scripts that corespond to the featured vignette !!!
55 |
56 | ## Access
57 |
58 | The meta PKN used with the older biorXiv version of COSMOS (2021) is available [here](http://metapkn.omnipathdb.org/).
59 |
60 | An updated meta PKN is available with the package (using data(meta_network) in R)
61 |
62 | ## Citation
63 | If you use cosmosR for your research please cite [COSMOS+ preprint](https://www.biorxiv.org/content/10.1101/2024.07.15.603538v2)
64 | Dugourd A, Lafrenz P, Mañanes D, Fallegger R, Kroger AC, Turei D, Shtylla B, Saez-Rodriguez J; Modeling causal signal propagation in multi-omic factor space with COSMOS; BioRxiv. 2024 Jul 17
65 | DOI: 10.1101/2024.07.15.603538
66 |
67 | The first publication of COSMOS is MSB can be found [here](https://www.embopress.org/doi/full/10.15252/msb.20209730):
68 |
69 | ## License
70 |
71 | The code is distributed under the GNU General Public License v3.0. The meta PKN is distributed under the Attribution-NonCommercial 4.0 International (CC-BY-NC 4.0) License.
72 |
--------------------------------------------------------------------------------
/data/HMDB_mapper_vec.RData:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/data/HMDB_mapper_vec.RData
--------------------------------------------------------------------------------
/data/meta_network.RData:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/data/meta_network.RData
--------------------------------------------------------------------------------
/data/toy_RNA.RData:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/data/toy_RNA.RData
--------------------------------------------------------------------------------
/data/toy_metabolic_input.RData:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/data/toy_metabolic_input.RData
--------------------------------------------------------------------------------
/data/toy_network.RData:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/data/toy_network.RData
--------------------------------------------------------------------------------
/data/toy_signaling_input.RData:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/data/toy_signaling_input.RData
--------------------------------------------------------------------------------
/inst/CITATION:
--------------------------------------------------------------------------------
1 | citEntry(
2 | entry="article",
3 | author = c(person("Aurelien", "Dugourd"),
4 | person("Christoph", "Kuppe"),
5 | person("Marco", "Sciacovelli"),
6 | person("Enio", "Gjerga"),
7 | person("Attila", "Gabor"),
8 | person("Kristina B.", "Embdal"),
9 | person("Vitor", "Vieira"),
10 | person("Dorte B.", "Bekker-Jensen"),
11 | person("Jennifer", "Kranz"),
12 | person("Eric M. J.", "Bindels"),
13 | person("Ana S. H.", "Costa"),
14 | person("Abel", "Sousa"),
15 | person("Pedro", "Beltrao"),
16 | person("Jesper V", "Olsen"),
17 | person("Christian", "Frezza"),
18 | person("Rafael", "Kramann"),
19 | person("Julio", "Saez-Rodriguez")),
20 | title = "Causal integration of multi-omics data with prior knowledge to generate mechanistic hypotheses",
21 | journal = "Molecular Systems Biology",
22 | year = 2021,
23 |
24 | textVersion = "Aurelien Dugourd, Christoph Kuppe, Marco Sciacovelli, Enio Gjerga, Attila Gabor, Kristina B. Emdal, Vitor Vieira, Dorte B. Bekker-Jensen, Jennifer Kranz, Eric. M. J. Bindels, Ana S. H. Costa, Abel sousa, Pedro Beltrao, Jesper V. Olsen, Christian Frezza, Rafael Kramann, Julio Saez-Rodriguez. 'Causal integration of multi-omics data with prior knowledge to generate mechanistic hypotheses.' Molecular Systems Biology. 2021."
25 | )
26 |
--------------------------------------------------------------------------------
/man/HMDB_mapper_vec.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/HMDB_mapper_vec.R
3 | \docType{data}
4 | \name{HMDB_mapper_vec}
5 | \alias{HMDB_mapper_vec}
6 | \title{Toy Input Transcription Data Set}
7 | \format{
8 | An object of class \dQuote{\code{character}} containing the marching between HMDB metabolite IDs and there coresponding metabolite names.
9 | }
10 | \source{
11 | \url{https://bioconductor.org/packages/release/data/annotation/html/metaboliteIDmapping.html}
12 | }
13 | \usage{
14 | data(HMDB_mapper_vec)
15 | }
16 | \description{
17 | This exemplary transcription data are the specific deregulated gene expression of the 786-O cell line from the NCI60 dataset.
18 | }
19 | \examples{
20 | data(HMDB_mapper_vec)
21 |
22 | }
23 | \keyword{datasets}
24 |
--------------------------------------------------------------------------------
/man/compress_same_children.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/compress_same_children.R
3 | \name{compress_same_children}
4 | \alias{compress_same_children}
5 | \title{Compress Network by Merging Nodes with Identical Children}
6 | \usage{
7 | compress_same_children(df, sig_input, metab_input)
8 | }
9 | \arguments{
10 | \item{df}{A data frame representing the network with three columns: source, target, and sign of interaction.}
11 |
12 | \item{sig_input}{A list of input node signatures to be considered for the merging process.}
13 |
14 | \item{metab_input}{A list of input metabolic signatures to be considered for the merging process.}
15 | }
16 | \value{
17 | A list containing the following elements:
18 | \item{compressed_network}{A data frame representing the compressed network.}
19 | \item{node_signatures}{A list of signatures of nodes in the network after the merging process.}
20 | \item{duplicated_signatures}{A list of duplicated signatures in the network after the merging process.}
21 | }
22 | \description{
23 | This function compresses a network by merging nodes that have the same children.
24 | The input network is represented as a data frame with three columns: source, target, and sign of interaction.
25 | The function returns a list containing the compressed network, node signatures, and duplicated signatures.
26 | }
27 | \examples{
28 | # Create a sample network
29 | df <- data.frame(source = c("A", "A", "B", "B"),
30 | target = c("C", "D", "C", "D"),
31 | sign_of_interaction = c(1, 1, 1, 1))
32 |
33 | # Define input node and metabolic signatures
34 | sig_input <- list()
35 | metab_input <- list()
36 |
37 | # Compress the network
38 | result <- compress_same_children(df, sig_input, metab_input)
39 | compressed_network <- result$compressed_network
40 |
41 | }
42 |
--------------------------------------------------------------------------------
/man/cosmos_data.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/cosmos_data.R
3 | \name{cosmos_data}
4 | \alias{cosmos_data}
5 | \title{Create Cosmos Data}
6 | \usage{
7 | cosmos_data(
8 | meta_network,
9 | tf_regulon = NULL,
10 | signaling_data,
11 | metabolic_data,
12 | expression_data,
13 | verbose = TRUE
14 | )
15 | }
16 | \arguments{
17 | \item{meta_network}{Prior knowledge network (PKN). By default COSMOS use a
18 | PKN derived from Omnipath, STITCHdb and Recon3D. See details on the data
19 | \code{\link{meta_network}}.}
20 |
21 | \item{tf_regulon}{Collection of transcription factor - target interactions.
22 | A default collection from dorothea can be obtained by the
23 | \code{\link{load_tf_regulon_dorothea}} function.}
24 |
25 | \item{signaling_data}{Numerical vector, where names are signaling nodes
26 | in the PKN and values are from \{1, 0, -1\}. Continuous data will be
27 | discretized using the \code{\link{sign}} function.}
28 |
29 | \item{metabolic_data}{Numerical vector, where names are metabolic nodes
30 | in the PKN and values are continuous values that represents log2 fold change
31 | or t-values from a differential analysis. These values are compared to
32 | the simulation results (simulated nodes can take value -1, 0 or 1).}
33 |
34 | \item{expression_data}{Numerical vector that represents the results of a
35 | differential gene expression analysis. Names are gene names using EntrezID
36 | starting with an X and values are log fold change or t-values. Genes with
37 | NA values are considered none expressed and they will be removed from the
38 | TF-gene expression interactions.}
39 |
40 | \item{verbose}{(default: TRUE) Reports details about the
41 | \code{\link{cosmos_data}} object.}
42 | }
43 | \value{
44 | \code{cosmos data} class instance.
45 | }
46 | \description{
47 | An S3 class that combines the required data into a comprehensive list. Use
48 | the \code{\link{preprocess_COSMOS_signaling_to_metabolism}} or
49 | \code{\link{preprocess_COSMOS_metabolism_to_signaling}} to create an instance.
50 | }
51 |
--------------------------------------------------------------------------------
/man/createLinearColors.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_pheatmap.R
3 | \name{createLinearColors}
4 | \alias{createLinearColors}
5 | \title{Create Linear Colors Based on Numeric Input}
6 | \usage{
7 | createLinearColors(
8 | numbers,
9 | withZero = T,
10 | maximum = 100,
11 | my_colors = c("royalblue3", "white", "red")
12 | )
13 | }
14 | \arguments{
15 | \item{numbers}{A numeric vector for which the color gradient is to be generated.}
16 |
17 | \item{withZero}{A logical value indicating whether zero should be included in the color gradient. Default is TRUE.}
18 |
19 | \item{maximum}{An integer specifying the maximum number of colors to be generated in the gradient. Default is 100.}
20 |
21 | \item{my_colors}{A character vector of length three specifying the colors to be used in the gradient. Default is c("royalblue3", "white", "red").}
22 | }
23 | \value{
24 | A character vector of colors representing the gradient based on the input numeric values.
25 | }
26 | \description{
27 | This function generates a gradient of colors based on the provided numeric values. The colors can be adjusted to include zero and are configurable with a specified maximum and custom color palette.
28 | }
29 | \examples{
30 | # Generate colors for a set of numbers including zero
31 | numbers <- c(-50, -20, 0, 20, 50)
32 | colors <- createLinearColors(numbers, withZero = TRUE, maximum = 100)
33 |
34 | }
35 |
--------------------------------------------------------------------------------
/man/decompress_moon_result.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{decompress_moon_result}
4 | \alias{decompress_moon_result}
5 | \title{Decompress Moon Result}
6 | \usage{
7 | decompress_moon_result(moon_res, meta_network_compressed_list, meta_network)
8 | }
9 | \arguments{
10 | \item{moon_res}{A data frame containing the results of a moon analysis.}
11 |
12 | \item{meta_network_compressed_list}{A list containing compressed meta network details,
13 | including node signatures and duplicated parents.}
14 |
15 | \item{meta_network}{A data frame representing the original meta network.}
16 | }
17 | \value{
18 | A data frame which merges the moon analysis results with the meta network data,
19 | including additional details about node signatures and handling of duplicated parents.
20 | }
21 | \description{
22 | This function decompresses the results obtained from moon analysis by incorporating
23 | node signatures and handling duplicated parents. It merges these details with the
24 | provided meta network data and returns a comprehensive data frame.
25 | }
26 | \examples{
27 | # Example usage (requires appropriate data structures for moon_res,
28 | # meta_network_compressed_list, and meta_network)
29 | # decompressed_result <- decompress_moon_result(moon_res, meta_network_compressed_list, meta_network)
30 |
31 | }
32 |
--------------------------------------------------------------------------------
/man/decompress_solution_network.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/compress_same_children.R
3 | \name{decompress_solution_network}
4 | \alias{decompress_solution_network}
5 | \title{Decompress Solution Network}
6 | \usage{
7 | decompress_solution_network(
8 | formatted_res,
9 | meta_network,
10 | node_signatures,
11 | duplicated_parents
12 | )
13 | }
14 | \arguments{
15 | \item{formatted_res}{A list containing the solution network and attribute table.}
16 |
17 | \item{meta_network}{A data frame representing the meta network.}
18 |
19 | \item{node_signatures}{A list of node signatures.}
20 |
21 | \item{duplicated_parents}{A list of duplicated parents from the compression process.}
22 | }
23 | \value{
24 | A list containing the following elements:
25 | \item{SIF}{A data frame representing the decompressed solution network.}
26 | \item{ATT}{A data frame containing the attributes of the decompressed solution network.}
27 | }
28 | \description{
29 | This function decompresses a solution network by mapping node signatures back to their original identifiers.
30 | The input is a formatted solution network, a meta network, node signatures, and duplicated parents.
31 | The function returns a list containing the decompressed solution network and attribute table.
