├── .gitignore ├── DESCRIPTION ├── NAMESPACE ├── README.md ├── data ├── qa_all.rda └── stats0.rda ├── inst ├── doc │ └── hg19_alias.tab ├── extdata │ ├── ALLphenoData.tsv │ ├── NCI60.Rda │ ├── nci.data │ └── nci.info └── script │ └── creatingNCI60.R └── vignettes ├── A01.1_IntroductionToR.Rmd ├── A01.2_IntroductionToBioconductor.Rmd ├── A01.3_BioconductorForSequenceAnalysis.Rmd ├── A01_Introduction.Rmd ├── B02.1.1_RNASeqLab.Rmd ├── B02.1_RNASeq.Rmd ├── B02.2.3_CopyNumber.Rmd ├── B02.2_CommonWorkFlows.Rmd ├── B02.3_MachineLearning.Rmd ├── B02.4_Annotation.Rmd ├── B02.5_GeneSetEnrichment.Rmd ├── C03.1_LargeData.Rmd ├── C03.2_CodeToPackages.Rmd ├── C03.3_ReproducibleResearch.Rmd ├── C03.4_Visualization.Rmd ├── D04.1_InstallIGV.Rmd ├── hg19_alias.tab └── our_figures ├── GRanges.png ├── GRangesList.png ├── RangeOperations.png ├── SequencingEcosystem.png ├── SequencingEcosystem_no_bioc_pkgs.png ├── Solexa-bridge-pcr.jpg ├── SummarizedExperiment.png ├── copy_number_QC_2.png ├── cross_validation.png ├── knn.png └── nrg2825-f2.jpg /.gitignore: -------------------------------------------------------------------------------- 1 | Agenda.docx 2 | bigdata 3 | vignettes/*html 4 | vignettes/*R 5 | vignettes/*.md 6 | vignettes/figure 7 | vignettes/cache 8 | -------------------------------------------------------------------------------- /DESCRIPTION: -------------------------------------------------------------------------------- 1 | Package: LearnBioconductor 2 | Type: Package 3 | Title: Learning R / Bioconductor for Sequence Analysis 4 | Version: 0.1.6 5 | Authors@R: c(person("Martin", "Morgan", role=c("aut", "cre"), 6 | email="mtmorgan@fhcrc.org"), 7 | person("Sonali", "Arora", role="aut")) 8 | License: Artistic-2.0 9 | VignetteBuilder: knitr 10 | Description: This course is directed at beginning and intermediate 11 | students who would like an introduction to the analysis and 12 | comprehension of high-throughput sequence data using R and 13 | Bioconductor. Day 1 focuses on learning essential background: an 14 | introduction to the R programming language; central concepts for 15 | effective use of Bioconductor software; and an overview of 16 | high-throughput sequence analysis work flows. Day 2 emphasizes use 17 | of Bioconductor for specific tasks: an RNA-seq differential 18 | expression work flow; exploratory, machine learning, and other 19 | statistical tasks; gene set enrichment; and annotation. Day 3 20 | transitions to understanding effective approaches for managing 21 | larger challenges: strategies for working with large data, writing 22 | re-usable functions, developing reproducible reports and work 23 | flows, and visualizing results. The course combines lectures with 24 | extensive hands-on practicals; students are required to bring a 25 | laptop with wireless internet access and a modern version of the 26 | Chrome or Safari web browser. 27 | Packaged: 2014-10-27 01:31:35 UTC; dante 28 | Author: Martin Morgan [aut, cre], 29 | Sonali Arora [aut] 30 | Maintainer: Martin Morgan 31 | Suggests: knitr, BiocStyle, BiocInstaller, ALL, 32 | BSgenome.Hsapiens.UCSC.hg19, BiocParallel, Biostrings, 33 | GenomicAlignments, GenomicFeatures, Gviz, MLSeq, PoiClaClu, 34 | RColorBrewer, RNAseqData.HNRNPC.bam.chr14, Rsamtools, ShortRead, 35 | TxDb.Hsapiens.UCSC.hg19.knownGene, VariantAnnotation, airway, 36 | class, cn.mops, dendextend, fission, genefilter, ggplot2, gplots, 37 | org.Hs.eg.db, sva, xtable, PoiClaClu, sva, fission, kernlab, 38 | e1071 39 | -------------------------------------------------------------------------------- /NAMESPACE: -------------------------------------------------------------------------------- 1 | exportPattern("^[^\\.]") 2 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | Learning R / Bioconductor for Sequence Analysis 2 | =============================================== 3 | 4 | Fred Hutchinson Cancer Research Center, Seattle, WA
5 | October 27-29 6 | 7 | Location: October 27 / 28: Arnold Building M1-A303; October 29: 8 | Thomas D1-080 9 | 10 | Contact: Martin Morgan (content, 11 | [mtmorgan@fhcrc.org](mailto:mtmorgan@fhcrc.org)); Melissa Alvendia 12 | (administration, [malvendia@fhcrc.org](mailto:malvendia@fhcrc.org)) 13 | 14 | This course is directed at beginning and intermediate users who would 15 | like an introduction to the analysis and comprehension of 16 | high-throughput sequence data using _R_ and 17 | _[Bioconductor](http://bioconductor.org)_. Day 1 focuses on learning 18 | essential background: an introduction to the _R_ programming language; 19 | central concepts for effective use of _Bioconductor_ software; and an 20 | overview of high-throughput sequence analysis work flows. Day 2 21 | emphasizes use of _Bioconductor_ for specific tasks: an RNA-seq 22 | differential expression work flow; exploratory, machine learning, and 23 | other statistical tasks; gene set enrichment; and annotation. Day 3 24 | transitions to understanding effective approaches for managing larger 25 | challenges: strategies for working with large data, writing re-usable 26 | functions, developing reproducible reports and work flows, and 27 | visualizing results. The course combines lectures with extensive 28 | hands-on practicals; students are required to bring a laptop with 29 | wireless internet access and a modern version of the Chrome or Safari 30 | web browser. 31 | 32 | Schedule (tentative) 33 | -------------------- 34 | 35 | Day 1: Learn _R_ / _Bioconductor_ 36 | 37 | - 9:00 - 10:30 Introduction to _R_: objects, functions, help! 38 | - 11:00 - 12:30 Introduction to _Bioconductor_: working with packages and classes 39 | - 1:30 - 5:00 (break: 3:00 - 3:30) Introduction to sequence analysis: 40 | typical work flow; data types and quality assessment; essential 41 | _Bioconductor_ packages 42 | 43 | Day 2: Use _R_ / _Bioconductor_ 44 | 45 | - 9:00 - 12:30 (break (10:30 - 11:00) An RNA-seq differential 46 | expression work flow (detail) 47 | - 1:30 - 2:00 Other work flows (survey): ChIP-seq, variants, copy 48 | number, epigenomics 49 | - 2:00 - 3:00 Machine learning; exploratory and other statistical 50 | analysis 51 | - 3:30 - 4:00 Annotating genes, genomes, and variants 52 | - 4:00 - 5:00 Approaches to gene set enrichment 53 | 54 | Day 3: Develop Skills and Best Practices 55 | 56 | - 9:00 - 10:30 Working with large data 57 | - 11:00 - 12:30 Organizing code in functions, files, and packages 58 | - 1:30 - 3:00 Reproducible reports and work flows: markdown 59 | - 3:30 - 4:30 Visualization 60 | - 4:30 - 5:00 Summary 61 | -------------------------------------------------------------------------------- /data/qa_all.rda: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/data/qa_all.rda -------------------------------------------------------------------------------- /data/stats0.rda: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/data/stats0.rda -------------------------------------------------------------------------------- /inst/doc/hg19_alias.tab: -------------------------------------------------------------------------------- 1 | gi|224384768|gb|CM000663.1| chr1 2 | gi|224384767|gb|CM000664.1| chr2 3 | gi|224384766|gb|CM000665.1| chr3 4 | gi|224384765|gb|CM000666.1| chr4 5 | gi|224384764|gb|CM000667.1| chr5 6 | gi|224384763|gb|CM000668.1| chr6 7 | gi|224384762|gb|CM000669.1| chr7 8 | gi|224384761|gb|CM000670.1| chr8 9 | gi|224384760|gb|CM000671.1| chr9 10 | gi|224384759|gb|CM000672.1| chr10 11 | gi|224384758|gb|CM000673.1| chr11 12 | gi|224384757|gb|CM000674.1| chr12 13 | gi|224384756|gb|CM000675.1| chr13 14 | gi|224384755|gb|CM000676.1| chr14 15 | gi|224384754|gb|CM000677.1| chr15 16 | gi|224384753|gb|CM000678.1| chr16 17 | gi|224384752|gb|CM000679.1| chr17 18 | gi|224384751|gb|CM000680.1| chr18 19 | gi|224384750|gb|CM000681.1| chr19 20 | gi|224384749|gb|CM000682.1| chr20 21 | gi|224384748|gb|CM000683.1| chr21 22 | gi|224384747|gb|CM000684.1| chr22 23 | -------------------------------------------------------------------------------- /inst/extdata/ALLphenoData.tsv: -------------------------------------------------------------------------------- 1 | id diagnosis sex age BT remission CR date.cr t(4;11) t(9;22) cyto.normal citog mol.biol fusion protein mdr kinet ccr relapse transplant f.u date last seen 2 | 1005 5/21/1997 M 53 B2 CR CR 8/6/1997 FALSE TRUE FALSE t(9;22) BCR/ABL p210 NEG dyploid FALSE FALSE TRUE BMT / DEATH IN CR NA 3 | 1010 3/29/2000 M 19 B2 CR CR 6/27/2000 FALSE FALSE FALSE simple alt. NEG NA POS dyploid FALSE TRUE FALSE REL 8/28/2000 4 | 3002 6/24/1998 F 52 B4 CR CR 8/17/1998 NA NA NA NA BCR/ABL p190 NEG dyploid FALSE TRUE FALSE REL 10/15/1999 5 | 4006 7/17/1997 M 38 B1 CR CR 9/8/1997 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 1/23/1998 6 | 4007 7/22/1997 M 57 B2 CR CR 9/17/1997 FALSE FALSE FALSE del(6q) NEG NA NEG dyploid FALSE TRUE FALSE REL 11/4/1997 7 | 4008 7/30/1997 M 17 B1 CR CR 9/27/1997 FALSE FALSE FALSE complex alt. NEG NA NEG hyperd. FALSE TRUE FALSE REL 12/15/1997 8 | 4010 10/30/1997 F 18 B1 CR CR 1/7/1998 FALSE FALSE FALSE complex alt. NEG NA POS hyperd. FALSE TRUE FALSE REL 3/5/1998 9 | 4016 2/10/2000 M 16 B1 CR CR 4/17/2000 FALSE FALSE FALSE simple alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 9/26/2000 10 | 6002 3/19/1997 M 15 B2 CR CR 6/9/1997 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 3/18/1998 11 | 8001 1/15/1997 M 40 B2 CR CR 3/26/1997 FALSE FALSE FALSE del(p15) BCR/ABL p190 NEG NA FALSE TRUE FALSE REL 7/11/1997 12 | 8011 8/21/1998 M 33 B3 CR CR 10/8/1998 FALSE FALSE FALSE del(p15/p16) BCR/ABL p190/p210 NEG dyploid FALSE FALSE TRUE BMT / DEATH IN CR NA 13 | 8012 10/22/1998 M 55 B3 CR CR 1/9/1999 FALSE FALSE FALSE simple alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 4/9/1999 14 | 8018 8/27/1999 M 5 B3 CR CR 10/18/1999 NA NA NA NA E2A/PBX1 NA NEG dyploid FALSE TRUE FALSE REL 5/23/2000 15 | 8024 7/20/2000 M 18 B2 CR DEATH IN CR NA FALSE FALSE FALSE simple alt. NEG NA POS dyploid NA NA NA NA NA 16 | 9008 12/17/1999 M 41 B3 CR CR 2/15/2000 FALSE TRUE FALSE t(9;22)+other BCR/ABL p190 NEG hyperd. TRUE FALSE TRUE BMT / CCR 00/09/01 17 | 9017 2/3/2000 F 27 B CR CR 3/23/2000 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 9/11/2001 18 | 11005 6/1/1998 M 27 B2 CR DEATH IN CR 8/3/1998 FALSE FALSE FALSE del(7q) + altro BCR/ABL p190 NEG dyploid FALSE FALSE FALSE DEATH IN CR NA 19 | 12006 2/20/1997 M 46 B3 REF REF NA FALSE TRUE FALSE t(9;22) BCR/ABL p210 NEG dyploid NA NA NA NA NA 20 | 12007 3/11/1997 M 37 B2 CR CR 5/7/1997 FALSE TRUE FALSE t(9;22)+ del(9) BCR/ABL NA NEG dyploid FALSE TRUE FALSE REL 9/26/1997 21 | 12012 5/21/1997 F 36 B3 REF REF NA FALSE TRUE FALSE t(9;22) BCR/ABL p190 NEG dyploid NA NA NA NA NA 22 | 12019 9/4/1997 M 53 B2 CR CR 11/11/1997 FALSE FALSE TRUE normal NEG NA POS dyploid TRUE FALSE FALSE CCR 6/6/2002 23 | 12026 5/29/1998 M 39 B2 REF REF NA FALSE TRUE FALSE t(9;22) BCR/ABL p190/p210 NA dyploid FALSE FALSE FALSE DEATH IN CR NA 24 | 14016 5/27/1999 M 53 B2 NA NA NA FALSE TRUE FALSE t(9;22) BCR/ABL p210 NEG NA NA NA NA NA NA 25 | 15001 9/3/1997 M 20 B1 CR CR 11/11/1997 FALSE FALSE TRUE normal NEG NA NEG dyploid TRUE FALSE FALSE CCR 6/21/2002 26 | 15004 2/10/2000 M 44 B1 CR CR 4/3/2000 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 12/19/2000 27 | 15005 5/4/2000 M 28 B2 CR CR 6/26/2000 FALSE TRUE FALSE t(9;22)+other BCR/ABL p190 POS hyperd. FALSE TRUE FALSE REL 11/15/2000 28 | 16004 4/19/1997 F 58 B1 CR CR 7/15/1997 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 12/9/1997 29 | 16009 7/11/2000 F 43 B2 NA NA NA NA NA NA NA NEG NA POS dyploid TRUE FALSE FALSE CCR / OFF 5/23/2002 30 | 19005 11/15/1997 F 48 B1 CR CR 2/3/1998 FALSE FALSE TRUE normal ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 2/4/1998 31 | 20002 5/9/1997 F 58 B2 CR CR 8/19/1997 FALSE TRUE FALSE t(9;22)+other BCR/ABL p190 NEG dyploid FALSE TRUE FALSE REL 12/15/1997 32 | 22009 8/10/1999 F 19 B NA NA NA FALSE FALSE FALSE simple alt. NEG NA NEG dyploid NA NA NA NA NA 33 | 22010 12/31/1999 F 26 B NA NA NA FALSE TRUE FALSE t(9;22) BCR/ABL p190/p210 NEG dyploid NA NA NA NA NA 34 | 22011 4/7/2000 M 19 B2 CR CR 