32 | }
33 | \examples{
34 | # Create a sample formatted_res
35 | formatted_res <- list(
36 | SIF = data.frame(source = c("parent_of_D1", "D"),
37 | target = c("D", "F"),
38 | interaction = c(1, 1),
39 | Weight = c(1, 1)),
40 | ATT = data.frame(Nodes = c("parent_of_D1", "D", "F"),
41 | NodeType = c("","",""),
42 | ZeroAct = c(0,0,0),
43 | UpAct = c(1,1,1),
44 | DownAct = c(0,0,0),
45 | AvgAct = c(1,1,1),
46 | measured = c(0,0,0),
47 | Activity = c(1,1,1))
48 | )
49 |
50 | # Create a sample meta_network
51 | meta_network <- data.frame(source = c("A", "B", "D"),
52 | target = c("D", "D", "F"),
53 | interaction_type = c(1, 1, 1))
54 |
55 | # Define node_signatures and duplicated_parents
56 | node_signatures <- list("A" = "parent_of_D1","B" = "parent_of_D1","D" = "parent_F1")
57 | duplicated_parents <- c("A" = "parent_of_D1","B" = "parent_of_D1")
58 |
59 | # Decompress the solution network
60 | result <- decompress_solution_network(formatted_res, meta_network, node_signatures, duplicated_parents)
61 | decompressed_network <- result[[1]]
62 | attribute_table <- result[[2]]
63 |
64 | }
65 |
--------------------------------------------------------------------------------
/man/decoupleRnival.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{decoupleRnival}
4 | \alias{decoupleRnival}
5 | \title{DecoupleRnival}
6 | \usage{
7 | decoupleRnival(
8 | upstream_input = NULL,
9 | downstream_input,
10 | meta_network,
11 | n_layers,
12 | n_perm = 1000,
13 | downstream_cutoff = 0,
14 | statistic = "norm_wmean"
15 | )
16 | }
17 | \arguments{
18 | \item{upstream_input}{A named vector with up_stream nodes and their corresponding activity.}
19 |
20 | \item{downstream_input}{A named vector with down_stream nodes and their corresponding activity.}
21 |
22 | \item{meta_network}{A network data frame containing signed directed prior knowledge of molecular interactions.}
23 |
24 | \item{n_layers}{The number of layers that will be propagated upstream.}
25 |
26 | \item{n_perm}{The number of permutations to use in decoupleR's algorithm.}
27 |
28 | \item{downstream_cutoff}{If downstream measurments should be included above a given threshold}
29 |
30 | \item{statistic}{the decoupleR stat to consider: "wmean", "norm_wmean", or "ulm"}
31 | }
32 | \value{
33 | A data frame containing the score of the nodes upstream of the
34 | downstream input based on the iterative propagation
35 | }
36 | \description{
37 | Iteratively propagate downstream input activity through a signed directed network
38 | using the weighted mean enrichment score from decoupleR package
39 | }
40 | \examples{
41 | # Example input data
42 | upstream_input <- c("A" = 1, "B" = -1, "C" = 0.5)
43 | downstream_input <- c("D" = 2, "E" = -1.5)
44 | meta_network <- data.frame(
45 | source = c("A", "A", "B", "C", "C", "D", "E"),
46 | target = c("B", "C", "D", "E", "D", "B", "A"),
47 | sign = c(1, -1, -1, 1, -1, -1, 1)
48 | )
49 |
50 | # Run the function with the example input data
51 | result <- decoupleRnival(upstream_input, downstream_input, meta_network, n_layers = 2, n_perm = 100)
52 |
53 | # View the results
54 | print(result)
55 | }
56 |
--------------------------------------------------------------------------------
/man/default_CARNIVAL_options.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/default_CARNIVAL_options.R
3 | \name{default_CARNIVAL_options}
4 | \alias{default_CARNIVAL_options}
5 | \title{Setting Default CARNIVAL Options}
6 | \usage{
7 | default_CARNIVAL_options(solver = NULL)
8 | }
9 | \arguments{
10 | \item{solver}{one of `cplex` (recommended, but require 3rd party tool), `cbc` (also require 3rd party tool) or `lpSolve` (only for small networks)}
11 | }
12 | \value{
13 | returns a list with all possible options implemented in CARNIVAL.
14 | see the documentation on \code{\link[CARNIVAL]{runCARNIVAL}}.
15 | }
16 | \description{
17 | Returns the default CARNIVAL options as a list. You can modify the elements
18 | of the list and then use it as an argument in \code{\link{run_COSMOS_metabolism_to_signaling}} or
19 | \code{\link{run_COSMOS_signaling_to_metabolism}}.
20 | If you choose CPLEX or CBC, you must modify then the solverPath field and point to
21 | the CPLEX/CBC executable (See Details).
22 | }
23 | \details{
24 | COSMOS is dependent on CARNIVAL for exhibiting the signalling pathway optimisation.
25 | CARNIVAL requires the interactive version of IBM Cplex, Gurobi or CBC-COIN solver
26 | as the network optimiser. The IBM ILOG Cplex is freely available through
27 | Academic Initiative \href{https://www.ibm.com/products/ilog-cplex-optimization-studio}{here}.
28 | Gurobi license is also free for academics, request a license following instructions
29 | \href{https://www.gurobi.com/downloads/end-user-license-agreement-academic/}{here}.
30 | The \href{https://projects.coin-or.org/Cbc}{CBC} solver is open source and freely
31 | available for any user, but has a significantly lower performance than CPLEX or
32 | Gurobi. Obtain CBC executable directly usable for cosmos
33 | \href{https://ampl.com/products/solvers/open-source/#cbc}{here}. Alternatively for
34 | small networks, users can rely on the freely available
35 | \href{https://cran.r-project.org/web/packages/lpSolve/index.html}{lpSolve R-package},
36 | which is automatically installed with the package.
37 | }
38 | \examples{
39 | # load and change default options:
40 | my_options = default_CARNIVAL_options(solver = "cplex")
41 |
42 | my_options$solverPath = "/Applications/CPLEX_Studio128/cplex/bin/x86-64_osx/cplex"
43 | my_options$threads = 2
44 | my_options$timelimit = 3600*15
45 | }
46 |
--------------------------------------------------------------------------------
/man/display_node_neighboorhood.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/display_node_neighboorhood.R
3 | \name{display_node_neighboorhood}
4 | \alias{display_node_neighboorhood}
5 | \title{display_node_neighboorhood}
6 | \usage{
7 | display_node_neighboorhood(central_node, sif, att, n = 100)
8 | }
9 | \arguments{
10 | \item{central_node}{character or character vector; node ID(s) around which a
11 | network will be branched out untill meansurments and input are reached}
12 |
13 | \item{sif}{df; COSMOS network solution in sif format like the first list
14 | element returned by the format_cosmos_res function}
15 |
16 | \item{att}{df; attributes of the nodes of the COMSOS network solution like
17 | the second list element returned by the format_cosmos_res function}
18 |
19 | \item{n}{numeric; maximum number of steps in the network to look for inputs
20 | and measurments}
21 | }
22 | \value{
23 | a visnetwork object
24 | }
25 | \description{
26 | display input and measurements within n steps of a given set of nodes
27 | }
28 | \examples{
29 | CARNIVAL_options <- cosmosR::default_CARNIVAL_options("lpSolve")
30 | data(toy_network)
31 | data(toy_signaling_input)
32 | data(toy_metabolic_input)
33 | data(toy_RNA)
34 | test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
35 | signaling_data = toy_signaling_input,
36 | metabolic_data = toy_metabolic_input,
37 | diff_expression_data = toy_RNA,
38 | maximum_network_depth = 15,
39 | remove_unexpressed_nodes = TRUE,
40 | CARNIVAL_options = CARNIVAL_options
41 | )
42 | test_result_for <- run_COSMOS_signaling_to_metabolism(data = test_for,
43 | CARNIVAL_options = CARNIVAL_options)
44 | test_result_for <- format_COSMOS_res(test_result_for)
45 | network_plot <- display_node_neighboorhood(central_node = 'MYC',
46 | sif = test_result_for[[1]],
47 | att = test_result_for[[2]],
48 | n = 7)
49 | network_plot
50 | }
51 |
--------------------------------------------------------------------------------
/man/extract_nodes_for_ORA.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_ORA.R
3 | \name{extract_nodes_for_ORA}
4 | \alias{extract_nodes_for_ORA}
5 | \title{Extract COSMOS nodes for ORA analysis}
6 | \usage{
7 | extract_nodes_for_ORA(sif, att)
8 | }
9 | \arguments{
10 | \item{sif}{df; COSMOS network solution in sif format like the first list
11 | element returned by the format_cosmos_res function}
12 |
13 | \item{att}{df; attributes of the nodes of the COMSOS network solution like
14 | the second list element returned by the format_cosmos_res function}
15 | }
16 | \value{
17 | List with 2 objects: the success and the background genes
18 | }
19 | \description{
20 | Function to extract the nodes that appear in the COSMOS output network and
21 | the background genes (all genes present in the prior knowledge network)
22 | }
23 | \examples{
24 | CARNIVAL_options <- cosmosR::default_CARNIVAL_options("lpSolve")
25 | data(toy_network)
26 | data(toy_signaling_input)
27 | data(toy_metabolic_input)
28 | data(toy_RNA)
29 | test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
30 | signaling_data = toy_signaling_input,
31 | metabolic_data = toy_metabolic_input,
32 | diff_expression_data = toy_RNA,
33 | maximum_network_depth = 15,
34 | remove_unexpressed_nodes = TRUE,
35 | CARNIVAL_options = CARNIVAL_options
36 | )
37 | test_result_for <- run_COSMOS_signaling_to_metabolism(data = test_for,
38 | CARNIVAL_options = CARNIVAL_options)
39 | test_result_for <- format_COSMOS_res(test_result_for)
40 | extreacted_nodes <- extract_nodes_for_ORA(
41 | sif = test_result_for[[1]],
42 | att = test_result_for[[2]]
43 | )
44 | }
45 |
--------------------------------------------------------------------------------
/man/figures/graphical_abstract.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/man/figures/graphical_abstract.png
--------------------------------------------------------------------------------
/man/figures/intro_data.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/man/figures/intro_data.png
--------------------------------------------------------------------------------
/man/figures/logo.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/man/figures/logo.png
--------------------------------------------------------------------------------
/man/figures/summary.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/man/figures/summary.png
--------------------------------------------------------------------------------
/man/filter_incohrent_TF_target.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{filter_incohrent_TF_target}
4 | \alias{filter_incohrent_TF_target}
5 | \title{filter_incohrent_TF_target}
6 | \usage{
7 | filter_incohrent_TF_target(
8 | decouplRnival_res,
9 | TF_reg_net,
10 | meta_network,
11 | RNA_input
12 | )
13 | }
14 | \arguments{
15 | \item{decouplRnival_res}{A data frame resulting from the decoupleRnival function.}
16 |
17 | \item{TF_reg_net}{A data frame containing prior knowledge of transcription factor (TF) regulatory interactions.}
18 |
19 | \item{meta_network}{A network data frame containing signed directed prior knowledge of molecular interactions.}
20 |
21 | \item{RNA_input}{A named vector containing differential gene expression data.}
22 | }
23 | \value{
24 | A network data frame containing the genes that are not incoherently regulated by TFs.
25 | }
26 | \description{
27 | Filters incoherent target genes from a regulatory network based on a decoupling analysis
28 | of upstream and downstream gene expression.
29 | }
30 | \examples{
31 | # Example input data
32 | upstream_input <- c("A" = 1, "B" = -1, "C" = 0.5)
33 | downstream_input <- c("D" = 2, "E" = -1.5)
34 | meta_network <- data.frame(
35 | source = c("A", "A", "B", "C", "C", "D", "E"),
36 | target = c("B", "D", "D", "E", "D", "B", "A"),
37 | interaction = c(-1, 1, -1, 1, -1, -1, 1)
38 | )
39 | RNA_input <- c("A" = 1, "B" = -1, "C" = 5, "D" = -0.7, "E" = -0.3)
40 |
41 | TF_reg_net <- data.frame(
42 | source = c("B"),
43 | target = c("D"),
44 | mor = c(-1)
45 | )
46 |
47 | # Run the decoupleRnival function to get the upstream influence scores
48 | upstream_scores <- decoupleRnival(upstream_input, downstream_input, meta_network, n_layers = 2, n_perm = 100)
49 |
50 | filtered_network <- filter_incohrent_TF_target(upstream_scores, TF_reg_net, meta_network, RNA_input)
51 |
52 | print(filtered_network)
53 | }
54 |
--------------------------------------------------------------------------------
/man/format_COSMOS_res.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/format_COSMOS_results.R
3 | \name{format_COSMOS_res}
4 | \alias{format_COSMOS_res}
5 | \title{format_COSMOS_res}
6 | \usage{
7 | format_COSMOS_res(cosmos_res, metab_mapping = NULL)
8 | }
9 | \arguments{
10 | \item{cosmos_res}{results of COSMOS run}
11 |
12 | \item{metab_mapping}{a named vector with HMDB Ids as names and desired metabolite names as values.}
13 | }
14 | \value{
15 | list with network and attribute tables.