5/19/2000 FALSE FALSE TRUE normal NEG NA NEG dyploid TRUE FALSE FALSE CCR 7/31/2002 35 | 22013 8/17/2000 M 32 B2 REF REF NA FALSE TRUE FALSE t(9;22)+other BCR/ABL p190/p210 NEG dyploid NA NA NA NA NA 36 | 24001 10/4/1996 F 17 B2 CR CR 12/20/1996 NA NA NA NA BCR/ABL p190 NEG dyploid FALSE TRUE FALSE REL 2/10/1997 37 | 24005 1/3/1997 F 45 B1 CR CR 4/8/1997 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 8/28/1997 38 | 24008 5/14/1997 F 20 B2 CR CR 7/31/1997 NA NA NA NA NEG NA NEG dyploid FALSE FALSE TRUE BMT / CCR 00/09/20+T12501 39 | 24010 6/3/1997 F 16 B2 CR CR 8/11/1997 FALSE TRUE FALSE t(9;22) BCR/ABL p190/p210 NEG dyploid FALSE TRUE TRUE BMT / REL 8/24/1998 40 | 24011 8/5/1997 F 51 B2 NA DEATH IN INDUCTION NA FALSE TRUE FALSE t(9;22) BCR/ABL p210 POS dyploid NA NA NA NA NA 41 | 24017 9/15/1998 M 57 B2 CR CR 12/7/1998 FALSE TRUE FALSE t(9;22)+other BCR/ABL p190 NEG hyperd. FALSE TRUE FALSE REL 2/22/2000 42 | 24018 2/18/1999 F 29 B2 CR CR 5/4/1999 NA NA NA NA NEG NA POS dyploid FALSE TRUE FALSE REL 7/22/2000 43 | 24019 5/4/1999 M 16 B4 CR CR 7/28/1999 FALSE FALSE FALSE t(1;19) + del(6q) E2A/PBX1 NA NEG dyploid FALSE TRUE FALSE REL 5/25/2000 44 | 24022 12/21/1999 F 32 B4 REF REF NA FALSE TRUE FALSE t(9;22) BCR/ABL p190 POS dyploid NA NA NA NA NA 45 | 25003 5/22/1998 M 15 B2 CR CR 8/4/1998 FALSE FALSE FALSE simple alt. NEG NA POS dyploid TRUE FALSE FALSE CCR 6/10/2002 46 | 25006 3/18/2000 NA NA B2 CR CR 5/8/2000 NA NA NA NA NEG NA NEG NA TRUE FALSE FALSE CCR 3/3/2002 47 | 26001 9/27/1997 M 21 B2 CR CR 12/11/1997 NA NA NA NA NEG NA POS dyploid TRUE FALSE FALSE CCR 7/31/2002 48 | 26003 2/18/1998 F 49 B4 CR CR 4/21/1998 FALSE FALSE FALSE del(p15/p16) BCR/ABL p210 NEG dyploid FALSE TRUE FALSE REL 7/1/1998 49 | 26005 8/21/1998 M 38 B2 CR CR 11/15/1998 NA NA NA NA p15/p16 NA NEG dyploid FALSE TRUE FALSE REL 3/4/2002 50 | 26008 8/25/1999 F 17 B1 CR CR 10/14/1999 FALSE FALSE TRUE normal ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 6/26/2000 51 | 27003 1/17/1998 F 26 B2 CR CR 3/16/1998 NA NA NA NA BCR/ABL p190/p210 POS dyploid FALSE TRUE FALSE REL 5/6/1998 52 | 27004 10/20/1998 F 48 B2 REF REF NA FALSE TRUE FALSE t(9;22)+del(p15) BCR/ABL p190 NEG dyploid NA NA NA NA NA 53 | 28001 10/19/1996 M 16 B3 REF REF NA FALSE FALSE TRUE normal NEG NA POS dyploid NA NA NA NA NA 54 | 28003 11/28/1996 M 18 B4 CR CR 1/17/1997 NA NA NA NA E2A/PBX1 NA NEG hyperd. TRUE FALSE FALSE CCR 12/31/2002 55 | 28005 12/27/1996 M 17 B3 CR CR 2/26/1997 FALSE FALSE FALSE complex alt. NEG NA POS hyperd. FALSE TRUE FALSE REL 3/6/1998 56 | 28006 1/13/1997 M 22 B3 CR CR 3/14/1997 FALSE FALSE FALSE del(6q) NEG NA NEG dyploid FALSE TRUE FALSE REL 1/16/1998 57 | 28007 2/21/1997 F 47 B3 CR CR 4/7/1997 FALSE FALSE TRUE normal NEG NA NEG dyploid TRUE FALSE FALSE CCR 3/22/2002 58 | 28019 2/10/1998 M 21 B4 CR CR 4/2/1998 NA NA NA NA BCR/ABL p190 NEG hyperd. TRUE FALSE TRUE BMT / CCR 3/21/2001 59 | 28021 3/18/1998 F 54 B3 CR DEATH IN CR 5/22/1998 FALSE TRUE FALSE t(9;22)+other BCR/ABL p190/p210 NEG hyperd. FALSE FALSE FALSE DEATH IN CR (ICR) NA 60 | 28023 2/26/1998 M 26 B3 CR CR 4/20/1998 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL / SNC 2/5/1999 61 | 28024 4/19/1998 F 19 B1 CR CR 6/17/1998 FALSE FALSE FALSE complex alt. NEG NA NEG hyperd. TRUE FALSE FALSE CCR 12/31/2002 62 | 28028 7/8/1998 M 47 B1 CR CR 9/3/1998 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 10/20/1999 63 | 28031 9/23/1998 M 18 B1 CR CR 12/7/1998 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 1/17/1999 64 | 28032 9/26/1998 F 52 B1 CR CR 10/30/1998 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid TRUE FALSE FALSE CCR 5/16/2002 65 | 28035 12/21/1998 M 27 B3 CR CR 2/12/1999 NA NA NA NA NEG NA POS hyperd. TRUE FALSE FALSE CCR 5/20/2002 66 | 28036 12/23/1998 M 52 B3 CR CR 3/8/1999 FALSE TRUE FALSE t(9;22) BCR/ABL p190 NEG dyploid FALSE TRUE FALSE REL 3/15/1999 67 | 28037 12/30/1998 M 18 B3 CR CR 3/4/1999 FALSE FALSE TRUE normal NEG NA POS dyploid FALSE TRUE FALSE REL 4/23/2001 68 | 28042 6/18/1999 M 18 B3 CR CR 8/9/1999 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL 11/30/2002 69 | 28043 6/28/1999 M 23 B3 CR CR 8/23/1999 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL 3/29/2001 70 | 28044 7/20/1999 M 16 B3 CR CR 9/20/1999 FALSE FALSE FALSE complex alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 12/23/1999 71 | 28047 4/17/2000 M NA B3 CR CR NA FALSE FALSE FALSE simple alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 9/4/2000 72 | 30001 1/16/1997 F 54 B3 NA DEATH IN INDUCTION NA FALSE TRUE FALSE t(9;22)+other BCR/ABL p190 NEG hyperd. NA NA NA NA NA 73 | 31007 2/20/1997 M 25 B1 CR CR 5/15/1997 TRUE FALSE FALSE t(4;11) ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 12/3/1998 74 | 31011 11/15/1997 M 31 B3 CR CR 1/21/1998 FALSE TRUE FALSE t(9;22) BCR/ABL p210 POS dyploid FALSE TRUE TRUE AUBMT / REL 6/1/1998 75 | 33005 2/10/1998 F 19 B1 CR CR 4/29/1998 FALSE FALSE FALSE complex alt. NEG NA NEG dyploid TRUE FALSE FALSE CCR 6/28/2002 76 | 36001 9/29/1997 F 24 B4 CR CR 12/5/1997 FALSE FALSE TRUE normal E2A/PBX1 NA NEG dyploid FALSE TRUE FALSE REL 1/7/1998 77 | 36002 4/1/1998 M 23 B2 CR CR 6/9/1998 FALSE FALSE TRUE normal NEG NA NEG hyperd. FALSE TRUE FALSE REL 3/29/2001 78 | 37013 5/28/1998 M NA B2 NA NA NA FALSE FALSE FALSE del(p15/p16) BCR/ABL NA NEG dyploid NA NA NA NA NA 79 | 43001 11/14/1996 M 41 B1 CR CR 1/30/1997 FALSE TRUE FALSE t(9;22) BCR/ABL p190/p210 POS dyploid FALSE TRUE FALSE REL 6/28/1998 80 | 43004 2/4/1997 F 37 B3 CR CR 4/1/1997 NA NA NA NA NEG NA NEG dyploid TRUE FALSE FALSE CCR 3/20/2001 81 | 43007 10/14/1997 M 54 B4 CR CR 12/30/1997 FALSE FALSE TRUE normal NEG NA NEG hyperd. TRUE FALSE FALSE CCR 5/29/2002 82 | 43012 1/15/1999 M 18 B4 CR CR 3/22/1999 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL 4/20/2000 83 | 48001 3/22/1997 M 19 B2 CR CR 5/20/1997 FALSE FALSE FALSE complex alt. NEG NA NEG hyperd. FALSE FALSE FALSE MUD / DEATH IN CR 12/18/1998 84 | 49006 8/12/1998 F 43 B2 CR CR 11/19/1998 NA NA NA NA BCR/ABL p210 NEG dyploid FALSE TRUE FALSE REL 4/26/1999 85 | 57001 1/29/1997 F 53 B3 NA DEATH IN INDUCTION NA FALSE FALSE TRUE normal NEG NA NEG hyperd. NA NA NA NA NA 86 | 62001 11/11/1997 F 50 B4 REF REF NA FALSE TRUE FALSE t(9;22)+other BCR/ABL NA NEG hyperd. NA NA NA NA NA 87 | 62002 1/15/1998 M 54 B4 NA DEATH IN INDUCTION NA FALSE TRUE FALSE t(9;22)+other BCR/ABL NA NEG hyperd. NA NA NA NA NA 88 | 62003 12/4/1998 M 53 B4 CR CR 1/28/1999 FALSE TRUE FALSE t(9;22)+other BCR/ABL p210 NEG hyperd. FALSE TRUE FALSE REL 8/8/2000 89 | 63001 7/8/1997 M 49 B1 CR CR 9/2/1997 NA NA NA NA ALL1/AF4 NA NEG dyploid FALSE TRUE FALSE REL 6/10/1998 90 | 64001 8/28/1997 M 20 B2 CR CR 10/27/1997 FALSE FALSE FALSE simple alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 1/21/1998 91 | 64002 10/21/1997 F 26 B2 CR CR 1/21/1998 NA NA NA NA NEG NA NEG hyperd. FALSE FALSE TRUE BMT / DEATH IN CR NA 92 | 65005 7/20/1999 M 22 B2 REF REF NA FALSE TRUE FALSE t(9;22)+del(p15/p16) BCR/ABL p190 NEG dyploid NA NA NA NA NA 93 | 68001 5/15/1997 M 36 B1 CR CR 7/22/1997 FALSE FALSE TRUE normal NEG NA NEG dyploid TRUE FALSE FALSE CCR 5/10/2002 94 | 68003 4/11/2000 F 27 B2 NA DEATH IN INDUCTION NA FALSE TRUE FALSE t(9;22)+other BCR/ABL p190 NEG NA NA NA NA NA NA 95 | 84004 9/25/1998 M 50 B CR CR 12/1/1998 NA NA NA NA BCR/ABL p190 NEG dyploid FALSE TRUE FALSE REL 1/25/1999 96 | LAL5 NA NA NA B NA NA NA NA NA NA NA E2A/PBX1 NA NA NA NA NA NA NA NA 97 | 1003 2/18/1997 M 31 T CR CR 4/29/1997 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 7/11/1997 98 | 1007 9/30/1998 F 16 T3 CR CR 11/30/1998 FALSE FALSE FALSE simple alt. NUP-98 NA NEG hyperd. FALSE FALSE TRUE BMT / DEATH IN CR NA 99 | 2020 3/23/2000 F 48 T2 NA DEATH IN INDUCTION NA FALSE FALSE FALSE complex alt. NEG NA NEG dyploid NA NA NA NA NA 100 | 4018 3/24/2000 M 17 T2 CR CR 5/23/2000 FALSE FALSE FALSE simple alt. NEG NA NEG dyploid TRUE FALSE FALSE CCR 5/14/2001 101 | 9002 5/14/1998 F 40 T3 CR CR 7/21/1998 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL / SNC 9/14/1999 102 | 10005 9/30/1997 M 22 T2 CR CR 1/26/1998 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL 10/10/1999 103 | 11002 12/23/1996 M 30 T CR CR 3/18/1997 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 4/14/1998 104 | 12008 3/22/1997 M 18 T4 REF REF NA FALSE FALSE FALSE del(6q) NEG NA NEG dyploid NA NA NA NA NA 105 | 15006 6/10/1998 M 22 T2 CR CR 8/11/1998 FALSE FALSE FALSE complex alt. NEG NA NEG hyperd. FALSE TRUE FALSE REL 9/15/2000 106 | 16002 4/10/1997 M 50 T3 CR CR 6/10/1997 NA NA NA NA NEG NA NEG hyperd. FALSE TRUE FALSE REL 12/7/1999 107 | 16007 11/1/1998 M 41 T3 CR CR 11/5/1998 NA NA NA NA NEG NA NEG dyploid TRUE FALSE FALSE CCR 1/8/2002 108 | 17003 4/8/1997 F 40 T REF REF NA NA NA NA NA NEG NA POS dyploid NA NA NA NA NA 109 | 18001 4/23/1997 F 28 T2 REF REF NA NA NA NA NA NEG NA NEG hyperd. NA NA NA NA NA 110 | 19002 1/29/1997 M 25 T3 CR CR 4/10/1997 FALSE FALSE FALSE simple alt. NEG NA NEG hyperd. FALSE TRUE FALSE REL 11/10/1997 111 | 19008 4/29/1998 F 16 T2 REF REF NA FALSE FALSE TRUE normal NEG NA POS dyploid NA NA NA NA NA 112 | 19014 3/24/1999 M 31 T2 CR CR 5/13/1999 FALSE FALSE FALSE simple alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 4/1/2000 113 | 19017 3/17/2000 M 14 T2 CR CR 6/15/2000 FALSE FALSE TRUE normal NEG NA POS dyploid FALSE TRUE FALSE REL 7/5/2001 114 | 20005 3/15/2000 M 24 T1 CR CR 5/5/2000 NA NA NA NA NEG NA NEG dyploid TRUE FALSE FALSE CCR 3/20/2002 115 | 24006 1/14/1997 F 19 T4 CR CR not known FALSE FALSE FALSE simple alt. NEG NA NEG dyploid TRUE FALSE FALSE CCR 6/5/2002 116 | 26009 8/26/1999 M 37 T CR CR 10/18/1999 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 6/12/2000 117 | 28008 3/27/1997 M 23 T2 CR CR 5/27/1997 FALSE FALSE TRUE normal NEG NA NEG hyperd. TRUE FALSE FALSE CCR 4/9/2002 118 | 28009 4/19/1997 F 30 T3 CR CR 6/13/1997 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL 6/30/1998 119 | 31015 12/3/1998 M 48 T2 NA DEATH IN INDUCTION NA NA NA NA NA NEG NA POS dyploid NA NA NA NA NA 120 | 37001 1/30/1997 M 22 T2 CR CR 4/3/1997 NA NA NA NA NEG NA NEG dyploid FALSE TRUE FALSE REL 1/7/1998 121 | 43006 6/17/1997 M 41 T2 REF REF NA FALSE FALSE FALSE simple alt. NEG NA POS dyploid NA NA NA NA NA 122 | 43015 2/29/2000 M 52 T2 CR CR 6/8/2000 FALSE FALSE TRUE normal NEG NA NEG dyploid FALSE TRUE FALSE REL 3/15/2002 123 | 44001 1/15/1997 M 32 T3 CR CR 3/7/1997 FALSE FALSE FALSE del(6q) NEG NA NEG NA FALSE TRUE FALSE REL 11/25/1997 124 | 49004 9/18/1997 M 24 T3 CR CR 11/11/1997 FALSE FALSE FALSE del(7q) NEG NA POS dyploid TRUE FALSE FALSE CCR 6/14/2001 125 | 56007 8/6/1999 M 37 T3 CR CR 9/24/1999 NA NA NA NA NEG NA NEG dyploid TRUE FALSE FALSE CCR 1/26/2001 126 | 64005 10/1/1998 M 19 T2 REF REF NA NA NA NA NA NEG NA NEG dyploid NA NA NA NA NA 127 | 65003 3/27/1998 M 30 T3 CR CR 5/27/1998 FALSE FALSE FALSE simple alt. NEG NA NEG dyploid FALSE TRUE FALSE REL 2/11/1999 128 | 83001 10/23/1998 M 29 T2 CR CR 12/21/1998 FALSE FALSE FALSE complex alt. NEG NA NEG hyperd. TRUE FALSE FALSE CCR 5/24/2002 129 | -------------------------------------------------------------------------------- /inst/extdata/NCI60.Rda: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/inst/extdata/NCI60.Rda -------------------------------------------------------------------------------- /inst/extdata/nci.info: -------------------------------------------------------------------------------- 1 | NCI microarray data (chap 14) 2 | 3 | Source and reference: 4 | 5 | http://genome-www.stanford.edu/nci60/ 6 | 7 | 8 | NCI microarray data 9 | 10 | 6830 genes 11 | missing values have been imputed via SVD 12 | 60 cell lines, labels are below 13 | 14 | 15 | CNS 16 | CNS 17 | CNS 18 | RENAL 19 | BREAST 20 | CNS 21 | CNS 22 | BREAST 23 | NSCLC 24 | NSCLC 25 | RENAL 26 | RENAL 27 | RENAL 28 | RENAL 29 | RENAL 30 | RENAL 31 | RENAL 32 | BREAST 33 | NSCLC 34 | RENAL 35 | UNKNOWN 36 | OVARIAN 37 | MELANOMA 38 | PROSTATE 39 | OVARIAN 40 | OVARIAN 41 | OVARIAN 42 | OVARIAN 43 | OVARIAN 44 | PROSTATE 45 | NSCLC 46 | NSCLC 47 | NSCLC 48 | LEUKEMIA 49 | K562B-repro 50 | K562A-repro 51 | LEUKEMIA 52 | LEUKEMIA 53 | LEUKEMIA 54 | LEUKEMIA 55 | LEUKEMIA 56 | COLON 57 | COLON 58 | COLON 59 | COLON 60 | COLON 61 | COLON 62 | COLON 63 | MCF7A-repro 64 | BREAST 65 | MCF7D-repro 66 | BREAST 67 | NSCLC 68 | NSCLC 69 | NSCLC 70 | MELANOMA 71 | BREAST 72 | BREAST 73 | MELANOMA 74 | MELANOMA 75 | MELANOMA 76 | MELANOMA 77 | MELANOMA 78 | MELANOMA 79 | -------------------------------------------------------------------------------- /inst/script/creatingNCI60.R: -------------------------------------------------------------------------------- 1 | rm(list=ls()) 2 | 3 | datafile <- system.file("extdata", "nci.data", package="LearnBioconductor") 4 | data <- read.table(datafile) 5 | data <- as.matrix(data) 6 | infofile <- system.file("extdata", "nci.info", package="LearnBioconductor") 7 | info <- read.table(infofile, skip=14) 8 | 9 | # drop subtypes that contain only 1 sample 10 | rem <- which(table(info)==1) 11 | rm_ind <- match(names(rem), info[,1]) 12 | info <- info[-rm_ind,] 13 | data <- data[, -rm_ind] 14 | 15 | showMethods("SummarizedExperiment") 16 | selectMethod(SummarizedExperiment, "matrix") 17 | 18 | NCI60 <- SummarizedExperiment(assays=list(data=data), colData=DataFrame(info)) 19 | save(NCI60, file="NCI60.Rda") 20 | -------------------------------------------------------------------------------- /vignettes/A01.1_IntroductionToR.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | knitr::opts_chunk$set(tidy=FALSE) 15 | ``` 16 | 17 | # Introduction to R 18 | 19 | Martin Morgan