16 | }
17 | \description{
18 | formats the network with readable names
19 | }
20 |
--------------------------------------------------------------------------------
/man/format_LR_ressource.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_decoupleR.R
3 | \name{format_LR_ressource}
4 | \alias{format_LR_ressource}
5 | \title{Format Ligand-Receptor Resource}
6 | \usage{
7 | format_LR_ressource(ligrec_ressource)
8 | }
9 | \arguments{
10 | \item{ligrec_ressource}{A data frame representing the ligand-receptor resource with columns for source and target gene symbols.}
11 | }
12 | \value{
13 | A data frame containing the formatted ligand-receptor gene set with columns:
14 | \item{gene}{The gene symbol from the ligand-receptor pairs.}
15 | \item{set}{The set identifier combining source and target gene symbols.}
16 | \item{mor}{Default value set to 1 for all entries.}
17 | \item{likelihood}{Default value set to 1 for all entries.}
18 | }
19 | \description{
20 | This function formats a ligand-receptor resource by creating a gene set
21 | with source-target pairs, converting it to a long format, and adding
22 | default values for 'mor' and 'likelihood'.
23 | }
24 | \examples{
25 | # Create a sample ligand-receptor resource
26 | ligrec_ressource <- data.frame(source_genesymbol = c("L1", "L2"),
27 | target_genesymbol = c("R1", "R2"))
28 |
29 | # Format the ligand-receptor resource
30 | formatted_geneset <- format_LR_ressource(ligrec_ressource)
31 |
32 | }
33 |
--------------------------------------------------------------------------------
/man/get_moon_scoring_network.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{get_moon_scoring_network}
4 | \alias{get_moon_scoring_network}
5 | \title{Get Moon Scoring Network}
6 | \usage{
7 | get_moon_scoring_network(
8 | upstream_node,
9 | meta_network,
10 | moon_scores,
11 | keep_upstream_node_peers = F
12 | )
13 | }
14 | \arguments{
15 | \item{upstream_node}{The node from which the network analysis starts.}
16 |
17 | \item{meta_network}{The complete network data.}
18 |
19 | \item{moon_scores}{Scores associated with each node in the network.}
20 |
21 | \item{keep_upstream_node_peers}{Logical; whether to keep peers of the upstream node. Default is FALSE.}
22 | }
23 | \value{
24 | A list with two elements:
25 | - `SIF`: A data frame representing the filtered meta network.
26 | - `ATT`: A data frame representing the updated moon scores.
27 | }
28 | \description{
29 | This function analyzes a given meta network based on moon scores and an upstream node.
30 | It filters and processes the network by controlling and observing neighbours
31 | according to specified parameters. The function returns a list containing a filtered
32 | network and updated moon scores.
33 | }
34 | \examples{
35 | # Example usage (requires appropriate data structures for meta_network and moon_scores)
36 | # result <- get_moon_scoring_network(upstream_node, meta_network, moon_scores)
37 |
38 | }
39 |
--------------------------------------------------------------------------------
/man/gmt_to_dataframe.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_ORA.R
3 | \name{gmt_to_dataframe}
4 | \alias{gmt_to_dataframe}
5 | \title{Convert gmt file to data frame}
6 | \usage{
7 | gmt_to_dataframe(gmtfile)
8 | }
9 | \arguments{
10 | \item{gmtfile}{a full path name of the gmt file to be converted}
11 | }
12 | \value{
13 | a two column data frame where the first column corresponds to omic
14 | features and the second column to associated terms (pathways).
15 | }
16 | \description{
17 | This function is designed to convert a gmt file (gene set file from MSigDB)
18 | into a two column data frame where the first column corresponds to omic
19 | features (genes) and the second column to associated terms (pathway the gene
20 | belongs to). One gene can belong to several pathways.
21 | }
22 |
--------------------------------------------------------------------------------
/man/load_tf_regulon_dorothea.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/load_tf_regulon_dorothea.R
3 | \name{load_tf_regulon_dorothea}
4 | \alias{load_tf_regulon_dorothea}
5 | \title{load transcription factor regulon}
6 | \usage{
7 | load_tf_regulon_dorothea(confidence = c("A", "B", "C"))
8 | }
9 | \arguments{
10 | \item{confidence}{strong vector (by default: c("A","B","C")). Subset of \{A, B,
11 | C, D, E\}. See the `dorothea` for the meaning of confidence levels.
12 | package for further details.}
13 | }
14 | \value{
15 | returns a PKN of a form of a data table. Each row is an interaction.
16 | Columns names are:
17 |
18 | - `tf` transcription factor
19 | - `confidence` class of confidence
20 | - `target` target gene
21 | - `sign` indicates if interaction is up (1) or down-regulation (-1).
22 | }
23 | \description{
24 | load the transcription factors from \code{DOROTHEA} package and converts
25 | gene symbols to EntrezID using org.Hs.eg.db
26 | }
27 | \examples{
28 | load_tf_regulon_dorothea()
29 | }
30 |
--------------------------------------------------------------------------------
/man/make_heatmap_color_palette.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_pheatmap.R
3 | \name{make_heatmap_color_palette}
4 | \alias{make_heatmap_color_palette}
5 | \title{Create Heatmap Color Palette}
6 | \usage{
7 | make_heatmap_color_palette(my_matrix)
8 | }
9 | \arguments{
10 | \item{my_matrix}{A numeric matrix for which the heatmap color palette is to be generated.}
11 | }
12 | \value{
13 | A character vector of colors representing the heatmap color palette based on the input matrix values.
14 | }
15 | \description{
16 | This function generates a color palette suitable for heatmaps based on the values in a matrix. It uses the `createLinearColors` function to generate separate color gradients for positive and negative values.
17 | }
18 | \examples{
19 | # Create a sample matrix
20 | my_matrix <- matrix(c(-3, -1, 0, 1, 3), nrow = 1)
21 |
22 | # Generate heatmap color palette
23 | heatmap_palette <- make_heatmap_color_palette(my_matrix)
24 |
25 | }
26 |
--------------------------------------------------------------------------------
/man/meta_network.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/meta_network.R
3 | \docType{data}
4 | \name{meta_network}
5 | \alias{meta_network}
6 | \title{Meta Prior Knowledge Network}
7 | \format{
8 | An object of class \dQuote{\code{tibble}} with 117065 rows
9 | (interactions) and three variables:
10 | \describe{
11 | \item{\code{source}}{Source node, either metabolite or protein}
12 | \item{\code{interaction}}{Type of interaction, 1 = Activation, -1 = Inhibition}
13 | \item{\code{target}}{Target node, either metabolite or protein}
14 | A detailed description of the identifier formatting can be found under
15 | \url{https://metapkn.omnipathdb.org/00__README.txt}.
16 | }
17 | }
18 | \source{
19 | The network is available in Omnipath:
20 | \url{https://metapkn.omnipathdb.org/metapkn__20200122.txt}, the scripts
21 | used for the build of the network are available under
22 | \url{https://github.com/saezlab/Meta_PKN}.
23 | }
24 | \usage{
25 | data(meta_network)
26 | }
27 | \description{
28 | Comprehensive Prior Knowledge Network (PKN), which combines signaling and
29 | metabolic interaction networks. The network was constructed using the Recon3D
30 | and STITCH metabolic networks as well as the signaling network from
31 | OmniPath.
32 | }
33 | \examples{
34 | data(meta_network)
35 |
36 | }
37 | \references{
38 | {
39 | Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
40 | Systems Biology}. \bold{17}, e9730.
41 | }
42 | }
43 | \keyword{datasets}
44 |
--------------------------------------------------------------------------------
/man/meta_network_cleanup.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{meta_network_cleanup}
4 | \alias{meta_network_cleanup}
5 | \title{meta_network_cleanup}
6 | \usage{
7 | meta_network_cleanup(meta_network)
8 | }
9 | \arguments{
10 | \item{meta_network}{A data frame with columns 'source', 'interaction', and 'target'.}
11 | }
12 | \value{
13 | A cleaned up meta network data frame.
14 | }
15 | \description{
16 | This function cleans up a meta network data frame by removing self-interactions,
17 | calculating the mean interaction values for duplicated source-target pairs, and
18 | keeping only interactions with values of 1 or -1.
19 | }
20 | \examples{
21 | # Create a meta network data frame
22 | example_meta_network <- data.frame(
23 | source = c("A", "B", "C", "D", "A", "B", "C", "D", "A"),
24 | interaction = c(1, 1, 1, -1, 1, -1, 1, -1, 1),
25 | target = c("B", "C", "D", "A", "C", "D", "A", "B", "B")
26 | )
27 |
28 | # Clean up the example meta network
29 | cleaned_meta_network <- meta_network_cleanup(example_meta_network)
30 | print(cleaned_meta_network)
31 |
32 | }
33 |
--------------------------------------------------------------------------------
/man/moon.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{moon}
4 | \alias{moon}
5 | \title{moon}
6 | \usage{
7 | moon(
8 | upstream_input = NULL,
9 | downstream_input,
10 | meta_network,
11 | n_layers,
12 | n_perm = 1000,
13 | downstream_cutoff = 0,
14 | statistic = "ulm",
15 | return_levels = F
16 | )
17 | }
18 | \arguments{
19 | \item{upstream_input}{A named vector with up_stream nodes and their corresponding activity.}
20 |
21 | \item{downstream_input}{A named vector with down_stream nodes and their corresponding activity.}
22 |
23 | \item{meta_network}{A network data frame containing signed directed prior knowledge of molecular interactions.}
24 |
25 | \item{n_layers}{The number of layers that will be propagated upstream.}
26 |
27 | \item{n_perm}{The number of permutations to use in decoupleR's algorithm.}
28 |
29 | \item{downstream_cutoff}{If downstream measurments should be included above a given threshold}
30 |
31 | \item{statistic}{the decoupleR stat to consider: "wmean", "norm_wmean", or "ulm"}
32 |
33 | \item{return_levels}{true or false, if true the layers that the protein belongs to will be returned alongside the scores}
34 | }
35 | \value{
36 | A data frame containing the score of the nodes upstream of the
37 | downstream input based on the iterative propagation
38 | }
39 | \description{
40 | Iteratively propagate downstream input activity through a signed directed network
41 | using the weighted mean enrichment score from decoupleR package
42 | }
43 | \examples{
44 | # Example input data
45 | upstream_input <- c("A" = 1, "B" = -1, "C" = 0.5)
46 | downstream_input <- c("D" = 2, "E" = -1.5)
47 | meta_network <- data.frame(
48 | source = c("A", "A", "B", "C", "C", "D", "E"),
49 | target = c("B", "C", "D", "E", "D", "B", "A"),
50 | sign = c(1, -1, -1, 1, -1, -1, 1)
51 | )
52 |
53 | # Run the function with the example input data
54 | result <- moon(upstream_input, downstream_input, meta_network, n_layers = 2, statistic = "wmean")
55 |
56 | # View the results
57 | print(result)
58 | }
59 |
--------------------------------------------------------------------------------
/man/prepare_metab_inputs.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/prepare_metab_inputs.R
3 | \name{prepare_metab_inputs}
4 | \alias{prepare_metab_inputs}
5 | \title{add metabolic compartment and metab__ prefix to metabolite IDs}
6 | \usage{
7 | prepare_metab_inputs(metab_input, compartment_codes)
8 | }
9 | \arguments{
10 | \item{metab_input}{a named vector with matebolic statistics as inputs and metabolite identifiers as names}
11 |
12 | \item{compartment_codes}{a character vector, the desired compartment codes to be added. Possible values are "r", "c", "e", "x", "m", "l", "n" and "g"}
13 | }
14 | \value{
15 | a named vector with the compartment code and prefixed added to the names
16 | }
17 | \description{
18 | This function adds metabolic compartments to the metabolic identifiers provided by the user.
19 | }
20 |
--------------------------------------------------------------------------------
/man/preprocess_COSMOS_metabolism_to_signaling.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/preprocess_COSMOS_metabolism_to_signaling.R
3 | \name{preprocess_COSMOS_metabolism_to_signaling}
4 | \alias{preprocess_COSMOS_metabolism_to_signaling}
5 | \title{Preprocess COSMOS Inputs For Metabolism to Signaling}
6 | \usage{
7 | preprocess_COSMOS_metabolism_to_signaling(
8 | meta_network = meta_network,
9 | tf_regulon = load_tf_regulon_dorothea(),
10 | signaling_data,
11 | metabolic_data,
12 | diff_expression_data = NULL,
13 | diff_exp_threshold = 1,
14 | maximum_network_depth = 8,
15 | expressed_genes = NULL,
16 | remove_unexpressed_nodes = TRUE,
17 | filter_tf_gene_interaction_by_optimization = TRUE,
18 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")
19 | )
20 | }
21 | \arguments{
22 | \item{meta_network}{prior knowledge network (PKN). A PKN released with COSMOS
23 | and derived from Omnipath, STITCHdb and Recon3D can be used. See details on
24 | the data \code{\link{meta_network}}.}
25 |
26 | \item{tf_regulon}{collection of transcription factor - target interactions.
27 | A default collection from dorothea can be obtained by the
28 | \code{\link{load_tf_regulon_dorothea}} function.}
29 |
30 | \item{signaling_data}{numerical vector, where names are signaling nodes
31 | in the PKN and values are from \{1, 0, -1\}. Continuous data will be
32 | discretized using the \code{\link{sign}} function.}
33 |
34 | \item{metabolic_data}{numerical vector, where names are metabolic nodes
35 | in the PKN and values are continuous values that represents log2 fold change
36 | or t-values from a differential analysis. These values are compared to
37 | the simulation results (simulated nodes can take value -1, 0 or 1)}
38 |
39 | \item{diff_expression_data}{(optional) numerical vector that represents the
40 | results of a differential gene expression analysis. Names are gene
41 | names using gene symbole and values are log fold change or
42 | t-values. We use the \dQuote{\code{diff_exp_threshold}} parameter to decide
43 | which genes changed significantly. Genes with NA values are considered none
44 | expressed and they will be removed from the TF-gene expression interactions.}
45 |
46 | \item{diff_exp_threshold}{threshold parameter (default 1) used to binarize
47 | the values of \dQuote{\code{diff_expression_data}}.}
48 |
49 | \item{maximum_network_depth}{integer > 0 (default: 8). Nodes that are further
50 | than \dQuote{\code{maximum_network_depth}} steps from the signaling nodes on
51 | the directed graph of the PKN are considered non-reachable and are removed.}
52 |
53 | \item{expressed_genes}{character vector. Names of nodes that are expressed. By
54 | default we consider all the nodes that appear in \code{diff_expression_data} with
55 | a numeric value (i.e. nodes with NA are removed)}
56 |
57 | \item{remove_unexpressed_nodes}{if TRUE (default) removes nodes from the PKN
58 | that are not expressed, see input \dQuote{\code{expressed_genes}}.}
59 |
60 | \item{filter_tf_gene_interaction_by_optimization}{(default:TRUE), if TRUE then runs
61 | a network optimization that estimates TF activity not included in the inputs
62 | and checks the consistency between the estimated activity and change in gene
63 | expression. Removes interactions where TF and gene expression are inconsistent}
64 |
65 | \item{CARNIVAL_options}{list that controls the options of CARNIVAL. See details
66 | in \code{\link{default_CARNIVAL_options}}.}
67 | }
68 | \value{
69 | cosmos_data object with the following fields:
70 | \describe{
71 | \item{\code{meta_network}}{Filtered PKN}
72 | \item{\code{tf_regulon}}{TF - target regulatory network}
73 | \item{\code{signaling_data_bin}}{Binarised signaling data}
74 | \item{\code{metabolic_data}}{Metabolomics data}
75 | \item{\code{diff_expression_data_bin}}{Binarized gene expression data}
76 | \item{\code{optimized_network}}{Initial optimized network if
77 | \code{filter_tf_gene_interaction_by_optimization is TRUE}}
78 | }
79 | }
80 | \description{
81 | Runs checks on the input data and simplifies the prior knowledge network.
82 | Simplification includes the removal of (1) nodes that are not reachable from
83 | signaling nodes and (2) interactions between transcription factors and target
84 | genes if the target gene does not respond or the response is contradictory
85 | with the change in the transcription factor activity.
86 | Optionally, further TF activities are estimated via network optimization via
87 | CARNIVAL and the interactions between TF and genes are filtered again.
88 | }
89 | \examples{
90 | data(toy_network)
91 | data(toy_signaling_input)
92 | data(toy_metabolic_input)
93 | data(toy_RNA)
94 | test_back <- preprocess_COSMOS_metabolism_to_signaling(
95 | meta_network = toy_network,
96 | signaling_data = toy_signaling_input,
97 | metabolic_data = toy_metabolic_input,
98 | diff_expression_data = toy_RNA,
99 | maximum_network_depth = 15,
100 | remove_unexpressed_nodes = TRUE,
101 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")
102 | )
103 | }
104 | \seealso{
105 | \code{\link{meta_network}} for meta PKN,
106 | \code{\link{load_tf_regulon_dorothea}} for tf regulon,
107 | \code{\link[CARNIVAL]{runCARNIVAL}}.
108 | }
109 |
--------------------------------------------------------------------------------
/man/preprocess_COSMOS_signaling_to_metabolism.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/preprocess_COSMOS_signaling_to_metabolism.R
3 | \name{preprocess_COSMOS_signaling_to_metabolism}
4 | \alias{preprocess_COSMOS_signaling_to_metabolism}
5 | \title{Preprocess COSMOS Inputs For Signaling to Metabolism}
6 | \usage{
7 | preprocess_COSMOS_signaling_to_metabolism(
8 | meta_network = meta_network,
9 | tf_regulon = load_tf_regulon_dorothea(),
10 | signaling_data,
11 | metabolic_data,
12 | diff_expression_data = NULL,
13 | diff_exp_threshold = 1,
14 | maximum_network_depth = 8,
15 | expressed_genes = NULL,
16 | remove_unexpressed_nodes = TRUE,
17 | filter_tf_gene_interaction_by_optimization = TRUE,
18 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")
19 | )
20 | }
21 | \arguments{
22 | \item{meta_network}{prior knowledge network (PKN). A PKN released with COSMOS
23 | and derived from Omnipath, STITCHdb and Recon3D can be used. See details on
24 | the data \code{\link{meta_network}}.}
25 |
26 | \item{tf_regulon}{collection of transcription factor - target interactions.
27 | A default collection from dorothea can be obtained by the
28 | \code{\link{load_tf_regulon_dorothea}} function.}
29 |
30 | \item{signaling_data}{numerical vector, where names are signaling nodes
31 | in the PKN and values are from \{1, 0, -1\}. Continuous data will be
32 | discretized using the \code{\link{sign}} function.}
33 |
34 | \item{metabolic_data}{numerical vector, where names are metabolic nodes
35 | in the PKN and values are continuous values that represents log2 fold change
36 | or t-values from a differential analysis. These values are compared to
37 | the simulation results (simulated nodes can take value -1, 0 or 1)}
38 |
39 | \item{diff_expression_data}{(optional) numerical vector that represents the
40 | results of a differential gene expression analysis. Names are gene
41 | names using gene symbole and values are log fold change or
42 | t-values. We use the \dQuote{\code{diff_exp_threshold}} parameter to decide
43 | which genes changed significantly. Genes with NA values are considered none
44 | expressed and they will be removed from the TF-gene expression interactions.}
45 |
46 | \item{diff_exp_threshold}{threshold parameter (default 1) used to binarize
47 | the values of \dQuote{\code{diff_expression_data}}.}
48 |
49 | \item{maximum_network_depth}{integer > 0 (default: 8). Nodes that are further
50 | than \dQuote{\code{maximum_network_depth}} steps from the signaling nodes on
51 | the directed graph of the PKN are considered non-reachable and are removed.}
52 |
53 | \item{expressed_genes}{character vector. Names of nodes that are expressed. By
54 | default we consider all the nodes that appear in \code{diff_expression_data} with
55 | a numeric value (i.e. nodes with NA are removed)}
56 |
57 | \item{remove_unexpressed_nodes}{if TRUE (default) removes nodes from the PKN
58 | that are not expressed, see input \dQuote{\code{expressed_genes}}.}
59 |
60 | \item{filter_tf_gene_interaction_by_optimization}{(default:TRUE), if TRUE then runs
61 | a network optimization that estimates TF activity not included in the inputs
62 | and checks the consistency between the estimated activity and change in gene
63 | expression. Removes interactions where TF and gene expression are inconsistent}
64 |
65 | \item{CARNIVAL_options}{list that controls the options of CARNIVAL. See details
66 | in \code{\link{default_CARNIVAL_options}}.}
67 | }
68 | \value{
69 | cosmos_data object with the following fields:
70 | \describe{
71 | \item{\code{meta_network}}{Filtered PKN}
72 | \item{\code{tf_regulon}}{TF - target regulatory network}
73 | \item{\code{signaling_data_bin}}{Binarised signaling data}
74 | \item{\code{metabolic_data}}{Metabolomics data}
75 | \item{\code{diff_expression_data_bin}}{Binarized gene expression data}
76 | \item{\code{optimized_network}}{Initial optimized network if
77 | \code{filter_tf_gene_interaction_by_optimization is TRUE}}
78 | }
79 | }
80 | \description{
81 | Runs checks on the input data and simplifies the prior knowledge network.
82 | Simplification includes the removal of (1) nodes that are not reachable from
83 | signaling nodes and (2) interactions between transcription factors and target
84 | genes if the target gene does not respond or the response is contradictory
85 | with the change in the transcription factor activity.
86 | Optionally, further TF activities are estimated via network optimization via
87 | CARNIVAL and the interactions between TF and genes are filtered again.
88 | }
89 | \examples{
90 | data(toy_network)
91 | data(toy_signaling_input)
92 | data(toy_metabolic_input)
93 | data(toy_RNA)
94 | test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
95 | signaling_data = toy_signaling_input,
96 | metabolic_data = toy_metabolic_input,
97 | diff_expression_data = toy_RNA,
98 | maximum_network_depth = 15,
99 | remove_unexpressed_nodes = TRUE,
100 | CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
101 | }
102 | \seealso{
103 | \code{\link{meta_network}} for meta PKN,
104 | \code{\link{load_tf_regulon_dorothea}} for tf regulon,
105 | \code{\link[CARNIVAL]{runCARNIVAL}}.
106 | }
107 |
--------------------------------------------------------------------------------
/man/print.cosmos_data.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/cosmos_data.R
3 | \name{print.cosmos_data}
4 | \alias{print.cosmos_data}
5 | \title{Print Cosmos Data Summary
6 | Print a summary of cosmos data.}
7 | \usage{
8 | \method{print}{cosmos_data}(x, ...)
9 | }
10 | \arguments{
11 | \item{x}{\code{\link{cosmos_data}} object. Use the
12 | \code{\link{preprocess_COSMOS_signaling_to_metabolism}} or
13 | \code{\link{preprocess_COSMOS_metabolism_to_signaling}} functions to
14 | create one.}
15 |
16 | \item{...}{Further print arguments passed to or from other methods.}
17 | }
18 | \value{
19 | input (invisible)
20 | }
21 | \description{
22 | Print Cosmos Data Summary
23 | Print a summary of cosmos data.
24 | }
25 | \seealso{
26 | \code{\link[base]{print}}, \code{\link{cosmos_data}}
27 | }
28 |
--------------------------------------------------------------------------------
/man/reduce_solution_network.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{reduce_solution_network}
4 | \alias{reduce_solution_network}
5 | \title{reduce_solution_network}
6 | \usage{
7 | reduce_solution_network(
8 | decoupleRnival_res,
9 | meta_network,
10 | cutoff,
11 | upstream_input,
12 | RNA_input = NULL,
13 | n_steps = 10
14 | )
15 | }
16 | \arguments{
17 | \item{decoupleRnival_res}{A data frame resulting from the decoupleRnival function.}
18 |
19 | \item{meta_network}{A network data frame containing signed directed prior knowledge of molecular interactions.}
20 |
21 | \item{cutoff}{The consistency threshold for filtering edges from the solution network.}
22 |
23 | \item{upstream_input}{A named vector with up_stream nodes and their corresponding activity.}
24 |
25 | \item{RNA_input}{A named vector containing differential gene expression data.}
26 |
27 | \item{n_steps}{The maximum number of steps from upstream input nodes to include in the solution network.}
28 | }
29 | \value{
30 | A list containing the solution network (SIF) and an attribute table (ATT) with gene expression data.
31 | }
32 | \description{
33 | Reduces a solution network based on a decoupling analysis of upstream and downstream gene expression,
34 | by filtering out edges that do not meet a consistency threshold, and limiting the network to a
35 | certain number of steps from upstream input nodes.