20 | October 27, 2014 21 | 22 | ## R 23 | 24 | Language and environment for statistical computing and graphics 25 | 26 | - Full-featured programming language 27 | - Interactive and *interpretted* -- convenient and forgiving 28 | - Coherent, extensive documentation 29 | - Statistical, e.g. `factor()`, `NA` 30 | - Extensible -- CRAN, Bioconductor, github, ... 31 | 32 | Vector, class, object 33 | 34 | - Efficient _vectorized_ calculations on 'atomic' vectors `logical`, 35 | `integer`, `numeric`, `complex`, `character`, `byte` 36 | - Atomic vectors are building blocks for more complicated _objects_ 37 | - `matrix` -- atomic vector with 'dim' attribute 38 | - `data.frame` -- list of equal length atomic vectors 39 | - Formal _classes_ represent complicated combinations of vectors, 40 | e.g., the return value of `lm()`, below 41 | 42 | Function, generic, method 43 | 44 | - Functions transform inputs to outputs, perhaps with side effects, 45 | e.g., `rnorm(1000)` 46 | - Argument matching first by name, then by position 47 | - Functions may define (some) arguments to have default values 48 | - _Generic_ functions dispatch to specific _methods_ based on class of 49 | argument(s), e.g., `print()`. 50 | - Methods are functions that implement specific generics, e.g., 51 | `print.factor`; methods are invoked _indirectly_, via the generic. 52 | 53 | Introspection 54 | 55 | - General properties, e.g., `class()`, `str()` 56 | - Class-specific properties, e.g., `dim()` 57 | 58 | Help 59 | 60 | - `?print`: help on the generic print 61 | - `?print.data.frame`: help on print method for objects of class 62 | data.frame. 63 | 64 | Example 65 | 66 | ```{r} 67 | x <- rnorm(1000) # atomic vectors 68 | y <- x + rnorm(1000, sd=.5) 69 | df <- data.frame(x=x, y=y) # object of class 'data.frame' 70 | plot(y ~ x, df) # generic plot, method plot.formula 71 | fit <- lm(y ~x, df) # object of class 'lm' 72 | methods(class=class(fit)) # introspection 73 | ``` 74 | 75 | ## Lab 76 | 77 | ### 1. _R_ data manipulation 78 | 79 | This exercise servers as a refresher / tutorial on basic input and 80 | manipulation of data. 81 | 82 | Input a file that contains ALL (acute lymphoblastic leukemia) patient 83 | information 84 | 85 | ```{r echo=TRUE, eval=FALSE} 86 | fname <- file.choose() ## "ALLphenoData.tsv" 87 | stopifnot(file.exists(fname)) 88 | pdata <- read.delim(fname) 89 | ``` 90 | ```{r echo=FALSE} 91 | fname <- system.file("extdata", "ALLphenoData.tsv", package="LearnBioconductor") 92 | stopifnot(file.exists(fname)) 93 | pdata <- read.delim(fname) 94 | ``` 95 | 96 | Check out the help page `?read.delim` for input options, and explore 97 | basic properties of the object you've created, for instance... 98 | 99 | ```{r ALL-properties} 100 | class(pdata) 101 | colnames(pdata) 102 | dim(pdata) 103 | head(pdata) 104 | summary(pdata$sex) 105 | summary(pdata$cyto.normal) 106 | ``` 107 | 108 | Remind yourselves about various ways to subset and access columns of a 109 | data.frame 110 | 111 | ```{r ALL-subset} 112 | pdata[1:5, 3:4] 113 | pdata[1:5, ] 114 | head(pdata[, 3:5]) 115 | tail(pdata[, 3:5], 3) 116 | head(pdata$age) 117 | head(pdata$sex) 118 | head(pdata[pdata$age > 21,]) 119 | ``` 120 | 121 | It seems from below that there are 17 females over 40 in the data set, 122 | but when sub-setting `pdata` to contain just those individuals 19 rows 123 | are selected. Why? What can we do to correct this? 124 | 125 | ```{r ALL-subset-NA} 126 | idx <- pdata$sex == "F" & pdata$age > 40 127 | table(idx) 128 | dim(pdata[idx,]) 129 | ``` 130 | 131 | Use the `mol.biol` column to subset the data to contain just 132 | individuals with 'BCR/ABL' or 'NEG', e.g., 133 | 134 | ```{r ALL-BCR/ABL-subset} 135 | bcrabl <- pdata[pdata$mol.biol %in% c("BCR/ABL", "NEG"),] 136 | ``` 137 | 138 | The `mol.biol` column is a factor, and retains all levels even after 139 | subsetting. How might you drop the unused factor levels? 140 | 141 | ```{r ALL-BCR/ABL-drop-unused} 142 | bcrabl$mol.biol <- factor(bcrabl$mol.biol) 143 | ``` 144 | 145 | The `BT` column is a factor describing B- and T-cell subtypes 146 | 147 | ```{r ALL-BT} 148 | levels(bcrabl$BT) 149 | ``` 150 | 151 | How might one collapse B1, B2, ... to a single type B, and likewise 152 | for T1, T2, ..., so there are only two subtypes, B and T 153 | 154 | ```{r ALL-BT-recode} 155 | table(bcrabl$BT) 156 | levels(bcrabl$BT) <- substring(levels(bcrabl$BT), 1, 1) 157 | table(bcrabl$BT) 158 | ``` 159 | 160 | Use `xtabs()` (cross-tabulation) to count the number of samples with 161 | B- and T-cell types in each of the BCR/ABL and NEG groups 162 | 163 | ```{r ALL-BCR/ABL-BT} 164 | xtabs(~ BT + mol.biol, bcrabl) 165 | ``` 166 | 167 | Use `aggregate()` to calculate the average age of males and females in 168 | the BCR/ABL and NEG treatment groups. 169 | 170 | ```{r ALL-aggregate} 171 | aggregate(age ~ mol.biol + sex, bcrabl, mean) 172 | ``` 173 | 174 | Use `t.test()` to compare the age of individuals in the BCR/ABL versus 175 | NEG groups; visualize the results using `boxplot()`. In both cases, 176 | use the `formula` interface. Consult the help page `?t.test` and re-do 177 | the test assuming that variance of ages in the two groups is 178 | identical. What parts of the test output change? 179 | 180 | ```{r ALL-age} 181 | t.test(age ~ mol.biol, bcrabl) 182 | boxplot(age ~ mol.biol, bcrabl) 183 | ``` 184 | 185 | ## Resources 186 | 187 | - [StackOverflow](http://stackoverflow.com/questions/tagged/r) for _R_ 188 | programming questions; also [R-help]() mailing list. 189 | 190 | Publications (General _R_) 191 | 192 | 193 | 194 | [R]: http://r-project.org 195 | -------------------------------------------------------------------------------- /vignettes/A01.2_IntroductionToBioconductor.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | knitr::opts_chunk$set(tidy=FALSE) 15 | ``` 16 | 17 | # Introduction to Bioconductor 18 | 19 | Martin Morgan
20 | October 29, 2014 21 | 22 | ## Bioconductor 23 | 24 | Analysis and comprehension of high-throughput genomic data 25 | 26 | - Statistical analysis: large data, technological artifacts, designed 27 | experiments; rigorous 28 | - Comprehension: biological context, visualization, reproducibility 29 | - High-throughput 30 | - Sequencing: RNASeq, ChIPSeq, variants, copy number, ... 31 | - Microarrays: expression, SNP, ... 32 | - Flow cytometry, proteomics, images, ... 33 | 34 | Packages, vignettes, work flows 35 | 36 | - 934 packages 37 | - Discover and navigate via [biocViews][] 38 | - Package 'landing page' 39 | - Title, author / maintainer, short description, citation, 40 | installation instructions, ..., download statistics 41 | - All user-visible functions have help pages, most with runnable 42 | examples 43 | - 'Vignettes' an important feature in Bioconductor -- narrative 44 | documents illustrating how to use the package, with integrated code 45 | - 'Release' (every six months) and 'devel' branches 46 | - [Support site](https://support.bioconductor.org); 47 | [videos](https://www.youtube.com/user/bioconductor), [recent 48 | courses](http://bioconductor.org/help/course-materials/) 49 | 50 | Objects 51 | 52 | - Represent complicated data types 53 | - Foster interoperability 54 | - S4 object system 55 | - Introspection: `getClass()`, `showMethods(..., where=search())`, 56 | `selectMethod()` 57 | - 'accessors' and other documented functions / methods for 58 | manipulation, rather than direct access to the object structure 59 | - Interactive help 60 | - `method?"substr,"` to select help on methods, `class?D` 61 | for help on classes 62 | 63 | Example 64 | 65 | ```{r Biostrings, message=FALSE} 66 | suppressPackageStartupMessages({ 67 | library(Biostrings) 68 | }) 69 | data(phiX174Phage) # sample data, see ?phiX174Phage 70 | phiX174Phage 71 | m <- consensusMatrix(phiX174Phage)[1:4,] # nucl. x position counts 72 | polymorphic <- which(colSums(m != 0) > 1) 73 | m[, polymorphic] 74 | ``` 75 | ```{r showMethods, eval=FALSE} 76 | showMethods(class=class(phiX174Phage), where=search()) 77 | ``` 78 | 79 | ## Core concepts 80 | 81 | ### Genomic ranges 82 | 83 | Genomic range 84 | 85 | - chromosome (`seqnames`), start, end, and optionally strand 86 | - Coordinates 87 | - 1-based 88 | - Closed -- start and end coordinates _included_ in range 89 | - Left-most -- start is always to the left of end, regardless of 90 | strand 91 | 92 | Why genomic ranges? 93 | 94 | - 'Annotation' 95 | - Many genome annotations are range-based 96 | - Simple ranges: exons, promoters, transcription factor binding 97 | sites, CpG islands, ... 98 | - Lists of ranges: gene models (exons-within-transcripts) 99 | - 'Data' 100 | - Reads themselves, or derived data 101 | - Simple ranges: ChIP-seq peaks, SNPs, ungapped reads, ... 102 | - List of ranges: gapped alignments, paired-end reads, ... 103 | 104 | Data objects 105 | 106 | - `r Biocpkg("GenomicRanges")`::_GRanges_ 107 | - `seqnames()` 108 | - `start()`, `end()`, `width()` 109 | - `strand()` 110 | - `mcols()`: 'metadata' associated with each range, stored as a 111 | `DataFrame` 112 | - Many very useful operations defined on ranges (later) 113 | 114 | - `r Biocpkg("GenomicRanges")`::_GRangesList_ 115 | - List-like (e.g., `length()`, `names()`, `[`, `[[`) 116 | - Each list element a _GRanges_ 117 | - Metadata at list and element-list levels 118 | - Very easy (fast) to `unlist()` and `relist()`. 119 | 120 | - `r Biocpkg("GenomicAlignments")`::_GAlignments_, _GAlignmentsList_, 121 | _GAlignemntPairs_; `r Biocpkg("VariantAnnotation")`::_VCF_, _VRanges_ 122 | - _GRanges_-like objects with more specialized roles 123 | 124 | Example: _GRanges_ 125 | 126 | ```{r eg-GRanges} 127 | ## 'Annotation' package; more later... 128 | suppressPackageStartupMessages({ 129 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 130 | }) 131 | promoters <- promoters(TxDb.Hsapiens.UCSC.hg19.knownGene) 132 | ## 'GRanges' with 2 metadata columns 133 | promoters 134 | head(table(seqnames(promoters))) 135 | table(strand(promoters)) 136 | seqinfo(promoters) 137 | ## vector-like access 138 | promoters[ seqnames(promoters) %in% c("chr1", "chr2") ] 139 | ## metadata 140 | mcols(promoters) 141 | length(unique(promoters$tx_name)) 142 | ``` 143 | 144 | ```{r eg-GRangesList} 145 | ## exons, grouped by transcript 146 | exByTx <- exonsBy(TxDb.Hsapiens.UCSC.hg19.knownGene, "tx", use.names=TRUE) 147 | ## list-like subsetting 148 | exByTx[1:10] # also logical, character, ... 149 | exByTx[["uc001aaa.3"]] # also numeric 150 | ## accessors return typed-List, e.g., IntegerList 151 | width(exByTx) 152 | log10(width(exByTx)) 153 | ## 'easy' to ask basic questions, e.g., ... 