36 | }
37 | \examples{
38 | # Example input data
39 | upstream_input <- c("A" = 1, "B" = -1, "C" = 0.5)
40 | downstream_input <- c("D" = 2, "E" = -1.5)
41 | meta_network <- data.frame(
42 | source = c("A", "A", "B", "C", "C", "D", "E"),
43 | target = c("B", "D", "D", "E", "D", "B", "A"),
44 | interaction = c(-1, 1, -1, 1, -1, -1, 1)
45 | )
46 | RNA_input <- c("A" = 1, "B" = -1, "C" = 5, "D" = 0.7, "E" = -0.3)
47 |
48 | # Run the decoupleRnival function to get the upstream influence scores
49 | upstream_scores <- decoupleRnival(upstream_input, downstream_input, meta_network, n_layers = 2, n_perm = 100)
50 |
51 | # Reduce the solution network based on the upstream influence scores
52 | reduced_network <- reduce_solution_network(upstream_scores, meta_network, 0.4, upstream_input, RNA_input, 3)
53 |
54 | # View the resulting solution network and attribute table
55 | print(reduced_network$SIF)
56 | print(reduced_network$ATT)
57 | }
58 |
--------------------------------------------------------------------------------
/man/run_COSMOS_metabolism_to_signaling.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/run_COSMOS_metabolism_to_signaling.R
3 | \name{run_COSMOS_metabolism_to_signaling}
4 | \alias{run_COSMOS_metabolism_to_signaling}
5 | \title{run COSMOS metabolism to signaling}
6 | \usage{
7 | run_COSMOS_metabolism_to_signaling(
8 | data,
9 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")
10 | )
11 | }
12 | \arguments{
13 | \item{data}{\code{\link{cosmos_data}} object. Use the
14 | \code{\link{preprocess_COSMOS_metabolism_to_signaling}} function to create
15 | an instance.}
16 |
17 | \item{CARNIVAL_options}{List that controls the options of CARNIVAL. See details
18 | in \code{\link{default_CARNIVAL_options}}.}
19 | }
20 | \value{
21 | List with the following elements:
22 | \describe{
23 | \item{\code{weightedSIF}}{The averaged networks found by
24 | optimization in a format of a Simple Interaction network, i.e. each row
25 | codes an edge}
26 | \item{\code{N_networks}}{Number of solutions found by the
27 | optimization}
28 | \item{\code{nodesAttributes}}{Estimated node properties}
29 | \item{\code{individual_networks}}{List of optimial networks found}
30 | \item{\code{individual_networks_node_attributes}}{Node activity in each
31 | network}
32 | }
33 | }
34 | \description{
35 | Runs COSMOS from metabolism to signaling. This function uses CARNIVAL to find
36 | a subset of the prior knowledge network based on optimization that (1)
37 | includes the most measured and input nodes and (2) which is in agreement with
38 | the data. Use \code{\link{preprocess_COSMOS_metabolism_to_signaling}} to
39 | prepare the the inputs, measurements and the prior knowledge network.
40 | }
41 | \examples{
42 | data(toy_network)
43 | data(toy_signaling_input)
44 | data(toy_metabolic_input)
45 | data(toy_RNA)
46 | test_back <- preprocess_COSMOS_metabolism_to_signaling(meta_network = toy_network,
47 | signaling_data = toy_signaling_input,
48 | metabolic_data = toy_metabolic_input,
49 | diff_expression_data = toy_RNA,
50 | maximum_network_depth = 15,
51 | remove_unexpressed_nodes = TRUE,
52 | CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
53 |
54 | test_result_back <- run_COSMOS_metabolism_to_signaling(data = test_back,
55 | CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
56 | }
57 | \seealso{
58 | \code{\link{preprocess_COSMOS_metabolism_to_signaling}},
59 | \code{\link[CARNIVAL]{runCARNIVAL}}, \code{\link{cosmos_data}}
60 | }
61 |
--------------------------------------------------------------------------------
/man/run_COSMOS_signaling_to_metabolism.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/run_COSMOS_signaling_to_metabolism.R
3 | \name{run_COSMOS_signaling_to_metabolism}
4 | \alias{run_COSMOS_signaling_to_metabolism}
5 | \title{run COSMOS signaling to metabolism}
6 | \usage{
7 | run_COSMOS_signaling_to_metabolism(
8 | data,
9 | CARNIVAL_options = default_CARNIVAL_options("lpSolve")
10 | )
11 | }
12 | \arguments{
13 | \item{data}{\code{\link{cosmos_data}} object. Use the
14 | \code{\link{preprocess_COSMOS_signaling_to_metabolism}} function to create
15 | an instance.}
16 |
17 | \item{CARNIVAL_options}{List that controls the options of CARNIVAL. See
18 | details in \code{\link{default_CARNIVAL_options}}.}
19 | }
20 | \value{
21 | List with the following elements:
22 | \describe{
23 | \item{\code{weightedSIF}}{The averaged networks found by
24 | optimization in a format of a Simple Interaction network, i.e. each row
25 | codes an edge}
26 | \item{\code{N_networks}}{Number of solutions found by the
27 | optimization}
28 | \item{\code{nodesAttributes}}{Estimated node properties}
29 | \item{\code{individual_networks}}{List of optimial networks found}
30 | \item{\code{individual_networks_node_attributes}}{Node activity in each
31 | network}
32 | }
33 | }
34 | \description{
35 | Runs COSMOS from signaling to metabolism. This function uses CARNIVAL to find
36 | a subset of the prior knowledge network based on optimisation that (1)
37 | includes the most measured and input nodes and (2) which is in agreement with
38 | the data. Use \code{\link{preprocess_COSMOS_signaling_to_metabolism}} to
39 | prepare inputs, measurements and prior knowledge network.
40 | }
41 | \examples{
42 | data(toy_network)
43 | data(toy_signaling_input)
44 | data(toy_metabolic_input)
45 | data(toy_RNA)
46 | test_for <- preprocess_COSMOS_signaling_to_metabolism(meta_network = toy_network,
47 | signaling_data = toy_signaling_input,
48 | metabolic_data = toy_metabolic_input,
49 | diff_expression_data = toy_RNA,
50 | maximum_network_depth = 15,
51 | remove_unexpressed_nodes = TRUE,
52 | CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
53 |
54 | test_result_for <- run_COSMOS_signaling_to_metabolism(data = test_for,
55 | CARNIVAL_options = default_CARNIVAL_options("lpSolve"))
56 | }
57 | \seealso{
58 | \code{\link{preprocess_COSMOS_metabolism_to_signaling}},
59 | \code{\link[CARNIVAL]{runCARNIVAL}}, \code{\link{cosmos_data}}
60 | }
61 |
--------------------------------------------------------------------------------
/man/toy_RNA.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/toy_RNA.R
3 | \docType{data}
4 | \name{toy_RNA}
5 | \alias{toy_RNA}
6 | \title{Toy Input Transcription Data Set}
7 | \format{
8 | An object of class \dQuote{\code{numeric}} containing the t-values of
9 | 9300 genes, which are named with gene symboles matching the toy network.
10 | }
11 | \source{
12 | \url{https://github.com/saezlab/COSMOS_MSB/blob/main/data/RNA_ttop_tumorvshealthy.csv}
13 | }
14 | \usage{
15 | data(toy_RNA)
16 | }
17 | \description{
18 | This exemplary transcription data are the specific deregulated gene expression of the 786-O cell line from the NCI60 dataset.
19 | }
20 | \examples{
21 | data(toy_RNA)
22 |
23 | }
24 | \references{
25 | {
26 | Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
27 | Systems Biology}. \bold{17}, e9730.
28 | }
29 | }
30 | \keyword{datasets}
31 |
--------------------------------------------------------------------------------
/man/toy_metabolic_input.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/toy_metabolic_input.R
3 | \docType{data}
4 | \name{toy_metabolic_input}
5 | \alias{toy_metabolic_input}
6 | \title{Toy Metabolic Input Data}
7 | \format{
8 | An object of class \dQuote{\code{numeric}} containing the t-values of
9 | 2 metabolites, which are named with metabolite HMDB Ids matching the
10 | toy network.
11 | }
12 | \source{
13 | Subset of:
14 | \url{https://github.com/saezlab/COSMOS_MSB/blob/main/data/metab_input_COSMOS.csv}
15 | }
16 | \usage{
17 | data(toy_metabolic_input)
18 | }
19 | \description{
20 | This metabolic data are a subset from the metabolic measurements of the 786-O cell line from the NCI60 dataset.
21 | }
22 | \examples{
23 | data(toy_metabolic_input)
24 |
25 | }
26 | \references{
27 | {
28 | Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
29 | Systems Biology}. \bold{17}, e9730.
30 | }
31 | }
32 | \keyword{datasets}
33 |
--------------------------------------------------------------------------------
/man/toy_network.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/toy_network.R
3 | \docType{data}
4 | \name{toy_network}
5 | \alias{toy_network}
6 | \title{Toy Input Network}
7 | \format{
8 | An object of class \dQuote{\code{data.frame}} with 19 rows
9 | (interactions) and three variables:
10 | \describe{
11 | \item{\code{source}}{Source node, either metabolite or protein}
12 | \item{\code{interaction}}{Type of interaction, 1 = Activation, -1 = Inhibition}
13 | \item{\code{target}}{Target node, either metabolite or protein}
14 | A detailed description of the identifier formatting can be found under
15 | \url{https://metapkn.omnipathdb.org/00__README.txt}.
16 | }
17 | }
18 | \source{
19 | The network data are available on github:
20 | \url{https://github.com/saezlab/COSMOS_MSB/tree/main/results/COSMOS_result/COSMOS_res_session.RData}.
21 | The toy_network is the combined network of the COSMOS network solutions
22 | CARNIVAL_Result2 and CARNIVAL_Result_rerun subsequently reduced to 19
23 | exemplary nodes.
24 | }
25 | \usage{
26 | data(toy_network)
27 | }
28 | \description{
29 | This signaling network is the reduced COSMOS network solution obtained in the
30 | cosmos test on 786-O NCI60 data. Here, this network solution is reused as an
31 | exemplary input prior knowledge network (PKN).
32 | }
33 | \examples{
34 | data(toy_network)
35 |
36 | }
37 | \references{
38 | {
39 | Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
40 | Systems Biology}. \bold{17}, e9730.
41 | }
42 | }
43 | \keyword{datasets}
44 |
--------------------------------------------------------------------------------
/man/toy_signaling_input.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/toy_signaling_input.R
3 | \docType{data}
4 | \name{toy_signaling_input}
5 | \alias{toy_signaling_input}
6 | \title{Toy Signaling Input}
7 | \format{
8 | An object of class \dQuote{\code{data.frame}} containing the normalised
9 | enrichment scores (NES) of 2 signaling proteins, which are named with their
10 | respective gene Entrez ID matching the toy network.
11 | }
12 | \source{
13 | Subset of:
14 | \url{https://github.com/saezlab/COSMOS_MSB/blob/main/data/signaling_input_COSMOS.csv}
15 | }
16 | \usage{
17 | data(toy_signaling_input)
18 | }
19 | \description{
20 | This signaling data are a subset of the footprint-based signaling activity
21 | estimates of transcription factors of the 786-O cell line from the NCI60 dataset.
22 | }
23 | \examples{
24 | data(toy_signaling_input)
25 |
26 | }
27 | \references{
28 | {
29 | Dugourd, A., Kuppe, C. and Sciacovelli, M. et. al. (2021) \emph{Molecular
30 | Systems Biology}. \bold{17}, e9730.
31 | }
32 | }
33 | \keyword{datasets}
34 |
--------------------------------------------------------------------------------
/man/translate_column_HMDB.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_decoupleR.R
3 | \name{translate_column_HMDB}
4 | \alias{translate_column_HMDB}
5 | \title{Translate Column Using HMDB Mapper}
6 | \usage{
7 | translate_column_HMDB(my_column, HMDB_mapper_vec)
8 | }
9 | \arguments{
10 | \item{my_column}{A vector of values to be translated.}
11 |
12 | \item{HMDB_mapper_vec}{A named vector where the names are the original identifiers and the values are the corresponding HMDB identifiers.}
13 | }
14 | \value{
15 | A vector with the translated values.
16 | }
17 | \description{
18 | This function translates the values in a column using a provided Human Metabolome Database (HMDB) mapper vector.
19 | It modifies the input values by replacing certain prefixes and suffixes according to specific rules.
20 | }
21 | \examples{
22 | # Create a sample column and HMDB mapper vector
23 | my_column <- c("Metab__1234_a", "Gene5678_b", "Metab__91011_c")
24 | HMDB_mapper_vec <- c("1234" = "HMDB00001", "5678" = "HMDB00002", "91011" = "HMDB00003")
25 |
26 | # Translate the column
27 | translated_column <- translate_column_HMDB(my_column, HMDB_mapper_vec)
28 |
29 | }
30 |
--------------------------------------------------------------------------------
/man/translate_res.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/decoupleRnival.R
3 | \name{translate_res}
4 | \alias{translate_res}
5 | \title{translate_res}
6 | \usage{
7 | translate_res(SIF, ATT, HMDB_mapper_vec = NULL)
8 | }
9 | \arguments{
10 | \item{SIF}{result SIF of decoupleRnival pipeline}
11 |
12 | \item{ATT}{result ATT of decoupleRnival pipeline}
13 |
14 | \item{HMDB_mapper_vec}{a named vector with HMDB Ids as names and desired metabolite names as values.}
15 | }
16 | \value{
17 | list with network and attribute tables.
18 | }
19 | \description{
20 | formats the network with readable names
21 | }
22 | \examples{
23 | # Create a meta network data frame
24 | example_SIF <- data.frame(
25 | source = c("GPX1", "Gene863__GPX1"),
26 | target = c("Gene863__GPX1", "Metab__HMDB0003337_c"),
27 | sign = c(1, 1)
28 | )
29 |
30 | example_ATT <- data.frame(
31 | Nodes = c("GPX1", "Gene863__GPX1","Metab__HMDB0003337_c"),
32 | sign = c(1, 1, 1)
33 | )
34 |
35 | example_SIF
36 |
37 | data("HMDB_mapper_vec")
38 |
39 | translated_res <- translate_res(example_SIF,example_ATT,HMDB_mapper_vec)
40 |
41 | translated_res$SIF
42 | }
43 |
--------------------------------------------------------------------------------
/man/wide_ulm_res.Rd:
--------------------------------------------------------------------------------
1 | % Generated by roxygen2: do not edit by hand
2 | % Please edit documentation in R/support_decoupleR.R
3 | \name{wide_ulm_res}
4 | \alias{wide_ulm_res}
5 | \title{Convert ULM Results to Wide Format}
6 | \usage{
7 | wide_ulm_res(ulm_result)
8 | }
9 | \arguments{
10 | \item{ulm_result}{A data frame representing the ULM results with columns: source, condition, and score.}
11 | }
12 | \value{
13 | A data frame in wide format where each row is a source and each column is a condition.
14 | }
15 | \description{
16 | This function converts the results from a ULM analysis to a wide format data frame.
17 | The input is a data frame with columns for source, condition, and score. The output
18 | is a data frame where each row represents a source and each column represents a condition,
19 | with the corresponding scores as values.
20 | }
21 | \examples{
22 | # Create a sample ULM result
23 | ulm_result <- data.frame(source = c("A", "A", "B", "B"),
24 | condition = c("cond1", "cond2", "cond1", "cond2"),
25 | score = c(0.5, 0.8, 0.3, 0.7))
26 |
27 | # Convert to wide format
28 | wide_ulm_result <- wide_ulm_res(ulm_result)
29 |
30 | }
31 |
--------------------------------------------------------------------------------
/pkgdown/_pkgdown.yml:
--------------------------------------------------------------------------------
1 | destination: docs
2 | navbar:
3 | structure:
4 | left: [home, intro, reference, articles, tutorials, news]
5 | right: [github, twitter, homepage]
6 | components:
7 | twitter:
8 | icon: "fab fa-twitter fa-lg"
9 | href: https://twitter.com/saezlab
10 | homepage:
11 | icon: "fas fa-university"
12 | href: https://saezlab.org
13 | type: dark
14 | bg: primary
15 | template:
16 | params:
17 | ganalytics: G-PV71G2Q3SF
18 | highlightcss: false
19 | bootstrap: 5
20 | bootswatch: cosmo
21 | bslib:
22 | primary: "#285FA9"
23 | assets: vignettes/MOFA_to_COSMOS_files, vignettes/net_compr_MOON_files
24 |
--------------------------------------------------------------------------------
/pkgdown/extra.css:
--------------------------------------------------------------------------------
1 | .contents p:first-of-type img {
2 | background-color: white;
3 | padding: 10px;
4 | margin-top: 15px;
5 | scale: 1.1;
6 | position: relative;
7 | left: 10%;
8 | }
9 |
10 | body {
11 | font-size: 117%!important;
12 | }
13 |
14 | .bg-primary {
15 | background-color: #285FA9!important;
16 | }
17 |
18 | nav[data-toggle="toc"] .nav > li > a {
19 | border-radius: 0px!important;
20 | padding-left: 1rem!important;
21 | }
22 |
23 | pre {
24 | border-radius: 0px!important;
25 | background-color: white!important;
26 | }
27 |
28 | code span, code a:any-link {
29 | color: #1f1c1b!important;
30 | }/* Normal */
31 | code span.al, code span.al a:any-link {
32 | color: #bf0303!important;
33 | background-color: #f7e6e6!important;
34 | font-weight: bold;
35 | } /* Alert */
36 | code span.an, code span.an a:any-link {
37 | color: #ca60ca!important;
38 | } /* Annotation */
39 | code span.at, code span.at a:any-link {
40 | color: #0057ae!important;
41 | } /* Attribute */
42 | code span.bn, code span.bn a:any-link {
43 | color: #b08000!important;
44 | } /* BaseN */
45 | code span.bu, code span.bu a:any-link {
46 | color: #644a9b!important;
47 | font-weight: bold;
48 | } /* BuiltIn */
49 | code span.cf, code span.cf a:any-link {
50 | color: #1f1c1b!important;
51 | font-weight: bold;
52 | } /* ControlFlow */
53 | code span.ch, code span.ch a:any-link {
54 | color: #924c9d!important;
55 | } /* Char */
56 | code span.cn, code span.cn a:any-link {
57 | color: #aa5500!important;
58 | }/* Constant */
59 | code span.co, code span.co a:any-link {
60 | color: #898887!important;
61 | }/* Comment */
62 | code span.cv, code span.cv a:any-link {
63 | color: #0095ff!important;
64 | }/* CommentVar */
65 | code span.do, code span.do a:any-link {
66 | color: #607880!important;
67 | }/* Documentation */
68 | code span.dt, code span.dt a:any-link {
69 | color: #0057ae!important;
70 | }/* DataType */
71 | code span.dv, code span.dv a:any-link {
72 | color: #b08000!important;
73 | }/* DecVal */
74 | code span.er, code span.er a:any-link {
75 | color: #bf0303!important;
76 | text-decoration: underline;
77 | } /* Error */
78 | code span.ex, code span.ex a:any-link {
79 | color: #0095ff!important;
80 | font-weight: bold;
81 | } /* Extension */
82 | code span.fl, code span.fl a:any-link {
83 | color: #b08000!important;
84 | }/* Float */
85 | code span.fu,
86 | code span.fu a:any-link {
87 | color: #644a9b!important;
88 | } /* Function */
89 | code span.im, code span.im a:any-link {
90 | color: #ff5500!important;
91 | }/* Import */
92 | code span.in, code span.in a:any-link {
93 | color: #b08000!important;
94 | }/* Information */
95 | code span.kw, code span.kw a:any-link {
96 | /*color: #1f1c1b!important;*/
97 | color: #007BA5!important;
98 | font-weight: bold;
99 | } /* Keyword */
100 | code span.op, code span.op a:any-link {
101 | /* color: #1f1c1b!important; */
102 | color: #5E5E5E!important;
103 | }/* Operator */
104 | code span.ot, code span.ot a:any-link {
105 | color: #006e28!important;
106 | }/* Other */
107 | code span.pp, code span.pp a:any-link {
108 | color: #006e28!important;
109 | }/* Preprocessor */
110 | code span.re, code span.re a:any-link {
111 | color: #0057ae!important;
112 | background-color: #e0e9f8!important;
113 | } /* RegionMarker */
114 | code span.sc, code span.sc a:any-link {
115 | color: #3daee9!important;
116 | }/* SpecialChar */
117 | code span.ss, code span.ss a:any-link {
118 | color: #ff5500!important;
119 | }/* SpecialString */
120 | /* code span.st, code span.st a:any-link {
121 | color: #bf0303!important;
122 | }/* String */
123 | code span.st, code span.st a:any-link {
124 | color: #20794d!important;
125 | }/* String */
126 | code span.va, code span.va a:any-link {
127 | color: #0057ae!important;
128 | }/* Variable */
129 | code span.vs, code span.vs a:any-link {
130 | color: #bf0303!important;
131 | }/* VerbatimString */
132 | code span.wa, code span.wa a:any-link {
133 | color: #bf0303!important;
134 | }/* Warning */
135 |
136 | .nav-text.text-muted {
137 | color: #ffffff!important;
138 | }
139 |
140 | .nav-item.active > .nav-link {
141 | background-color: #e9ecef!important;
142 | color: #285FA9!important;
143 | }
144 |
145 | .navbar-dark input[type="search"] {
146 | background-color: #e9ecef!important;
147 | color: #212529!important;
148 | }
149 |
150 | .template-home > .row > #main > p > img {
151 | background-color: #ffffff!important;
152 | padding-top: 20px;
153 | }
154 |
155 | .template-home > .row > #main > .section > .page-header > img {
156 | display: none!important;
157 | }
158 |
159 | .template-home h1, .template-home h2, .template-home h3, .template-home h4,
160 | .template-home h5, .template-home h6, .template-article .page-header h1 {
161 | font-weight: 700!important;
162 | }
163 |
164 | h1#omnipathr {
165 | font-size: 4.125rem!important;
166 | }
167 |
168 | img.logo {
169 | background-color: #ffffff!important;
170 | }
171 |
172 | p.abstract {
173 | font-size: calc(1.375rem + 1.5vw)!important;
174 | }
175 |
176 | h4.author {
177 | font-size: 1.25rem!important;
178 | margin-top: 1rem!important;
179 | }
180 |
181 | @media (min-width: 1200px){
182 | p.abstract {
183 | font-size: 2.5rem!important;
184 | }
185 | }
186 |
187 | .author_afil {
188 | font-size: small!important;
189 | }
190 |
191 | .nav-item {
192 | margin-left: 10px!important;
193 | }
194 |
--------------------------------------------------------------------------------
/pkgdown/favicon/apple-touch-icon-120x120.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/apple-touch-icon-120x120.png
--------------------------------------------------------------------------------
/pkgdown/favicon/apple-touch-icon-152x152.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/apple-touch-icon-152x152.png
--------------------------------------------------------------------------------
/pkgdown/favicon/apple-touch-icon-180x180.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/apple-touch-icon-180x180.png
--------------------------------------------------------------------------------
/pkgdown/favicon/apple-touch-icon-60x60.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/apple-touch-icon-60x60.png
--------------------------------------------------------------------------------
/pkgdown/favicon/apple-touch-icon-76x76.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/apple-touch-icon-76x76.png
--------------------------------------------------------------------------------
/pkgdown/favicon/apple-touch-icon.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/apple-touch-icon.png
--------------------------------------------------------------------------------
/pkgdown/favicon/favicon-16x16.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/favicon-16x16.png
--------------------------------------------------------------------------------
/pkgdown/favicon/favicon-32x32.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/favicon-32x32.png
--------------------------------------------------------------------------------
/pkgdown/favicon/favicon.ico:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/pkgdown/favicon/favicon.ico
--------------------------------------------------------------------------------
/tests/testthat.R:
--------------------------------------------------------------------------------
1 | library(testthat)
2 | library(cosmosR)
3 |
4 | test_check("cosmosR")
5 |
--------------------------------------------------------------------------------
/tests/testthat/test-check_COSMOS_inputs.R:
--------------------------------------------------------------------------------
1 |
2 | test_that("test data", {
3 |
4 | # should not work with dataframe
5 | test_data = dplyr::tibble(ID = c("A","B"),activity = c(0.1,5))
6 | expect_error(check_COSMOS_inputs(signaling_data = test_data))
7 |
8 | # should pass with named vector
9 | test_data = c(X0125=4,X123684=-4)
10 | expect_true(check_COSMOS_inputs(signaling_data = test_data))
11 |
12 | })
13 |
14 | test_that("test meta network", {
15 |
16 | test_data = dplyr::tibble(source = c("A","B"),interaction = c(-1,1), target = c("C","D"))
17 | expect_true(check_COSMOS_inputs(meta_network = test_data))
18 |
19 | })
20 |
21 |
22 | test_that("test regulon network", {
23 |
24 | test_data = dplyr::tibble(tf = c("A","B"),sign = c(-1,1), target = c("C","D"))
25 | expect_true(check_COSMOS_inputs(tf_regulon = test_data))
26 |
27 | })
28 |
--------------------------------------------------------------------------------
/tests/testthat/test-cosmos_data.R:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 | test_that("construct cosmos_data", {
5 |
6 | meta_network_test <- cosmosR:::meta_network_test
7 | signaling_input_test <- cosmosR:::signaling_input_test
8 | expression_data_test <- cosmosR:::expression_data_test
9 | metabolic_data_test <- cosmosR:::metabolic_data_test
10 |
11 | cos <- new_cosmos_data(meta_network = meta_network_test,
12 | tf_regulon = load_tf_regulon_dorothea(),
13 | signaling_data = signaling_input_test,
14 | expression_data = expression_data_test,
15 | metabolic_data = metabolic_data_test)
16 | expect_true(is(cos,"cosmos_data"))
17 | })
18 |
19 |
20 |
21 | test_that("validate cosmos_data", {
22 |
23 | cos <- new_cosmos_data(meta_network = meta_network_test,
24 | tf_regulon = load_tf_regulon_dorothea(),
25 | signaling_data = signaling_input_test,
26 | expression_data = expression_data_test,
27 | metabolic_data = metabolic_data_test)
28 | expect_true({
29 | validate_cosmos_data(cos)
30 | TRUE
31 | })
32 | })
33 |
34 |
35 |
36 |
37 |
--------------------------------------------------------------------------------
/tests/testthat/test-load_tf_regulon_dorothea.R:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 | test_that("tf conversion", {
5 |
6 | reg = load_tf_regulon_dorothea(confidence = c("A","B","C"))
7 | # make sure A,B,C still contains a lot of interactions
8 | expect_true(nrow(reg) > 1)
9 | # should return 3 columns, correct names
10 | expect_true(ncol(reg) == 3)
11 | expect_true(all(c("tf","sign","target") %in% colnames(reg)))