154 | hist(unlist(log10(width(exByTx)))) # widths of exons 155 | exByTx[which.max(max(width(exByTx)))] # transcript with largest exon 156 | exByTx[which.max(elementLengths(exByTx))] # transcript with most exons 157 | ``` 158 | 159 | There are many neat range-based operations (more later)! 160 | 161 | ![Range Operations](our_figures/RangeOperations.png) 162 | 163 | Some detail 164 | 165 | - _GRanges_ and friends use data structures defined in `r Biocpkg("S4Vectors")`, 166 | `r Biocpkg("IRanges")` 167 | - These data structures can handle relatively large data easily, e.g., 168 | 1-10 million ranges 169 | - Basic concepts are built on _R_'s vector and list; _List_ instances 170 | are implemented to be efficient when there are long lists of a few 171 | elements each. 172 | - Takes a little getting used to, but very powerful 173 | 174 | ### Integrated containers 175 | 176 | What is an experiment? 177 | 178 | - 'Assays' 179 | - Regions-of-interest x samples 180 | - E.g., read counts, expression values 181 | - Regions-of-interest 182 | - Microarrays: probeset or gene identifiers 183 | - Sequencing: genomic ranges 184 | - Samples 185 | - Experimental inforamtion, covariates 186 | - Overall experimental description 187 | 188 | Why integrate? 189 | 190 | - Avoid errors when manipulating data 191 | - Case study: [reproducible research]() 192 | 193 | Data objects 194 | 195 | - `r Biocpkg("Biobase")`::_ExpressionSet_ 196 | - Assays (`exprs()`): matrix of expression values 197 | - Regions-of-interest (`featureData(); fData()`): probeset or gene 198 | identifiers 199 | - Samples (`phenoData(); pData()`: `data.frame` of relevant 200 | information 201 | - Experiment data (`exptData()`): Instance of class `MIAME`. 202 | - `r Biocpkg("GenomicRanges")`::_SummarizedExperiment_ 203 | - Assays (`assay(), assays()`): arbitrary matrix-like object 204 | - Regions-of-interest (`rowData()`): `GRanges` or `GRangesList`; 205 | use `GRangesList` with names and 0-length elements to represent 206 | assays without ranges. 207 | - Samples (`colData()`): `DataFrame` of relevant information. 208 | - Experiment data (`exptData()`): `List` of arbitrary information. 209 | 210 | ![SummarizedExperiment](our_figures/SummarizedExperiment.png) 211 | 212 | Example: `ExpressionSet` (see vignettes in `r Biocpkg("Biobase")`). 213 | 214 | ```{r eg-ExpressionSet} 215 | suppressPackageStartupMessages({ 216 | library(ALL) 217 | }) 218 | data(ALL) 219 | ALL 220 | ## 'Phenotype' (sample) and 'feature' data 221 | head(pData(ALL)) 222 | head(featureNames(ALL)) 223 | ## access to pData columns; matrix-like subsetting; exprs() 224 | ALL[, ALL$sex %in% "M"] 225 | range(exprs(ALL)) 226 | ## 30% 'most variable' features (c.f., genefilter::varFilter) 227 | iqr <- apply(exprs(ALL), 1, IQR) 228 | ALL[iqr > quantile(iqr, 0.7), ] 229 | ``` 230 | 231 | Example: `SummarizedExperiment` (see vignettes in `r Biocpkg("GenomicRanges")`). 232 | 233 | ```{r eg-SummarizedExperiment} 234 | 235 | suppressPackageStartupMessages({ 236 | library(airway) 237 | }) 238 | data(airway) 239 | airway 240 | ## column and row data 241 | colData(airway) 242 | rowData(airway) 243 | ## access colData; matrix-like subsetting; assay() / assays() 244 | airway[, airway$dex %in% "trt"] 245 | head(assay(airway)) 246 | assays(airway) 247 | ## library size 248 | colSums(assay(airway)) 249 | hist(rowMeans(log10(assay(airway)))) 250 | ``` 251 | 252 | ## Lab 253 | 254 | ### GC content 255 | 256 | 1. Calculate the GC content of human chr1 in the hg19 build, 257 | excluding regions where the sequence is "N". You will need to 258 | 259 | 1. Load the `r Biocannopkg("BSgenome.Hsapiens.UCSC.hg19")` 260 | 2. Extract, using `[[`, chromosome 1 ("chr1"). 261 | 3. Use `alphabetFrequency()` to calculate the count or frequency 262 | of the nucleotides in chr1 263 | 4. Use standard _R_ functions to calculate the GC content. 264 | 265 | ```{r gc-reference} 266 | library(BSgenome.Hsapiens.UCSC.hg19) 267 | chr1seq <- BSgenome.Hsapiens.UCSC.hg19[["chr1"]] 268 | chr1alf <- alphabetFrequency(chr1seq) 269 | chr1gc <- sum(chr1alf[c("G", "C")]) / sum(chr1alf[c("A", "C", "G", "T")]) 270 | ``` 271 | 272 | 2. Calculate the GC content of 'exome' (approximately, all genic 273 | regions) on chr1. You will need to 274 | 275 | 1. Load the `r Biocannopkg("TxDb.Hsapiens.UCSC.hg19.knownGene")` 276 | package. 277 | 2. Use `genes()` to extract genic regions of all genes, then 278 | subsetting operations to restrict to chromosome 1. 279 | 3. Use `getSeq,BSgenome-method` to extract sequences from 280 | chromosome 1 of the BSgenome object. 281 | 4. Use `alphabetFrequency()` (with the argument `collapse=TRUE` -- 282 | why?) and standard _R_ operations to extract the gc content of 283 | the genes. 284 | 285 | ```{r gc-exons-1} 286 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 287 | genes <- genes(TxDb.Hsapiens.UCSC.hg19.knownGene) 288 | genes1 <- genes[seqnames(genes) %in% "chr1"] 289 | seq1 <- getSeq(BSgenome.Hsapiens.UCSC.hg19, genes1) 290 | alf1 <- alphabetFrequency(seq1, collapse=TRUE) 291 | gc1 <- sum(alf1[c("G", "C")]) / sum(alf1[c("A", "C", "G", "T")]) 292 | ``` 293 | 294 | How does the GC content just calculated compare to the average of 295 | the GC content of each exon? Answer this using 296 | `alphabetFrequency()` but with `collapse=FALSE)`, and adjust the 297 | calculation of GC content to act on a matrix, rather than 298 | vector. Why are these numbers different? 299 | 300 | ```{r gc-exons-2} 301 | alf2 <- alphabetFrequency(seq1, collapse=FALSE) 302 | gc2 <- rowSums(alf2[, c("G", "C")]) / rowSums(alf2[,c("A", "C", "G", "T")]) 303 | ``` 304 | 305 | 3. Plot a histogram of per-gene GC content, annotating with 306 | information about chromosome and exome GC content. Use base 307 | graphics `hist()`, `abline()`, `plot(density(...))`, 308 | `plot(ecdf(...))`, etc. (one example is below). If this is too 309 | easy, prepare a short presentation for the class illustrating how 310 | to visualize this type of information using another _R_ graphics 311 | package, e.g., `r CRANpkg("ggplot2")`, `{r CRANpkg("ggvis")`, or 312 | `{r CRANpkg("lattice")}. 313 | 314 | ```{r gc-denisty} 315 | plot(density(gc2)) 316 | abline(v=c(chr1gc, gc1), col=c("red", "blue"), lwd=2) 317 | ``` 318 | 319 | ### Integrated containers 320 | 321 | This exercise illustrates how integrated containers can be used to 322 | effectively manage data; it does _NOT_ represent a suitable way to 323 | analyze RNASeq differential expression data. 324 | 325 | 1. Load the `r Biocpkg("airway")` package and `airway` data 326 | set. Explore it a litte, e.g., determining its dimensions (number 327 | of regions of interest and samples), the information describing 328 | samples, and the range of values in the `count` assay. The data are 329 | from an RNA-seq experiment. The `colData()` describe treatment 330 | groups and other information. The `assay()` is the (raw) number of 331 | short reads overlapping each region of interest, in each 332 | sample. The solution to this exercise is summarized above. 333 | 334 | 2. Create a subset of the data set that contains only the 30% most 335 | variable (using IQR as a metric) observations. Plot the 336 | distribution of asinh-transformed (a log-like transformation, 337 | except near 0) row mean counts 338 | 339 | ```{r airway-plot} 340 | iqr <- apply(assay(airway), 1, IQR) 341 | airway1 <- airway[iqr > quantile(iqr, 0.7),] 342 | plot(density(rowMeans(asinh(assay(airway1))))) 343 | ``` 344 | 345 | 3. Use the `r Biocpkg("genefilter")` package `rowttests` function 346 | (consult it's help page!) to compare asinh-transformed read counts 347 | between the two `dex` treatment groups for each row. Explore the 348 | result in various ways, e.g., finding the 'most' differentially 349 | expressed genes, the genes with largest (absolute) difference 350 | between treatment groups, adding adjusted _P_ values (via 351 | `p.adjust()`, in the _stats_ package), etc. Can you obtain the 352 | read counts for each treatment group, for the most differentially 353 | expressed gene? 354 | 355 | ```{r airway-rowttest} 356 | suppressPackageStartupMessages({ 357 | library(genefilter) 358 | }) 359 | ttest <- rowttests(asinh(assay(airway1)), airway1$dex) 360 | ttest$p.adj <- p.adjust(ttest$p.value, method="BH") 361 | ttest[head(order(ttest$p.adj)),] 362 | split(assay(airway1)[order(ttest$p.adj)[1], ], airway1$dex) 363 | ``` 364 | 365 | 4. Add the statistics of differential expression to the `airway1` 366 | _SummarizedExperiment_. Confirm that the statistics have been 367 | added. 368 | 369 | ```{r airway-merge} 370 | mcols(rowData(airway1)) <- ttest 371 | head(mcols(airway1)) 372 | ``` 373 | 374 | # Resources 375 | 376 | - [Web site][Bioconductor] -- install, learn, use, develop _R_ / 377 | _Bioconductor_ packages 378 | - [Support](http://support.bioconductor.org) -- seek help and 379 | guidance; also 380 | - [biocViews](http://bioconductor.org/packages/release/BiocViews.html) 381 | -- discover packages 382 | - Package landing pages, e.g., 383 | [GenomicRanges](http://bioconductor.org/packages/release/bioc/html/GenomicRanges.html), 384 | including title, description, authors, installation instructions, 385 | vignettes (e.g., GenomicRanges '[How 386 | To](http://bioconductor.org/packages/release/bioc/vignettes/GenomicRanges/inst/doc/GenomicRangesHOWTOs.pdf)'), 387 | etc. 388 | - [Course](http://bioconductor.org/help/course-materials/) and other 389 | [help](http://bioconductor.org/help/) material (e.g., videos, EdX 390 | course, community blogs, ...) 391 | 392 | Publications (General _Bioconductor_) 393 | 394 | - Lawrence M, Huber W, Pagès H, Aboyoun P, Carlson M, et al. (2013) 395 | Software for Computing and Annotating Genomic Ranges. PLoS Comput 396 | Biol 9(8): e1003118. doi: 397 | [10.1371/journal.pcbi.1003118][GRanges.bib] 398 | 399 | Other 400 | 401 | - Lawrence, M. 2014. Software for Enabling Genomic Data 402 | Analysis. Bioc2014 conference [slides][Lawrence.bioc2014.bib]. 403 | 404 | [R]: http://r-project.org 405 | [Bioconductor]: http://bioconductor.org 406 | [GRanges.bib]: https://doi.org/10.1371/journal.pcbi.1003118 407 | [Scalable.bib]: http://arxiv.org/abs/1409.2864 408 | [Lawrence.bioc2014.bib]: 409 | http://bioconductor.org/help/course-materials/2014/BioC2014/Lawrence_Talk.pdf 410 | 411 | [AnnotationData]: http://bioconductor.org/packages/release/BiocViews.html#___AnnotationData 412 | [AnnotationDbi]: http://bioconductor.org/packages/release/bioc/html/AnnotationDbi.html 413 | [AnnotationHub]: http://bioconductor.org/packages/release/bioc/html/AnnotationHub.html 414 | [BSgenome.Hsapiens.UCSC.hg19]: http://bioconductor.org/packages/release/data/annotation/html/BSgenome.Hsapiens.UCSC.hg19.html 415 | [BSgenome]: http://bioconductor.org/packages/release/bioc/html/BSgenome.html 416 | [BiocParallel]: http://bioconductor.org/packages/release/bioc/html/BiocParallel.html 417 | [Biostrings]: http://bioconductor.org/packages/release/bioc/html/Biostrings.html 418 | [Bsgenome.Hsapiens.UCSC.hg19]: http://bioconductor.org/packages/release/data/annotation/html/Bsgenome.Hsapiens.UCSC.hg19.html 419 | [CNTools]: http://bioconductor.org/packages/release/bioc/html/CNTools.html 420 | [ChIPQC]: http://bioconductor.org/packages/release/bioc/html/ChIPQC.html 421 | [ChIPpeakAnno]: http://bioconductor.org/packages/release/bioc/html/ChIPpeakAnno.html 422 | [DESeq2]: http://bioconductor.org/packages/release/bioc/html/DESeq2.html 423 | [DiffBind]: http://bioconductor.org/packages/release/bioc/html/DiffBind.html 424 | [GenomicAlignments]: http://bioconductor.org/packages/release/bioc/html/GenomicAlignments.html 425 | [GenomicFiles]: http://bioconductor.org/packages/release/bioc/html/GenomicFiles.html 426 | [GenomicRanges]: http://bioconductor.org/packages/release/bioc/html/GenomicRanges.html 427 | [Homo.sapiens]: http://bioconductor.org/packages/release/data/annotation/html/Homo.sapiens.html 428 | [IRanges]: http://bioconductor.org/packages/release/bioc/html/IRanges.html 429 | [KEGGREST]: http://bioconductor.org/packages/release/bioc/html/KEGGREST.html 430 | [PSICQUIC]: http://bioconductor.org/packages/release/bioc/html/PSICQUIC.html 431 | [Rsamtools]: http://bioconductor.org/packages/release/bioc/html/Rsamtools.html 432 | [Rsubread]: http://bioconductor.org/packages/release/bioc/html/Rsubread.html 433 | [ShortRead]: http://bioconductor.org/packages/release/bioc/html/ShortRead.html 434 | [SomaticSignatures]: http://bioconductor.org/packages/release/bioc/html/SomaticSignatures.html 435 | [TxDb.Hsapiens.UCSC.hg19.knownGene]: http://bioconductor.org/packages/release/data/annotation/html/TxDb.Hsapiens.UCSC.hg19.knownGene.html 436 | [VariantAnnotation]: http://bioconductor.org/packages/release/bioc/html/VariantAnnotation.html 437 | [VariantFiltering]: http://bioconductor.org/packages/release/bioc/html/VariantFiltering.html 438 | [VariantTools]: http://bioconductor.org/packages/release/bioc/html/VariantTools.html 439 | [biocViews]: http://bioconductor.org/packages/release/BiocViews.html#___Software 440 | [biomaRt]: http://bioconductor.org/packages/release/bioc/html/biomaRt.html 441 | [cn.mops]: http://bioconductor.org/packages/release/bioc/html/cn.mops.html 442 | [edgeR]: http://bioconductor.org/packages/release/bioc/html/edgeR.html 443 | [ensemblVEP]: http://bioconductor.org/packages/release/bioc/html/ensemblVEP.html 444 | [h5vc]: http://bioconductor.org/packages/release/bioc/html/h5vc.html 445 | [limma]: http://bioconductor.org/packages/release/bioc/html/limma.html 446 | [metagenomeSeq]: http://bioconductor.org/packages/release/bioc/html/metagenomeSeq.html 447 | [org.Hs.eg.db]: http://bioconductor.org/packages/release/data/annotation/html/org.Hs.eg.db.html 448 | [org.Sc.sgd.db]: http://bioconductor.org/packages/release/data/annotation/html/org.Sc.sgd.db.html 449 | [phyloseq]: http://bioconductor.org/packages/release/bioc/html/phyloseq.html 450 | [rtracklayer]: http://bioconductor.org/packages/release/bioc/html/rtracklayer.html 451 | [snpStats]: http://bioconductor.org/packages/release/bioc/html/snpStats.html 452 | [Gviz]: http://bioconductor.org/packages/release/bioc/html/Gviz.html 453 | [epivizr]: http://bioconductor.org/packages/release/bioc/html/epivizr.html 454 | [ggbio]: http://bioconductor.org/packages/release/bioc/html/ggbio.html 455 | [OmicCircos]: http://bioconductor.org/packages/release/bioc/html/OmicCircos.html 456 | 457 | 458 | -------------------------------------------------------------------------------- /vignettes/A01.3_BioconductorForSequenceAnalysis.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | knitr::opts_chunk$set(tidy=FALSE) 15 | ``` 16 | 17 | # Bioconductor for Sequence Analysis 18 | 19 | Martin Morgan
20 | October 29, 2014 21 | 22 | ## Sequence analysis work flows 23 | 24 | 1. Experimental design 25 | - Keep it simple, e.g., 'control' and 'treatment' groups 26 | - Replicate within treatments! 27 | 2. Wet-lab sequence preparation 28 | - Record covariates, including processing day -- likely 'batch effects' 29 | 3. (Illumina) Sequencing (Bentley et al., 2008, 30 | [doi:10.1038/nature07517](doi:10.1038/nature07517)) 31 | 32 | - Primary output: FASTQ files of short reads and their [quality 33 | scores](http://en.wikipedia.org/wiki/FASTQ_format#Encoding) 34 | 4. Alignment 35 | - Choose to match task, e.g., [Rsubread][], Bowtie2 good for ChIPseq, 36 | some forms of RNAseq; BWA, GMAP better for variant calling 37 | - Primary output: BAM files of aligned reads 38 | 5. Reduction 39 | - e.g., RNASeq 'count table' (simple spreadsheets), DNASeq called 40 | variants (VCF files), ChIPSeq peaks (BED, WIG files) 41 | 6. Analysis 42 | - Differential expression, peak identification, ... 43 | 7. Comprehension 44 | - Biological context 45 | 46 | Data movement 47 | 48 | ![Alt Sequencing Ecosystem](our_figures/SequencingEcosystem_no_bioc_pkgs.png) 49 | 50 | ## Sequence data representations 51 | 52 | ### DNA / amino acid sequences: FASTA files 53 | 54 | Input & manipulation: [Biostrings][] 55 | 56 | >NM_078863_up_2000_chr2L_16764737_f chr2L:16764737-16766736 57 | gttggtggcccaccagtgccaaaatacacaagaagaagaaacagcatctt 58 | gacactaaaatgcaaaaattgctttgcgtcaatgactcaaaacgaaaatg 59 | ... 60 | atgggtatcaagttgccccgtataaaaggcaagtttaccggttgcacggt 61 | >NM_001201794_up_2000_chr2L_8382455_f chr2L:8382455-8384454 62 | ttatttatgtaggcgcccgttcccgcagccaaagcactcagaattccggg 63 | cgtgtagcgcaacgaccatctacaaggcaatattttgatcgcttgttagg 64 | ... 65 | 66 | Whole genomes: `2bit` and `.fa` formats: [rtracklayer][], 67 | [Rsamtools][]; [BSgenome][] 68 | 69 | ### Reads: FASTQ files 70 | 71 | Input & manipulation: [ShortRead][] `readFastq()`, `FastqStreamer()`, 72 | `FastqSampler()` 73 | 74 | @ERR127302.1703 HWI-EAS350_0441:1:1:1460:19184#0/1 75 | CCTGAGTGAAGCTGATCTTGATCTACGAAGAGAGATAGATCTTGATCGTCGAGGAGATGCTGACCTTGACCT 76 | + 77 | HHGHHGHHHHHHHHDGG>CE?=896=: 78 | @ERR127302.1704 HWI-EAS350_0441:1:1:1460:16861#0/1 79 | GCGGTATGCTGGAAGGTGCTCGAATGGAGAGCGCCAGCGCCCCGGCGCTGAGCCGCAGCCTCAGGTCCGCCC 80 | + 81 | DE?DD>ED4>EEE>DE8EEEDE8B?EB<@3;BA79?,881B?@73;1?######################## 82 | 83 | - Quality scores: 'phred-like', encoded. See 84 | [wikipedia](http://en.wikipedia.org/wiki/FASTQ_format#Encoding) 85 | 86 | ### Aligned reads: BAM files (e.g., ERR127306_chr14.bam) 87 | 88 | Input & manipulation: 'low-level' [Rsamtools][], `scanBam()`, 89 | `BamFile()`; 'high-level' [GenomicAlignments][] 90 | 91 | - Header 92 | 93 | @HD VN:1.0 SO:coordinate 94 | @SQ SN:chr1 LN:249250621 95 | @SQ SN:chr10 LN:135534747 96 | @SQ SN:chr11 LN:135006516 97 | ... 98 | @SQ SN:chrY LN:59373566 99 | @PG ID:TopHat VN:2.0.8b CL:/home/hpages/tophat-2.0.8b.Linux_x86_64/tophat --mate-inner-dist 150 --solexa-quals --max-multihits 5 --no-discordant --no-mixed --coverage-search --microexon-search --library-type fr-unstranded --num-threads 2 --output-dir tophat2_out/ERR127306 /home/hpages/bowtie2-2.1.0/indexes/hg19 fastq/ERR127306_1.fastq fastq/ERR127306_2.fastq 100 | 101 | - Alignments: ID, flag, alignment and mate 102 | 103 | ERR127306.7941162 403 chr14 19653689 3 72M = 19652348 -1413 ... 104 | ERR127306.22648137 145 chr14 19653692 1 72M = 19650044 -3720 ... 105 | ERR127306.933914 339 chr14 19653707 1 66M120N6M = 19653686 -213 ... 106 | ERR127306.11052450 83 chr14 19653707 3 66M120N6M = 19652348 -1551 ... 107 | ERR127306.24611331 147 chr14 19653708 1 65M120N7M = 19653675 -225 ... 108 | ERR127306.2698854 419 chr14 19653717 0 56M120N16M = 19653935 290 ... 109 | ERR127306.2698854 163 chr14 19653717 0 56M120N16M = 19653935 2019 ... 110 | 111 | - Alignments: sequence and quality 112 | 113 | ... GAATTGATCAGTCTCATCTGAGAGTAACTTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCC *'%%%%%#&&%''#'&%%%)&&%%$%%'%%'&*****$))$)'')'%)))&)%%%%$'%%%%&"))'')%)) 114 | ... TTGATCAGTCTCATCTGAGAGTAACTTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCCCAG '**)****)*'*&*********('&)****&***(**')))())%)))&)))*')&***********)**** 115 | ... TGAGAGTAACTTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCCCAGCAGCCTCTGGTTTCT '******&%)&)))&")')'')'*((******&)&'')'))$))'')&))$)**&&**************** 116 | ... TGAGAGTAACTTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCCCAGCAGCCTCTGGTTTCT ##&&(#')$')'%&&#)%$#$%"%###&!%))'%%''%'))&))#)&%((%())))%)%)))%********* 117 | ... GAGAGTAACTTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCCCAGCAGCCTCTGGTTTCTT )&$'$'$%!&&%&&#!'%'))%''&%'&))))''$""'%'%&%'#'%'"!'')#&)))))%$)%)&'"'))) 118 | ... TTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCCCAGCAGCCTCTGGTTTCTTCATGTGGCT ++++++++++++++++++++++++++++++++++++++*++++++**++++**+**''**+*+*'*)))*)# 119 | ... TTTGTACCCATCACTGATTCCTTCTGAGACTGCCTCCACTTCCCCAGCAGCCTCTGGTTTCTTCATGTGGCT ++++++++++++++++++++++++++++++++++++++*++++++**++++**+**''**+*+*'*)))*)# 120 | 121 | - Alignments: Tags 122 | 123 | ... AS:i:0 XN:i:0 XM:i:0 XO:i:0 XG:i:0 NM:i:0 MD:Z:72 YT:Z:UU NH:i:2 CC:Z:chr22 CP:i:16189276 HI:i:0 124 | ... AS:i:0 XN:i:0 XM:i:0 XO:i:0 XG:i:0 NM:i:0 MD:Z:72 YT:Z:UU NH:i:3 CC:Z:= CP:i:19921600 HI:i:0 125 | ... AS:i:0 XN:i:0 XM:i:0 XO:i:0 XG:i:0 NM:i:4 MD:Z:72 YT:Z:UU XS:A:+ NH:i:3 CC:Z:= CP:i:19921465 HI:i:0 126 | ... AS:i:0 XN:i:0 XM:i:0 XO:i:0 XG:i:0 NM:i:4 MD:Z:72 YT:Z:UU XS:A:+ NH:i:2 CC:Z:chr22 CP:i:16189138 HI:i:0 127 | ... AS:i:0 XN:i:0 XM:i:0 XO:i:0 XG:i:0 NM:i:5 MD:Z:72 YT:Z:UU XS:A:+ NH:i:3 CC:Z:= CP:i:19921464 HI:i:0 128 | ... AS:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:72 NM:i:0 XS:A:+ NH:i:5 CC:Z:= CP:i:19653717 HI:i:0 129 | ... AS:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:72 NM:i:0 XS:A:+ NH:i:5 CC:Z:= CP:i:19921455 HI:i:1 130 | 131 | ### Called variants: VCF files 132 | 133 | Input and manipulation: [VariantAnnotation][] `readVcf()`, 134 | `readInfo()`, `readGeno()` selectively with `ScanVcfParam()`. 135 | 136 | - Header 137 | 138 | ##fileformat=VCFv4.2 139 | ##fileDate=20090805 140 | ##source=myImputationProgramV3.1 141 | ##reference=file:///seq/references/1000GenomesPilot-NCBI36.fasta 142 | ##contig= 143 | ##phasing=partial 144 | ##INFO= 145 | ##INFO= 146 | ... 147 | ##FILTER= 148 | ##FILTER= 149 | ... 150 | ##FORMAT= 151 | ##FORMAT= 152 | 153 | - Location 154 | 155 | #CHROM POS ID REF ALT QUAL FILTER ... 156 | 20 14370 rs6054257 G A 29 PASS ... 157 | 20 17330 . T A 3 q10 ... 158 | 20 1110696 rs6040355 A G,T 67 PASS ... 159 | 20 1230237 . T . 47 PASS ... 160 | 20 1234567 microsat1 GTC G,GTCT 50 PASS ... 161 | 162 | - Variant INFO 163 | 164 | #CHROM POS ... INFO ... 165 | 20 14370 ... NS=3;DP=14;AF=0.5;DB;H2 ... 166 | 20 17330 ... NS=3;DP=11;AF=0.017 ... 167 | 20 1110696 ... NS=2;DP=10;AF=0.333,0.667;AA=T;DB ... 168 | 20 1230237 ... NS=3;DP=13;AA=T ... 169 | 20 1234567 ... NS=3;DP=9;AA=G ... 170 | 171 | - Genotype FORMAT and samples 172 | 173 | ... POS ... FORMAT NA00001 NA00002 NA00003 174 | ... 14370 ... GT:GQ:DP:HQ 0|0:48:1:51,51 1|0:48:8:51,51 1/1:43:5:.,. 175 | ... 17330 ... GT:GQ:DP:HQ 0|0:49:3:58,50 0|1:3:5:65,3 0/0:41:3 176 | ... 1110696 ... GT:GQ:DP:HQ 1|2:21:6:23,27 2|1:2:0:18,2 2/2:35:4 177 | ... 1230237 ... GT:GQ:DP:HQ 0|0:54:7:56,60 0|0:48:4:51,51 0/0:61:2 178 | ... 1234567 ... GT:GQ:DP 0/1:35:4 0/2:17:2 1/1:40:3 179 | 180 | ### Genome annotations: BED, WIG, GTF, etc. files 181 | 182 | Input: [rtracklayer][] `import()` 183 | 184 | - BED: range-based annotation (see 185 | http://genome.ucsc.edu/FAQ/FAQformat.html for definition of this and 186 | related formats) 187 | - WIG / bigWig: dense, continuous-valued data 188 | - GTF: gene model 189 | 190 | - Component coordinates 191 | 192 | 7 protein_coding gene 27221129 27224842 . - . ... 193 | ... 194 | 7 protein_coding transcript 27221134 27224835 . - . ... 195 | 7 protein_coding exon 27224055 27224835 . - . ... 196 | 7 protein_coding CDS 27224055 27224763 . - 0 ... 197 | 7 protein_coding start_codon 27224761 27224763 . - 0 ... 198 | 7 protein_coding exon 27221134 27222647 . - . ... 199 | 7 protein_coding CDS 27222418 27222647 . - 2 ... 200 | 7 protein_coding stop_codon 27222415 27222417 . - 0 ... 201 | 7 protein_coding UTR 27224764 27224835 . - . ... 202 | 7 protein_coding UTR 27221134 27222414 . - . ... 203 | 204 | - Annotations 205 | 206 | gene_id "ENSG00000005073"; gene_name "HOXA11"; gene_source "ensembl_havana"; gene_biotype "protein_coding"; 207 | ... 208 | ... transcript_id "ENST00000006015"; transcript_name "HOXA11-001"; transcript_source "ensembl_havana"; tag "CCDS"; ccds_id "CCDS5411"; 209 | ... exon_number "1"; exon_id "ENSE00001147062"; 210 | ... exon_number "1"; protein_id "ENSP00000006015"; 211 | ... exon_number "1"; 212 | ... exon_number "2"; exon_id "ENSE00002099557"; 213 | ... exon_number "2"; protein_id "ENSP00000006015"; 214 | ... exon_number "2"; 215 | ... 