12 | # should not contain missing values
13 | expect_true(all(complete.cases(reg)))
14 |
15 | })
16 |
--------------------------------------------------------------------------------
/tests/testthat/test-preprocess_COSMOS.R:
--------------------------------------------------------------------------------
1 |
2 | test_that("test cosmos preprocessing (signaling to metabolism)", {
3 |
4 | meta_network_test <- cosmosR:::meta_network_test
5 | signaling_input_test <- cosmosR:::signaling_input_test
6 | expression_data_test <- cosmosR:::expression_data_test
7 | metabolic_data_test <- cosmosR:::metabolic_data_test
8 |
9 | res <- preprocess_COSMOS_signaling_to_metabolism(meta_network = meta_network_test,
10 | signaling_data = signaling_input_test,
11 | metabolic_data = metabolic_data_test,
12 | diff_expression_data = expression_data_test,
13 | filter_tf_gene_interaction_by_optimization = FALSE)
14 | # check list:
15 | expect_length(res, 10)
16 | expect_true(all(c("meta_network",
17 | "tf_regulon",
18 | "signaling_data",
19 | "signaling_data_bin",
20 | "metabolic_data",
21 | "metabolic_data_bin",
22 | "expression_data",
23 | "diff_expression_data_bin",
24 | "optimized_network") %in% names(res)))
25 |
26 |
27 | # checking network
28 | expect_equal(ncol(res$meta_network), 3)
29 | # expect_equal(nrow(res$meta_network), 44005)
30 | expect_true(all(colnames(res$meta_network) %in% c("source","interaction","target")))
31 |
32 | # checking tf_regulon
33 | expect_true(all(c("tf","target","sign") %in% colnames(res$tf_regulon)))
34 |
35 | # checking signaling
36 | expect_true(is.vector(res$signaling_data_bin))
37 | #expect_equal(length(res$signaling_data_bin),103)
38 | #expect_equal(sum(grepl("^X",names(res$signaling_data_bin))),103)
39 | expect_true(all(res$signaling_data_bin %in% c(-1,0,1)))
40 |
41 | # checking metabolic
42 | expect_true(is.vector(res$metabolic_data))
43 | #expect_equal(length(res$metabolic_data),35)
44 | #expect_equal(sum(grepl("^XMetab",names(res$metabolic_data))),35)
45 |
46 | # check diff_expression_data_bin
47 | expect_true(is.vector(res$diff_expression_data_bin))
48 | #expect_equal(length(res$diff_expression_data_bin),15919)
49 | #expect_equal(sum(grepl("^X",names(res$diff_expression_data_bin))),15919)
50 | expect_true(all(res$diff_expression_data_bin %in% c(-1,0,1)))
51 | })
52 |
53 |
54 | test_that("test cosmos preprocessing (signaling to metabolism)", {
55 |
56 |
57 | meta_network_test <- cosmosR:::meta_network_test
58 | signaling_input_test <- cosmosR:::signaling_input_test
59 | expression_data_test <- cosmosR:::expression_data_test
60 | metabolic_data_test <- cosmosR:::metabolic_data_test
61 |
62 | res <- preprocess_COSMOS_metabolism_to_signaling(meta_network = meta_network_test,
63 | signaling_data = signaling_input_test,
64 | metabolic_data = metabolic_data_test,
65 | diff_expression_data = expression_data_test,
66 | filter_tf_gene_interaction_by_optimization = FALSE)
67 | # check list:
68 | expect_length(res, 10)
69 | expect_true(all(c("meta_network",
70 | "tf_regulon",
71 | "signaling_data",
72 | "signaling_data_bin",
73 | "metabolic_data",
74 | "metabolic_data_bin",
75 | "expression_data",
76 | "diff_expression_data_bin",
77 | "optimized_network") %in% names(res)))
78 |
79 |
80 | # checking network
81 | expect_equal(ncol(res$meta_network), 3)
82 | #expect_equal(nrow(res$meta_network), 42103)
83 | expect_true(all(colnames(res$meta_network) %in% c("source","interaction","target")))
84 |
85 | # checking tf_regulon
86 | expect_true(all(c("tf","target","sign") %in% colnames(res$tf_regulon)))
87 |
88 | # checking signaling
89 | expect_true(is.vector(res$signaling_data_bin))
90 | #expect_equal(length(res$signaling_data_bin),107)
91 | #expect_equal(sum(grepl("^X",names(res$signaling_data_bin))),107)
92 | expect_true(all(res$signaling_data_bin %in% c(-1,0,1)))
93 |
94 | # checking metabolic
95 | expect_true(is.vector(res$metabolic_data))
96 | #expect_equal(length(res$metabolic_data),29)
97 | #expect_equal(sum(grepl("^XMetab",names(res$metabolic_data))),29)
98 |
99 | # check diff_expression_data_bin
100 | expect_true(is.vector(res$diff_expression_data_bin))
101 | #expect_equal(length(res$diff_expression_data_bin),15919)
102 | #expect_equal(sum(grepl("^X",names(res$diff_expression_data_bin))),15919)
103 | expect_true(all(res$diff_expression_data_bin %in% c(-1,0,1)))
104 | })
105 |
--------------------------------------------------------------------------------
/tests/testthat/test-run_COSMOS_metabolism_to_signaling.R:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 | test_that("test run COSMOS signaling to metabolism", {
5 |
6 | meta_network <- cosmosR:::meta_network_test
7 | signaling_data <- cosmosR:::signaling_input_test
8 | expression_data <- cosmosR:::expression_data_test
9 | metabolic_data <- cosmosR:::metabolic_data_test
10 |
11 |
12 | res <- preprocess_COSMOS_metabolism_to_signaling(signaling_data = signaling_data,
13 | meta_network = meta_network,
14 | metabolic_data = metabolic_data,
15 | diff_expression_data = expression_data,
16 | remove_unexpressed_nodes = FALSE,
17 | maximum_network_depth = 15,
18 | filter_tf_gene_interaction_by_optimization = FALSE)
19 |
20 | CARNIVAL_options = CARNIVAL::defaultLpSolveCarnivalOptions()
21 |
22 | res_network = run_COSMOS_metabolism_to_signaling(data = res,
23 | CARNIVAL_options = CARNIVAL_options)
24 |
25 | expect_length(res_network, 5)
26 | expect_true(all(c("weightedSIF",
27 | "N_networks",
28 | "nodesAttributes",
29 | "individual_networks",
30 | "individual_networks_node_attributes") %in% names(res_network)))
31 |
32 | })
33 |
--------------------------------------------------------------------------------
/tests/testthat/test-run_COSMOS_signaling_to_metabolism.R:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 | test_that("test run COSMOS signaling to metabolism", {
5 |
6 |
7 | meta_network <- cosmosR:::meta_network_test
8 | signaling_data <- cosmosR:::signaling_input_test
9 | expression_data <- cosmosR:::expression_data_test
10 | metabolic_data <- cosmosR:::metabolic_data_test
11 |
12 |
13 | res <- preprocess_COSMOS_signaling_to_metabolism(signaling_data = signaling_data,
14 | meta_network = meta_network,
15 | metabolic_data = metabolic_data,
16 | diff_expression_data = expression_data,
17 | remove_unexpressed_nodes = FALSE,
18 | maximum_network_depth = 15,
19 | filter_tf_gene_interaction_by_optimization = FALSE)
20 |
21 | CARNIVAL_options = CARNIVAL::defaultLpSolveCarnivalOptions()
22 |
23 |
24 | #cplex_file <- "/Applications/CPLEX_Studio128/cplex/bin/x86-64_osx/cplex"
25 | #skip_if(!file.exists(cplex_file),"CPLEX optimization based test skipped.")
26 |
27 | #CARNIVAL_options$solverPath = cplex_file
28 | res_network = run_COSMOS_signaling_to_metabolism(data = res,
29 | CARNIVAL_options = CARNIVAL_options)
30 |
31 |
32 | expect_length(res_network, 5)
33 | expect_true(all(c("weightedSIF",
34 | "N_networks",
35 | "nodesAttributes",
36 | "individual_networks",
37 | "individual_networks_node_attributes") %in% names(res_network)))
38 |
39 | })
40 |
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Activity_estimations_for_factor_4_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Activity_estimations_for_factor_4_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Activity_estimations_for_factor_4_2.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Activity_estimations_for_factor_4_2.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_2.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_2.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_3.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Chekc_cross_correlation_of_omics_compared_to_variance_epxlained_3.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Compare_MOFA_models_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Compare_MOFA_models_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Factor_weights_per_view_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Factor_weights_per_view_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Plot_Ligand_receptor_activity_estimates_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Plot_Ligand_receptor_activity_estimates_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Plot_Metabolite_factor_weights_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Plot_Metabolite_factor_weights_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Plot_RNA_weights_and_protein_weights_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Plot_RNA_weights_and_protein_weights_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Plot_RNA_weights_and_protein_weights_2.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Plot_RNA_weights_and_protein_weights_2.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Plot_TF_activity_estimates_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Plot_TF_activity_estimates_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Preparing_MOFA_input_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Preparing_MOFA_input_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Preparing_MOFA_input_2.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Preparing_MOFA_input_2.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Preparing_MOFA_input_3.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Preparing_MOFA_input_3.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Preparing_MOFA_input_4.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Preparing_MOFA_input_4.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Preparing_MOFA_input_5.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Preparing_MOFA_input_5.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Run_moon_rec_to_TFmetab_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Run_moon_rec_to_TFmetab_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Supp_figure_condition_prot_RNA_correlation_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Supp_figure_condition_prot_RNA_correlation_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Supp_figure_single_prot_RNA_correlation_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Supp_figure_single_prot_RNA_correlation_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Total_variance_per_factor_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Total_variance_per_factor_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/Variance_per_view_per_factor_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/Variance_per_view_per_factor_1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/correlation_RNA/prot-1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/correlation_RNA/prot-1.png
--------------------------------------------------------------------------------
/vignettes/M2Cf/figs/run_moon_TF_to_lig_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/M2Cf/figs/run_moon_TF_to_lig_1.png
--------------------------------------------------------------------------------
/vignettes/MOFA_to_COSMOS.Rmd:
--------------------------------------------------------------------------------
1 | ---
2 | title: "MOFA to COSMOS tutorial"
3 | output: rmarkdown::html_vignette
4 | resource_files:
5 | - MOFA_to_COSMOS_files
6 | vignette: >
7 | %\VignetteIndexEntry{MOFA to COSMOS tutorial}
8 | %\VignetteEngine{knitr::rmarkdown}
9 | %\VignetteEncoding{UTF-8}
10 | ---
11 |
12 | ```{r, include = FALSE}
13 | knitr::opts_chunk$set(
14 | collapse = TRUE,
15 | comment = "#>"
16 | )
17 | ```
18 |
19 | ```{r, results='asis',echo=FALSE}
20 | #install.packages("htmltools")
21 | #install.packages("markdown")
22 | # Include the external HTML file
23 | htmltools::includeMarkdown("MOFA_to_COSMOS.md")
24 | ```
25 |
26 |
--------------------------------------------------------------------------------
/vignettes/NCI60_tutorial.Rmd:
--------------------------------------------------------------------------------
1 | ---
2 | title: "NCI60 tutorial"
3 | output: rmarkdown::html_vignette
4 | resource_files:
5 | - net_compr_MOON_files
6 | vignette: >
7 | %\VignetteIndexEntry{NCI60 tutorial}
8 | %\VignetteEngine{knitr::rmarkdown}
9 | %\VignetteEncoding{UTF-8}
10 | ---
11 |
12 | ```{r, include = FALSE}
13 | knitr::opts_chunk$set(
14 | collapse = TRUE,
15 | comment = "#>"
16 | )
17 | ```
18 |
19 | ```{r, results='asis',echo=FALSE}
20 | # htmltools::tags$iframe(
21 | # src = base64enc::dataURI(file="../pkgdown/assets/NCI.html", mime="text/html; charset=UTF-8"),
22 | # style="border:0; position:relative; top:0; left:0; right:0; bottom:0; width:100%; height:100vh"
23 | # )
24 | #
25 | htmltools::includeMarkdown("net_compr_MOON.md")
26 | ```
27 |
--------------------------------------------------------------------------------
/vignettes/net_compr_MOON.md:
--------------------------------------------------------------------------------
1 |
2 | ## libraries, data loading and feature selection
3 |
4 | # We advise to instal from github to get the latest version of the tool.