216 | 217 | ## Sequence data in _R_ / _Bioconductor_ 218 | 219 | ### Role for Bioconductor 220 | 221 | - Pre-processing and alignment 222 | - Data reduction 223 | - Statistical analysis 224 | - Comprehension -- integrative and 'down stream' analysis 225 | 226 | ![Alt Sequencing Ecosystem](our_figures/SequencingEcosystem.png) 227 | 228 | ### Sequences 229 | 230 | _Biostrings_ classes for DNA or amino acid sequences 231 | 232 | - XString, XStringSet, e.g., DNAString (genomes), 233 | DNAStringSet (reads) 234 | 235 | Methods 236 | 237 | - [Cheat sheat](http://bioconductor.org/packages/release/bioc/vignettes/Biostrings/inst/doc/BiostringsQuickOverview.pdf) 238 | - Manipulation, e.g., `reverseComplement()` 239 | - Summary, e.g., `letterFrequency()` 240 | - Matching, e.g., `matchPDict()`, `matchPWM()` 241 | 242 | Related packages 243 | 244 | - [BSgenome][] 245 | - Whole-genome representations 246 | - Model organism or custom 247 | - 'Masks' to exclude regions (pre-computed or arbitrary) from 248 | calculations 249 | - [BSgenome][]: Whole-genome sequence representation & manipulation) 250 | - [Rsamtools][]: `FaFile` class for indexed on-disk representation 251 | - [rtracklayer][]: UCSC '2bit' (`TwoBitFile`) class for indexed 252 | on-disk representation 253 | 254 | Example 255 | 256 | - Whole-genome sequences are distrubuted by ENSEMBL, NCBI, and others 257 | as FASTA files; model organism whole genome sequences are packaged 258 | into more user-friendly `BSgenome` packages. The following 259 | calculates GC content across chr14. 260 | 261 | ```{r BSgenome-require, message=FALSE} 262 | suppressPackageStartupMessages({ 263 | library(BSgenome.Hsapiens.UCSC.hg19) 264 | }) 265 | chr14_range = GRanges("chr14", IRanges(1, seqlengths(Hsapiens)["chr14"])) 266 | chr14_dna <- getSeq(Hsapiens, chr14_range) 267 | letterFrequency(chr14_dna, "GC", as.prob=TRUE) 268 | ``` 269 | 270 | ### Ranges 271 | 272 | Ranges represent: 273 | - Data, e.g., aligned reads, ChIP peaks, SNPs, CpG islands, ... 274 | - Annotations, e.g., gene models, regulatory elements, methylated 275 | regions 276 | - Ranges are defined by chromosome, start, end, and strand 277 | - Often, metadata is associated with each range, e.g., quality of 278 | alignment, strength of ChIP peak 279 | 280 | Many common biological questions are range-based 281 | - What reads overlap genes? 282 | - What genes are ChIP peaks nearest? 283 | - ... 284 | 285 | 286 | The [GenomicRanges][] package defines essential classes and methods 287 | 288 | - `GRanges` 289 | 290 | ![Alt ](our_figures/GRanges.png) 291 | 292 | - `GRangesList` 293 | 294 | ![Alt ](our_figures/GRangesList.png) 295 | 296 | #### Range operations 297 | 298 | ![Alt Ranges Algebra](our_figures/RangeOperations.png) 299 | 300 | Ranges 301 | - IRanges 302 | - `start()` / `end()` / `width()` 303 | - List-like -- `length()`, subset, etc. 304 | - 'metadata', `mcols()` 305 | - GRanges 306 | - 'seqnames' (chromosome), 'strand' 307 | - `Seqinfo`, including `seqlevels` and `seqlengths` 308 | 309 | Intra-range methods 310 | - Independent of other ranges in the same object 311 | - GRanges variants strand-aware 312 | - `shift()`, `narrow()`, `flank()`, `promoters()`, `resize()`, 313 | `restrict()`, `trim()` 314 | - See `?"intra-range-methods"` 315 | 316 | Inter-range methods 317 | - Depends on other ranges in the same object 318 | - `range()`, `reduce()`, `gaps()`, `disjoin()` 319 | - `coverage()` (!) 320 | - see `?"inter-range-methods"` 321 | 322 | Between-range methods 323 | - Functions of two (or more) range objects 324 | - `findOverlaps()`, `countOverlaps()`, ..., `%over%`, `%within%`, 325 | `%outside%`; `union()`, `intersect()`, `setdiff()`, `punion()`, 326 | `pintersect()`, `psetdiff()` 327 | 328 | Example 329 | 330 | ```{r ranges, message=FALSE} 331 | suppressPackageStartupMessages({ 332 | library(GenomicRanges) 333 | }) 334 | gr <- GRanges("A", IRanges(c(10, 20, 22), width=5), "+") 335 | shift(gr, 1) # 1-based coordinates! 336 | range(gr) # intra-range 337 | reduce(gr) # inter-range 338 | coverage(gr) 339 | setdiff(range(gr), gr) # 'introns' 340 | ``` 341 | 342 | IRangesList, GRangesList 343 | - List: all elements of the same type 344 | - Many *List-aware methods, but a common 'trick': apply a vectorized 345 | function to the unlisted representaion, then re-list 346 | 347 | grl <- GRangesList(...) 348 | orig_gr <- unlist(grl) 349 | transformed_gr <- FUN(orig) 350 | transformed_grl <- relist(, grl) 351 | 352 | Reference 353 | 354 | - Lawrence M, Huber W, Pagès H, Aboyoun P, Carlson M, et al. (2013) 355 | Software for Computing and Annotating Genomic Ranges. PLoS Comput 356 | Biol 9(8): e1003118. doi:10.1371/journal.pcbi.1003118 357 | 358 | ### [GenomicAlignments][] (Aligned reads) 359 | 360 | Classes -- GenomicRanges-like behaivor 361 | 362 | - GAlignments, GAlignmentPairs, GAlignmentsList 363 | - SummarizedExperiment 364 | - Matrix where rows are indexed by genomic ranges, columns by a 365 | DataFrame. 366 | 367 | Methods 368 | 369 | - `readGAlignments()`, `readGAlignmentsList()` 370 | - Easy to restrict input, iterate in chunks 371 | - `summarizeOverlaps()` 372 | 373 | Example 374 | 375 | - Find reads supporting the junction identified above, at position 376 | 19653707 + 66M = 19653773 of chromosome 14 377 | 378 | ```{r bam-require} 379 | suppressPackageStartupMessages({ 380 | library(GenomicRanges) 381 | library(GenomicAlignments) 382 | library(Rsamtools) 383 | }) 384 | 385 | ## our 'region of interest' 386 | roi <- GRanges("chr14", IRanges(19653773, width=1)) 387 | ## sample data 388 | suppressPackageStartupMessages({ 389 | library('RNAseqData.HNRNPC.bam.chr14') 390 | }) 391 | bf <- BamFile(RNAseqData.HNRNPC.bam.chr14_BAMFILES[[1]], asMates=TRUE) 392 | ## alignments, junctions, overlapping our roi 393 | paln <- readGAlignmentsList(bf) 394 | j <- summarizeJunctions(paln, with.revmap=TRUE) 395 | j_overlap <- j[j %over% roi] 396 | 397 | ## supporting reads 398 | paln[j_overlap$revmap[[1]]] 399 | ``` 400 | 401 | ### [VariantAnnotation][] (called variants) 402 | 403 | Classes -- GenomicRanges-like behavior 404 | 405 | - VCF -- 'wide' 406 | - VRanges -- 'tall' 407 | 408 | Functions and methods 409 | 410 | - I/O and filtering: `readVcf()`, `readGeno()`, `readInfo()`, 411 | `readGT()`, `writeVcf()`, `filterVcf()` 412 | - Annotation: `locateVariants()` (variants overlapping ranges), 413 | `predictCoding()`, `summarizeVariants()` 414 | - SNPs: `genotypeToSnpMatrix()`, `snpSummary()` 415 | 416 | Example 417 | 418 | - Read variants from a VCF file, and annotate with respect to a known 419 | gene model 420 | 421 | ```{r vcf, message=FALSE} 422 | ## input variants 423 | suppressPackageStartupMessages({ 424 | library(VariantAnnotation) 425 | }) 426 | fl <- system.file("extdata", "chr22.vcf.gz", package="VariantAnnotation") 427 | vcf <- readVcf(fl, "hg19") 428 | seqlevels(vcf) <- "chr22" 429 | ## known gene model 430 | suppressPackageStartupMessages({ 431 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 432 | }) 433 | coding <- locateVariants(rowData(vcf), 434 | TxDb.Hsapiens.UCSC.hg19.knownGene, 435 | CodingVariants()) 436 | head(coding) 437 | ``` 438 | 439 | Related packages 440 | 441 | - [ensemblVEP][] 442 | - Forward variants to Ensembl Variant Effect Predictor 443 | - [VariantTools][], [h5vc][] 444 | - Call variants 445 | - [VariantFiltering][] 446 | - Filter variants using criteria such as coding consequence, MAF, 447 | ..., inheritance model 448 | 449 | Reference 450 | 451 | - Obenchain, V, Lawrence, M, Carey, V, Gogarten, S, Shannon, P, and 452 | Morgan, M. VariantAnnotation: a Bioconductor package for exploration 453 | and annotation of genetic variants. Bioinformatics, first published 454 | online March 28, 2014 455 | [doi:10.1093/bioinformatics/btu168](http://bioinformatics.oxfordjournals.org/content/early/2014/04/21/bioinformatics.btu168) 456 | 457 | ### [rtracklayer][] (Genome annotations) 458 | 459 | - Import BED, GTF, WIG, etc 460 | - Export GRanges to BED, GTF, WIG, ... 461 | - Access UCSC genome browser 462 | 463 | ## A sequence analysis package tour 464 | 465 | This very open-ended topic points to some of the most prominent 466 | Bioconductor packages for sequence analysis. Use the opportunity in 467 | this lab to explore the package vignettes and help pages highlighted 468 | below; many of the material will be covered in greater detail in 469 | subsequent labs and lectures. 470 | 471 | Basics 472 | 473 | - Bioconductor packages are listed on the [biocViews][] page. Each 474 | package has 'biocViews' (tags from a controlled vocabulary) 475 | associated with it; these can be searched to identify appropriately 476 | tagged packages, as can the package title and author. 477 | - Each package has a 'landing page', e.g., for 478 | [GenomicRanges][]. Visit this landing page, and note the 479 | description, authors, and installation instructions. Packages are 480 | often written up in the scientific literature, and if available the 481 | corresponding citation is present on the landing page. Also on the 482 | landing page are links to the vignettes and reference manual and, at 483 | the bottom, an indication of cross-platform availability and 484 | download statistics. 485 | - A package needs to be installed once, using the instructions on the 486 | landing page. Once installed, the package can be loaded into an R 487 | session 488 | 489 | ```{r require} 490 | suppressPackageStartupMessages({ 491 | library(GenomicRanges) 492 | }) 493 | ``` 494 | 495 | and the help system queried interactively, as outlined above: 496 | 497 | ```{r help, eval=FALSE} 498 | help(package="GenomicRanges") 499 | vignette(package="GenomicRanges") 500 | vignette(package="GenomicRanges", "GenomicRangesHOWTOs") 501 | ?GRanges 502 | ``` 503 | 504 | Domain-specific analysis -- explore the landing pages, vignettes, and 505 | reference manuals of two or three of the following packages. 506 | 507 | - Important packages for analysis of differential expression include 508 | [edgeR][] and [DESeq2][]; both have excellent vignettes for 509 | exploration. Additional research methods embodied in Bioconductor 510 | packages can be discovered by visiting the [biocViews][] web page, 511 | searching for the 'DifferentialExpression' view term, and narrowing 512 | the selection by searching for 'RNA seq' and similar. 513 | - Popular ChIP-seq packages include [DiffBind][] for comparison of 514 | peaks across samples, [ChIPQC][] for quality assessment, and 515 | [ChIPpeakAnno][] for annotating results (e.g., discovering nearby 516 | genes). What other ChIP-seq packages are listed on the [biocViews][] 517 | page? 518 | - Working with called variants (VCF files) is facilitated by packages 519 | such as [VariantAnnotation][], [VariantFiltering][], [ensemblVEP][], 520 | and [SomaticSignatures][]; packages for calling variants include, 521 | e.g., [h5vc][] and [VariantTools][]. 522 | - Several packages identify copy number variants from sequence data, 523 | including [cn.mops][]; from the [biocViews][] page, what other copy 524 | number packages are available? The [CNTools][] package provides some 525 | useful facilities for comparison of segments across samples. 526 | - Microbiome and metagenomic analysis is facilitated by packages such 527 | as [phyloseq][] and [metagenomeSeq][]. 528 | - Metabolomics, chemoinformatics, image analysis, and many other 529 | high-throughput analysis domains are also represented in 530 | Bioconductor; explore these via biocViews and title searches. 531 | 532 | Working with sequences, alignments, common web file formats, and raw 533 | data; these packages rely very heavily on the [IRanges][] / 534 | [GenomicRanges][] infrastructure that we will encounter later in the 535 | course. 536 | 537 | - The [Biostrings][] package is used to represent DNA and other 538 | sequences, with many convenient sequence-related functions. Check 539 | out the functions documented on the help page `?consensusMatrix`, 540 | for instance. Also check out the [BSgenome][] package for working 541 | with whole genome sequences, e.g., `?"getSeq,BSgenome-method"` 542 | - The [GenomicAlignments][] package is used to input reads aligned to 543 | a reference genome. See for instance the `?readGAlignments` help 544 | page and `vigentte(package="GenomicAlignments", 545 | "summarizeOverlaps")` 546 | - [rtracklayer][]'s `import` and `export` functions can read in many 547 | common file types, e.g., BED, WIG, GTF, ..., in addition to querying 548 | and navigating the UCSC genome browser. Check out the `?import` page 549 | for basic usage. 