5 | # if (!requireNamespace("devtools", quietly = TRUE))
6 | # install.packages("devtools")
7 | #
8 | # devtools::install_github("saezlab/cosmosR")
9 |
10 | library(cosmosR)
11 | library(reshape2)
12 | library(readr)
13 |
14 | data("meta_network")
15 |
16 | meta_network <- meta_network_cleanup(meta_network)
17 |
18 | ## Warning: `summarise_each()` was deprecated in dplyr 0.7.0.
19 | ## ℹ Please use `across()` instead.
20 | ## ℹ The deprecated feature was likely used in the cosmosR package.
21 | ## Please report the issue at .
22 | ## This warning is displayed once every 8 hours.
23 | ## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
24 | ## generated.
25 |
26 | ## Warning: `funs()` was deprecated in dplyr 0.8.0.
27 | ## ℹ Please use a list of either functions or lambdas:
28 | ##
29 | ## # Simple named list: list(mean = mean, median = median)
30 | ##
31 | ## # Auto named with `tibble::lst()`: tibble::lst(mean, median)
32 | ##
33 | ## # Using lambdas list(~ mean(., trim = .2), ~ median(., na.rm = TRUE))
34 | ## ℹ The deprecated feature was likely used in the cosmosR package.
35 | ## Please report the issue at .
36 | ## This warning is displayed once every 8 hours.
37 | ## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
38 | ## generated.
39 |
40 | load("data/cosmos/cosmos_inputs.RData")
41 |
42 | names(cosmos_inputs)
43 |
44 | ## [1] "786-0" "A498" "A549/ATCC" "ACHN" "BT-549"
45 | ## [6] "CAKI-1" "CCRF-CEM" "COLO 205" "DU-145" "EKVX"
46 | ## [11] "HCC-2998" "HCT-116" "HCT-15" "HL-60(TB)" "HOP-62"
47 | ## [16] "HOP-92" "HS 578T" "HT29" "IGROV1" "K-562"
48 | ## [21] "KM12" "LOX IMVI" "M14" "MALME-3M" "MCF7"
49 | ## [26] "MDA-MB-435" "MOLT-4" "NCI-H226" "NCI-H23" "NCI-H322M"
50 | ## [31] "NCI-H460" "NCI-H522" "NCI/ADR-RES" "OVCAR-3" "OVCAR-4"
51 | ## [36] "OVCAR-5" "OVCAR-8" "PC-3" "RPMI-8226" "SF-268"
52 | ## [41] "SF-295" "SF-539" "SK-MEL-28" "SK-MEL-5" "SK-OV-3"
53 | ## [46] "SN12C" "SNB-19" "SNB-75" "SR" "SW-620"
54 | ## [51] "T-47D" "TK-10" "U251" "UACC-257" "UACC-62"
55 | ## [56] "UO-31"
56 |
57 | cell_line <- "786-0"
58 |
59 | #see scripts/prepare_cosmos_inputs.R
60 | sig_input <- cosmos_inputs[[cell_line]]$TF_scores
61 | metab_input <- cosmos_inputs[[cell_line]]$metabolomic
62 | RNA_input <- cosmos_inputs[[cell_line]]$RNA
63 |
64 | #Choose which compartment to assign to the metabolic measurments
65 | metab_input <- prepare_metab_inputs(metab_input, c("c","m"))
66 |
67 | ## [1] "Adding compartment codes."
68 |
69 | ##Filter significant inputs
70 | sig_input <- sig_input[abs(sig_input) > 2]
71 | # metab_input <- metab_input[abs(metab_input) > 2]
72 |
73 | ## Filter inputs and prior knowledge network
74 |
75 | #Remove genes that are not expressed from the meta_network
76 | meta_network <- cosmosR:::filter_pkn_expressed_genes(names(RNA_input), meta_pkn = meta_network)
77 |
78 | ## [1] "COSMOS: removing unexpressed nodes from PKN..."
79 | ## [1] "COSMOS: 20357 interactions removed"
80 |
81 | #Filter inputs and prune the meta_network to only keep nodes that can be found downstream of the inputs
82 | #The number of step is quite flexible, 7 steps already covers most of the network
83 |
84 | n_steps <- 6
85 |
86 | # in this step we prune the network to keep only the relevant part between upstream and downstream nodes
87 | sig_input <- cosmosR:::filter_input_nodes_not_in_pkn(sig_input, meta_network)
88 |
89 | ## [1] "COSMOS: 12 input/measured nodes are not in PKN any more: CEBPA, ESR1, FOS, FOXA1, GATA3, HNF4A and 6 more."
90 |
91 | meta_network <- cosmosR:::keep_controllable_neighbours(meta_network, n_steps, names(sig_input))
92 |
93 | ## [1] "COSMOS: removing nodes that are not reachable from inputs within 6 steps"
94 | ## [1] "COSMOS: 26540 from 37740 interactions are removed from the PKN"
95 |
96 | metab_input <- cosmosR:::filter_input_nodes_not_in_pkn(metab_input, meta_network)
97 |
98 | ## [1] "COSMOS: 195 input/measured nodes are not in PKN any more: Metab__HMDB0011747_c, Metab__HMDB0000755_c, Metab__HMDB0000905_c, Metab__HMDB0001191_c, Metab__HMDB0000355_c, Metab__HMDB0000479_c and 189 more."
99 |
100 | meta_network <- cosmosR:::keep_observable_neighbours(meta_network, n_steps, names(metab_input))
101 |
102 | ## [1] "COSMOS: removing nodes that are not observable by measurements within 6 steps"
103 | ## [1] "COSMOS: 3657 from 11200 interactions are removed from the PKN"
104 |
105 | sig_input <- cosmosR:::filter_input_nodes_not_in_pkn(sig_input, meta_network)
106 |
107 | ## [1] "COSMOS: 4 input/measured nodes are not in PKN any more: CTCF, EPAS1, ETS1, USF1 and 0 more."
108 |
109 | #compress the network
110 | meta_network_compressed_list <- compress_same_children(meta_network, sig_input = sig_input, metab_input = metab_input)
111 |
112 | meta_network_compressed <- meta_network_compressed_list$compressed_network
113 |
114 | node_signatures <- meta_network_compressed_list$node_signatures
115 |
116 | duplicated_parents <- meta_network_compressed_list$duplicated_signatures
117 |
118 | meta_network_compressed <- meta_network_cleanup(meta_network_compressed)
119 |
120 | ## run MOON ot score the and contextualise the PKN
121 |
122 | load("support/dorothea_reg.RData")
123 |
124 | meta_network_TF_to_metab <- meta_network_compressed
125 |
126 | before <- 1
127 | after <- 0
128 | i <- 1
129 | while (before != after & i < 10) {
130 | before <- length(meta_network_TF_to_metab[,1])
131 | moon_res <- moon(upstream_input = sig_input,
132 | downstream_input = metab_input,
133 | meta_network = meta_network_TF_to_metab,
134 | n_layers = n_steps,
135 | statistic = "ulm")
136 |
137 | meta_network_TF_to_metab <- filter_incohrent_TF_target(moon_res, dorothea_reg, meta_network_TF_to_metab, RNA_input)
138 | after <- length(meta_network_TF_to_metab[,1])
139 | i <- i + 1
140 | }
141 |
142 | ## [1] 2
143 | ## [1] 3
144 | ## [1] 4
145 | ## [1] 5
146 | ## [1] 6
147 | ## [1] 2
148 | ## [1] 3
149 | ## [1] 4
150 | ## [1] 5
151 | ## [1] 6
152 | ## [1] 2
153 | ## [1] 3
154 | ## [1] 4
155 | ## [1] 5
156 | ## [1] 6
157 |
158 | if(i < 10)
159 | {
160 | print(paste("Converged after ",paste(i-1," iterations", sep = ""),sep = ""))
161 | } else
162 | {
163 | print(paste("Interupted after ",paste(i," iterations. Convergence uncertain.", sep = ""),sep = ""))
164 | }
165 |
166 | ## [1] "Converged after 3 iterations"
167 |
168 | #####
169 | write_csv(moon_res, file = paste("results/moon/",paste(cell_line, "_ATT_decouplerino_full.csv",sep = ""), sep = ""))
170 |
171 | source("scripts/support_decompression.R")
172 | moon_res <- decompress_moon_result(moon_res, meta_network_compressed_list, meta_network_TF_to_metab)
173 |
174 | plot(density(moon_res$score))
175 | abline(v = 1)
176 | abline(v = -1)
177 |
178 | 
179 |
180 | solution_network <- reduce_solution_network(decoupleRnival_res = moon_res,
181 | meta_network = meta_network,
182 | cutoff = 0.5,
183 | upstream_input = sig_input,
184 | RNA_input = RNA_input,
185 | n_steps = n_steps)
186 |
187 | ## [1] "COSMOS: removing nodes that are not reachable from inputs within 6 steps"
188 | ## [1] "COSMOS: 458 from 1747 interactions are removed from the PKN"
189 |
190 | SIF <- solution_network$SIF
191 | names(SIF)[3] <- "sign"
192 | ATT <- solution_network$ATT
193 |
194 | data("HMDB_mapper_vec")
195 |
196 | translated_res <- translate_res(SIF,ATT,HMDB_mapper_vec)
197 |
198 | SIF <- translated_res[[1]]
199 | ATT <- translated_res[[2]]
200 |
201 | write_csv(SIF, file = paste("results/",paste(cell_line, "_dec_compressed_SIF.csv",sep = ""), sep = ""))
202 | write_csv(ATT, file = paste("results/",paste(cell_line, "_dec_compressed_ATT.csv",sep = ""), sep = ""))
203 |
204 | sessionInfo()
205 |
206 | ## R version 4.2.0 (2022-04-22)
207 | ## Platform: aarch64-apple-darwin20 (64-bit)
208 | ## Running under: macOS Monterey 12.6
209 | ##
210 | ## Matrix products: default
211 | ## BLAS: /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/lib/libRblas.0.dylib
212 | ## LAPACK: /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/lib/libRlapack.dylib
213 | ##
214 | ## locale:
215 | ## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
216 | ##
217 | ## attached base packages:
218 | ## [1] stats graphics grDevices utils datasets methods base
219 | ##
220 | ## other attached packages:
221 | ## [1] readr_2.1.4 reshape2_1.4.4 cosmosR_1.5.2
222 | ##
223 | ## loaded via a namespace (and not attached):
224 | ## [1] Rcpp_1.0.10 highr_0.10 plyr_1.8.8 pillar_1.9.0
225 | ## [5] compiler_4.2.0 prettyunits_1.1.1 tools_4.2.0 progress_1.2.2
226 | ## [9] bit_4.0.5 digest_0.6.31 evaluate_0.20 lifecycle_1.0.3
227 | ## [13] tibble_3.2.1 lattice_0.20-45 pkgconfig_2.0.3 rlang_1.1.0
228 | ## [17] igraph_1.4.2 Matrix_1.5-3 cli_3.6.1 rstudioapi_0.14
229 | ## [21] yaml_2.3.7 parallel_4.2.0 xfun_0.42 fastmap_1.1.1
230 | ## [25] withr_2.5.0 dplyr_1.1.3 stringr_1.5.0 knitr_1.42
231 | ## [29] generics_0.1.3 vctrs_0.6.1 hms_1.1.3 bit64_4.0.5
232 | ## [33] grid_4.2.0 tidyselect_1.2.0 glue_1.6.2 R6_2.5.1
233 | ## [37] fansi_1.0.4 parallelly_1.34.0 vroom_1.6.1 rmarkdown_2.21
234 | ## [41] tzdb_0.3.0 tidyr_1.3.0 purrr_1.0.1 decoupleR_2.9.1
235 | ## [45] magrittr_2.0.3 htmltools_0.5.5 utf8_1.2.3 stringi_1.7.12
236 | ## [49] crayon_1.5.2
237 |
--------------------------------------------------------------------------------
/vignettes/net_compr_MOON_files/figure-markdown_strict/extract_subnetwork_from_scored_MOON_network_1.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/saezlab/cosmosR/ef9d5aacc056c79c730382974a98b7d04e62c69d/vignettes/net_compr_MOON_files/figure-markdown_strict/extract_subnetwork_from_scored_MOON_network_1.png
--------------------------------------------------------------------------------