550 | - The [ShortRead][] and [Rsamtools][] packages can be used for 551 | lower-level access to FASTQ and BAM files, respectively. Explore the 552 | [ShortRead vignette](http://bioconductor.org/packages/release/bioc/vignettes/ShortRead/inst/doc/Overview.pdf) 553 | and Scalable Genomics labs to see approaches to effectively 554 | processing the large files. 555 | 556 | Visualization 557 | 558 | - The [Gviz][] package provides great tools for visualizing local 559 | genomic coordinates and associated data. 560 | - [epivizr][] drives the [epiviz](http://epiviz.cbcb.umd.edu/) genome 561 | browser from within R; [rtracklayer][] provides easy ways to 562 | transfer data to and manipulate UCSC browser sessions. 563 | - Additionl packages include [ggbio][], [OmicCircos][], ... 564 | 565 | ## Lab 566 | 567 | ### Short read quality assessment 568 | 569 | `fastqc` is a Java program commonly used for summarizing quality of 570 | fastq files. It has a straight-forward graphical user interface. Here 571 | we will use the command-line version. 572 | 573 | 1. From within _Rstudio_, choose 'Tools --> Shell...', or log on to 574 | your Amazon machine instance using a Mac / linux terminal or on 575 | Windows the PuTTY program. 576 | 577 | 2. Run fastqc on sample fastq files, sending the output to the 578 | `~/fastqc_report` directory. 579 | 580 | fastqc fastq/*fastq --threads 8 --outdir=fastqc_reports 581 | 582 | 3. Study the quality report and resulting on-line [documentation](FIXME): 583 | In the Files tab, click on `fastqc_reports`. Click on the HTML file 584 | there and then click on "View in Web Browser". 585 | 586 | `r Biocpkg("ShortRead")` provides similar functionality, but from 587 | within _R_. The following shows that _R_ can handle large data, and 588 | illustrates some of the basic ways in which one might interact with 589 | functionality provided by a _Bioconductor_ package. 590 | 591 | ```{r ShortRead, messages=FALSE} 592 | ## 1. attach ShortRead and BiocParallel 593 | suppressPackageStartupMessages({ 594 | library(ShortRead) 595 | library(BiocParallel) 596 | }) 597 | 598 | ## 2. create a vector of file paths 599 | fls <- dir("~/fastq", pattern="*fastq", full=TRUE) 600 | 601 | ``` 602 | 603 | ```{r fakestats, eval=FALSE} 604 | ## 3. collect statistics 605 | stats0 <- qa(fls) 606 | ``` 607 | 608 | ```{r realstats, echo=FALSE, results="hide"} 609 | data(stats0) 610 | ``` 611 | 612 | ```{r browseStats} 613 | ## 4. generate and browse the report 614 | if (interactive()) 615 | browseURL(report(stats0)) 616 | ``` 617 | 618 | Check out the qa report from all lanes 619 | 620 | ```{r ShortRead-qa-all} 621 | data(qa_all) 622 | if (interactive()) 623 | browseURL(report(qa_all)) 624 | ``` 625 | 626 | ### Alignments (and genomic annotations) 627 | 628 | This data is from the `r Biocannopkg("airway")` Bioconductor 629 | annotation package; see the 630 | [vignette](http://bioconductor.org/packages/release/data/experiment/vignettes/airway/inst/doc/airway.html) 631 | for details 632 | 633 | Integrative Genomics Viewer 634 | 635 | 1. Start IGV and select the "hg19" genome. 636 | 637 | 2. The sequence names used in the reference genome differ from those 638 | used by IGV to represent the identical genome. We need to map 639 | between these different sequence names, following the instructions 640 | for 641 | [Creating a Chromosome Name Alias File](http://www.broadinstitute.org/software/igv/LoadData/#aliasfile). 642 | 643 | Copy the file `hg19_alias.tab` from the location specified in class 644 | into the directory `/igv/genomes/`. Restart IGV. 645 | 646 | 3. Start igv. 647 | 648 | 4. Choose hg19 from the drop-down menu at the top left of 649 | the screen 650 | 651 | 5. Use `File -> Load from URL` menu to load a bam file. The URLs will 652 | be provided during class. 653 | 654 | 6. Zoom in to a particular gene, e.g., SPARCL1, by entering the gene 655 | symbol in the box toward the center of the browser window. Adjust 656 | the zoom until reads come in to view, and interpret the result. 657 | 658 | _Bioconductor_: we'll explore how to map between different types of 659 | identifiers, how to navigate genomic coordinates, and how to query BAM 660 | files for aligned reads. 661 | 662 | 1. Attach 'Annotation' packages containing information about gene 663 | symbols `r Biocannopkg("org.Hs.eg.db")` and genomic coordinates 664 | (e.g., genes, exons, cds, transcripts) `r 665 | Biocannopkg(TxDb.Hsapiens.UCSC.hg19.knownGene)`. Arrange for the 666 | 'seqlevels' (chromosome names) in the TxDb package to match those 667 | in the BAM files. 668 | 669 | 2. Use an appropriate `org.*` package to map from gene symbol to 670 | Entrez gene id, and the appropriate `TxDb.*` package to retrieve 671 | gene coordinates of the SPARCL1 gene. N.B. -- The following uses a 672 | single gene symbol, but we could have used 1, 2, or all gene 673 | symbols in a _vectorized_ fashion. 674 | 675 | 3. Attach the `r Biocpkg("GenomicAlignments")` package for working 676 | with aligned reads. Use `range()` to get the genomic coordinates 677 | spanning the first and last exon of SPARCL1. Input paired reads 678 | overlapping SPARCL1. 679 | 680 | 4. What questions can you easily answer about these alignments? E.g., 681 | how many reads overlap this region of interest? 682 | 683 | ```{r setup-view, message=FALSE, warning=FALSE} 684 | ## 1.a 'Annotation' packages 685 | suppressPackageStartupMessages({ 686 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 687 | library(org.Hs.eg.db) 688 | }) 689 | 690 | ## 1.b -- map 'seqlevels' as recorded in the TxDb file to those in the 691 | ## BAM file 692 | fl <- "~/igv/genomes/hg19_alias.tab" 693 | map <- with(read.delim(fl, header=FALSE, stringsAsFactors=FALSE), 694 | setNames(V1, V2)) 695 | seqlevels(TxDb.Hsapiens.UCSC.hg19.knownGene, force=TRUE) <- map 696 | 697 | ## 2. Symbol -> Entrez ID -> Gene coordinates 698 | sym2eg <- select(org.Hs.eg.db, "SPARCL1", "ENTREZID", "SYMBOL") 699 | exByGn <- exonsBy(TxDb.Hsapiens.UCSC.hg19.knownGene, "gene") 700 | sparcl1exons <- exByGn[[sym2eg$ENTREZID]] 701 | 702 | ## 3. Aligned reads 703 | suppressPackageStartupMessages({ 704 | library(GenomicAlignments) 705 | }) 706 | 707 | fl <- "~/bam/SRR1039508_sorted.bam" 708 | sparcl1gene <- range(sparcl1exons) 709 | param <- ScanBamParam(which=sparcl1gene) 710 | aln <- readGAlignmentPairs(fl, param=param) 711 | ``` 712 | 713 | 5. As another exercise we ask how many of the reads we've input are 714 | compatible with the known gene model. We have to find the 715 | transcripts that belong to our gene, and then exons grouped by 716 | transcript 717 | 718 | ```{r compatibleAlignments, warning=FALSE} 719 | ## 5.a. exons-by-transcript for our gene of interest 720 | txids <- select(TxDb.Hsapiens.UCSC.hg19.knownGene, sym2eg$ENTREZID, 721 | "TXID", "GENEID")$TXID 722 | exByTx <- exonsBy(TxDb.Hsapiens.UCSC.hg19.knownGene, "tx")[txids] 723 | 724 | ## 5.b compatible alignments 725 | hits <- findCompatibleOverlaps(query=aln, subject=exByTx) 726 | good <- seq_along(aln) %in% queryHits(hits) 727 | table(good) 728 | ``` 729 | 730 | 6. Finally, let's go from gene model to protein coding 731 | sequence. (a) Extract CDS regions grouped by transcript, select just 732 | transcripts we're interested in, (b) attach and then extract the coding 733 | sequence from the appropriate reference genome. Translating the 734 | coding sequences to proteins. 735 | 736 | ```{r coding-sequence, warning=FALSE} 737 | ## reset seqlevels 738 | restoreSeqlevels(TxDb.Hsapiens.UCSC.hg19.knownGene) 739 | 740 | ## a. cds coordinates, grouped by transcript 741 | txids <- select(TxDb.Hsapiens.UCSC.hg19.knownGene, sym2eg$ENTREZID, 742 | "TXID", "GENEID")$TXID 743 | cdsByTx <- cdsBy(TxDb.Hsapiens.UCSC.hg19.knownGene, "tx")[txids] 744 | 745 | ## b. coding sequence from relevant reference genome 746 | suppressPackageStartupMessages({ 747 | library(BSgenome.Hsapiens.UCSC.hg19) 748 | }) 749 | 750 | dna <- extractTranscriptSeqs(BSgenome.Hsapiens.UCSC.hg19, cdsByTx) 751 | protein <- translate(dna) 752 | ``` 753 | 754 | ### Working with genomic ranges 755 | 756 | Visit the "GenomicRanges HOWTOs" vignette. 757 | 758 | ```{r GenomicRanges-howtos, eval=FALSE} 759 | browseVignettes("GenomicRanges") 760 | ``` 761 | 762 | Read section 1, and 763 | do exercises 2.2, 2.4, 2.5, 2.8, 2.12, and 2.13. Perhaps select 764 | additional topics of particular interest to you. 765 | 766 | ## Resources 767 | 768 | _R_ / _Bioconductor_ 769 | 770 | - [Web site][Bioconductor] -- install, learn, use, develop _R_ / 771 | _Bioconductor_ packages 772 | - [Support](http://support.bioconductor.org) -- seek help and 773 | guidance; also 774 | [StackOverflow](http://stackoverflow.com/questions/tagged/r) for _R_ 775 | programming questions 776 | - [biocViews](http://bioconductor.org/packages/release/BiocViews.html) 777 | -- discover packages 778 | - Package landing pages, e.g., 779 | [GenomicRanges](http://bioconductor.org/packages/release/bioc/html/GenomicRanges.html), 780 | including title, description, authors, installation instructions, 781 | vignettes (e.g., GenomicRanges '[How 782 | To](http://bioconductor.org/packages/release/bioc/vignettes/GenomicRanges/inst/doc/GenomicRangesHOWTOs.pdf)'), 783 | etc. 784 | - [Course](http://bioconductor.org/help/course-materials/) and other 785 | [help](http://bioconductor.org/help/) material (e.g., videos, EdX 786 | course, community blogs, ...) 787 | 788 | Publications and presentations 789 | 790 | - Lawrence M, Huber W, Pagès H, Aboyoun P, Carlson M, et al. (2013) 791 | Software for Computing and Annotating Genomic Ranges. PLoS Comput 792 | Biol 9(8): e1003118. doi: 793 | [10.1371/journal.pcbi.1003118][GRanges.bib] 794 | 795 | - Lawrence, M. 2014. Software for Enabling Genomic Data 796 | Analysis. Bioc2014 conference [slides](http://bioconductor.org/help/course-materials/2014/BioC2014/Lawrence_Talk.pdf). 797 | 798 | 799 | 800 | [R]: http://r-project.org 801 | [Bioconductor]: http://bioconductor.org 802 | [GRanges.bib]: https://doi.org/10.1371/journal.pcbi.1003118 803 | [Scalable.bib]: http://arxiv.org/abs/1409.2864 804 | [Lawrence.bioc2014.bib]: 805 | http://bioconductor.org/help/course-materials/2014/BioC2014/Lawrence_Talk.pdf 806 | 807 | 808 | [AnnotationData]: http://bioconductor.org/packages/release/BiocViews.html#___AnnotationData 809 | [AnnotationDbi]: http://bioconductor.org/packages/release/bioc/html/AnnotationDbi.html 810 | [AnnotationHub]: http://bioconductor.org/packages/release/bioc/html/AnnotationHub.html 811 | [BSgenome.Hsapiens.UCSC.hg19]: http://bioconductor.org/packages/release/data/annotation/html/BSgenome.Hsapiens.UCSC.hg19.html 812 | [BSgenome]: http://bioconductor.org/packages/release/bioc/html/BSgenome.html 813 | [Biostrings]: http://bioconductor.org/packages/release/bioc/html/Biostrings.html 814 | [Bsgenome.Hsapiens.UCSC.hg19]: http://bioconductor.org/packages/release/data/annotation/html/Bsgenome.Hsapiens.UCSC.hg19.html 815 | [CNTools]: http://bioconductor.org/packages/release/bioc/html/CNTools.html 816 | [ChIPQC]: http://bioconductor.org/packages/release/bioc/html/ChIPQC.html 817 | [ChIPpeakAnno]: http://bioconductor.org/packages/release/bioc/html/ChIPpeakAnno.html 818 | [DESeq2]: http://bioconductor.org/packages/release/bioc/html/DESeq2.html 819 | [DiffBind]: http://bioconductor.org/packages/release/bioc/html/DiffBind.html 820 | [GenomicAlignments]: http://bioconductor.org/packages/release/bioc/html/GenomicAlignments.html 821 | [GenomicRanges]: http://bioconductor.org/packages/release/bioc/html/GenomicRanges.html 822 | [Homo.sapiens]: http://bioconductor.org/packages/release/data/annotation/html/Homo.sapiens.html 823 | [IRanges]: http://bioconductor.org/packages/release/bioc/html/IRanges.html 824 | [KEGGREST]: http://bioconductor.org/packages/release/bioc/html/KEGGREST.html 825 | [PSICQUIC]: http://bioconductor.org/packages/release/bioc/html/PSICQUIC.html 826 | [Rsamtools]: http://bioconductor.org/packages/release/bioc/html/Rsamtools.html 827 | [Rsubread]: http://bioconductor.org/packages/release/bioc/html/Rsubread.html 828 | [ShortRead]: http://bioconductor.org/packages/release/bioc/html/ShortRead.html 829 | [SomaticSignatures]: http://bioconductor.org/packages/release/bioc/html/SomaticSignatures.html 830 | [TxDb.Hsapiens.UCSC.hg19.knownGene]: http://bioconductor.org/packages/release/data/annotation/html/TxDb.Hsapiens.UCSC.hg19.knownGene.html 831 | [VariantAnnotation]: http://bioconductor.org/packages/release/bioc/html/VariantAnnotation.html 832 | [VariantFiltering]: http://bioconductor.org/packages/release/bioc/html/VariantFiltering.html 833 | [VariantTools]: http://bioconductor.org/packages/release/bioc/html/VariantTools.html 834 | [biocViews]: http://bioconductor.org/packages/release/BiocViews.html#___Software 835 | [biomaRt]: http://bioconductor.org/packages/release/bioc/html/biomaRt.html 836 | [cn.mops]: http://bioconductor.org/packages/release/bioc/html/cn.mops.html 837 | [edgeR]: http://bioconductor.org/packages/release/bioc/html/edgeR.html 838 | [ensemblVEP]: http://bioconductor.org/packages/release/bioc/html/ensemblVEP.html 839 | [h5vc]: http://bioconductor.org/packages/release/bioc/html/h5vc.html 840 | [limma]: http://bioconductor.org/packages/release/bioc/html/limma.html 841 | [metagenomeSeq]: http://bioconductor.org/packages/release/bioc/html/metagenomeSeq.html 842 | [org.Hs.eg.db]: http://bioconductor.org/packages/release/data/annotation/html/org.Hs.eg.db.html 843 | [org.Sc.sgd.db]: http://bioconductor.org/packages/release/data/annotation/html/org.Sc.sgd.db.html 844 | [phyloseq]: http://bioconductor.org/packages/release/bioc/html/phyloseq.html 845 | [rtracklayer]: http://bioconductor.org/packages/release/bioc/html/rtracklayer.html 846 | [snpStats]: http://bioconductor.org/packages/release/bioc/html/snpStats.html 847 | [Gviz]: http://bioconductor.org/packages/release/bioc/html/Gviz.html 848 | [epivizr]: http://bioconductor.org/packages/release/bioc/html/epivizr.html 849 | [ggbio]: http://bioconductor.org/packages/release/bioc/html/ggbio.html 850 | [OmicCircos]: http://bioconductor.org/packages/release/bioc/html/OmicCircos.html 851 | -------------------------------------------------------------------------------- /vignettes/A01_Introduction.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Learning R / Bioconductor for Sequence Analysis 17 | 18 | This package contains training material for a Fall, 2014 introductory 19 | _R_ / _Bioconductor_ course "Learning _R_ / _Bioconductor_ for Sequence 20 | Analysis", offered October 27-29, Seattle, WA. 21 | 22 | This course is directed at beginning and intermediate users who 23 | would like an introduction to the analysis and comprehension of 24 | high-throughput sequence data using _R_ and _Bioconductor_. Day 1 focuses 25 | on learning essential background: an introduction to the _R_ programming 26 | language; central concepts for effective use of _Bioconductor_ software; 27 | and an overview of high-throughput sequence analysis work flows. Day 2 28 | emphasizes use of _Bioconductor_ for specific tasks: an RNA-seq 29 | differential expression work flow; exploratory, machine learning, and 30 | other statistical tasks; gene set enrichment; and annotation. Day 3 31 | transitions to understanding effective approaches for managing larger 32 | challenges: strategies for working with large data, writing re-usable 33 | functions, developing reproducible reports and work flows, and 34 | visualizing results. 35 | 36 | The course combines lectures with extensive hands-on practicals; 37 | students are required to bring a laptop with wireless internet access 38 | and a modern version of the Chrome or Safari web browser. 39 | 40 | ## Schedule (tentative) 41 | 42 | Day 1: Learn _R_ / _Bioconductor_ 43 | 44 | - 9:00 - 10:30 [Introduction to _R_](A01.1_IntroductionToR.html): 45 | objects, functions, help! 46 | - 11:00 - 12:30 47 | [Introduction to _Bioconductor_](A01.2_IntroductionToBioconductor.html): 48 | working with packages and classes 49 | - 1:30 - 5:00 (break: 3:00 - 3:30) 50 | [Introduction to sequence analysis](A01.3_BioconductorForSequenceAnalysis.html): 51 | typical work flow; data types and quality assessment; essential 52 | _Bioconductor_ packages 53 | 54 | Day 2: Use _R_ / _Bioconductor_ 55 | 56 | - 9:00 - 12:30 (break: 10:30 - 11:00) An RNA-seq differential 57 | expression work flow ([lecture](B02.1_RNASeq.html); 58 | [practical](B02.1.1_RNASeqLab.html)) 59 | - 1:30 - 2:00 [Other work flows](B02.2_CommonWorkFlows.html) (survey): 60 | ChIP-seq, variants, copy number, epigenomics 61 | - 2:00 - 3:00 [Machine learning](B02.3_MachineLearning.html); 62 | exploratory and other statistical analysis 63 | - 3:30 - 4:00 64 | [Annotating genes, genomes, and variants](B02.4_Annotation.html) 65 | - 4:00 - 5:00 66 | [Approaches to gene set enrichment](B02.5_GeneSetEnrichment.html) 67 | 68 | Day 3: Develop Skills and Best Practices 69 | 70 | - 9:00 - 10:30 [Working with large data](C03.1_LargeData.html) 71 | - 11:00 - 12:30 72 | [Organizing code in functions, files, and packages](C03.2_CodeToPackages.html) 73 | - 1:30 - 3:00 74 | [Reproducible reports and work flows](C03.3_ReproducibleResearch.html) 75 | - 3:30 - 4:30 [Visualization](C03.4_Visualization.html) 76 | - 4:30 - 5:00 Summary 77 | -------------------------------------------------------------------------------- /vignettes/B02.1_RNASeq.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r style, echo = FALSE, results = 'asis'} 8 | BiocStyle::markdown() 9 | knitr::opts_chunk$set(tidy=FALSE) 10 | ``` 11 | 12 | ```{r setup, echo=FALSE} 13 | library(LearnBioconductor) 14 | stopifnot(BiocInstaller::biocVersion() == "3.0") 15 | ``` 16 | 17 | # RNA-Seq Work Flows 18 | 19 | Martin Morgan, Sonali Arora
20 | October 28, 2014 21 | 22 | ## 7-step work flow 23 | 24 | 25 | ### 1. Experimental design 26 | 27 | Keep it simple 28 | 29 | - Classical experimental designs 30 | - Time series 31 | - Without missing values, where possible 32 | - Intended analysis must be feasbile -- can the available samples and 33 | hypothesis of interest be combined to formulate a testable 34 | statistical hypothesis? 35 | 36 | Replicate 37 | 38 | - Extent of replication determines nuance of biological question. 39 | - No replication (1 sample per treatment): qualitative description 40 | with limited statistical options. 41 | - 3-5 replicates per treatment: designed experimental manipulation 42 | with cell lines or other well-defined entities; 2-fold (?) 43 | change in average expression between groups. 44 | - 10-50 replicates per treatment: population studies, e.g., cancer 45 | cell lines. 46 | - 1000's of replicates: prospective studies, e.g., SNP discovery 47 | - One resource: `r Biocpkg("RNASeqPower")` 48 | 49 | Avoid confounding experimental factors with other factors 50 | 51 | - Common problems: samples from one treatment all on the same flow 52 | cell; samples from treatment 1 processed first, treatment 2 53 | processed second, etc. 54 | 55 | Record co-variates 56 | 57 | Be aware of _batch effects_ 58 | 59 | - Leek et al., 2010, Nature Reviews Genetics 11 60 | [733-739](http://www.nature.com/nrg/journal/v11/n10/abs/nrg2825.html), 61 | Leek & Story PLoS Genet 3(9): 62 | [e161](https://doi.org/10.1371/journal.pgen.0030161). 63 | - Scientific finding: pervasive batch effects 64 | - Statistical insights: surrogate variable analysis: identify and 65 | build surrogate variables; remove known batch effects 66 | - Benefits: reduce dependence, stabilize error rate estimates, and 67 | improve reproducibility 68 | - _combat_ software / `r Biocpkg("sva")` _Bioconductor_ package 69 | 70 | ![](our_figures/nrg2825-f2.jpg) 71 | HapMap samples from one facility, ordered by date of processing. 72 | 73 | ### 2. Wet-lab 74 | 75 | Confounding factors 76 | 77 | - Record or avoid 78 | 79 | Artifacts of your _particular_ protocols 80 | 81 | - Sequence contaminants 82 | - Enrichment bias, e.g., non-uniform transcript representation. 83 | - PCR artifacts -- adapter contaminants, sequence-specific 84 | amplification bias, ... 85 | 86 | ### 3. Sequencing 87 | 88 | Axes of variation 89 | 90 | - Single- versus paired-end 91 | - Length: 50-200nt 92 | - Number of reads per sample 93 | 94 | Application-specific, e.g., 95 | 96 | - ChIP-seq: short, single-end reads are usually sufficient 97 | - RNA-seq, known genes: single- or paired-end reads 98 | - RNA-seq, transcripts or novel variants: paired-end reads 99 | - Copy number: single- or paired-end reads 100 | - Structural variants: paired-end reads 101 | - Variants: depth via longer, paired-end reads 102 | - Microbiome: long paired-end reads (overlapping ends) 103 | 104 | ### 4. Alignment 105 | 106 | Alignment strategies 107 | 108 | - _de novo_ 109 | - No reference genome; considerable sequencing and computational 110 | resources 111 | - Genome 112 | - Established reference genome 113 | - Splice-aware aligners 114 | - Novel transcript discovery 115 | - Transcriptome 116 | - Established reference genome; reliable gene model 117 | - Simple aligners 118 | - Known gene / transcript expression 119 | 120 | Splice-aware aligners (and _Bioconductor_ wrappers) 121 | 122 | - [Bowtie2][] (`r Biocpkg("Rbowtie")`) 123 | - [STAR][] ([doi](https://doi.org/10.1093/bioinformatics/bts635)) 124 | - [GMAP/GSNAP][] (`r Biocpkg("gmapR")`) 125 | - subread ([doi](https://doi.org/10.1093/nar/gkt214)) 126 | (`r Biocpkg("Rsubread")`) 127 | - Systematic evaluation (Engstrom et al., 2013, 128 | [doi](https://doi.org/10.1038/nmeth.2722)) 129 | 130 | ### (5a. Bowtie2 / tophat / Cufflinks / Cuffdiff) 131 | 132 | - [tophat][] uses [Bowtie2][] to perform basic single- and paired-end 133 | alignments, then uses algorithms to place difficult-to-align reads 134 | near to their well-aligned mates. 135 | - [Cufflinks][] ([doi](https://doi.org/10.1038/nprot.2012.016)) 136 | takes _tophat_ output and estimate existing and novel transcript 137 | abundance. 138 | [How Cufflinks Works](http://cufflinks.cbcb.umd.edu/howitworks.html) 139 | - [Cuffdiff][] assesses statistical significance of estimated 140 | abundances between experimental groups 141 | 142 | ### 5. Reduction to 'count tables' 143 | 144 | - Use known gene model to count aligned reads overlapping regions of 145 | interest / gene models 146 | - Gene model can be public (e.g., UCSC, NCBI, ENSEMBL) or _ad hoc_ (gff file) 147 | - `GenomicAlignments::summarizeOverlaps()` 148 | - [HTSeq](http://www-huber.embl.de/users/anders/HTSeq/doc/overview.html), 149 | [htseq-count](http://www-huber.embl.de/users/anders/HTSeq/doc/count.html) 150 | 151 | ### Step 6. Analysis 152 | 153 | Summarization 154 | 155 | - Counts _per se_, rather than a summary (RPKM, FRPKM, ...), are 156 | relevant for analysis 157 | - For a given gene, larger counts imply more information; RPKM etc., 158 | treat all estimates as equally informative. 159 | - Comparison is across samples at _each_ region of interest; all 160 | samples have the same region of interest, so modulo library size 161 | there is no need to correct for, e.g., gene length or mapability. 162 | 163 | Normalization 164 | 165 | - Libraries differ in size (total counted reads per sample) for 166 | un-interesting reasons; we need to account for differences in 167 | library size in statistical analysis. 168 | - Total number of counted reads per sample is _not_ a good estimate of 169 | library size. It is un-necessarily influenced by regions with large 170 | counts, and can introduce bias and correlation across 171 | genes. Instead, use a robust measure of library size that takes 172 | account of skew in the distribution of counts (simplest: trimmed 173 | geometric mean; more advanced / appropriate encountered in the lab). 174 | - Library size (total number of counted reads) differs between 175 | samples, and should be included _as a statistical offset_ in 176 | analysis of differential expression, rather than 'dividing by' the 177 | library size early in an analysis. 178 | 179 | Appropriate error model 180 | 181 | - Count data is _not_ distributed normally or as a Poisson process, 182 | but rather as negative binomial. 183 | - Result of a combination Poisson (`shot' noise, i.e., within-sample 184 | technical and sampling variation in read counts) with variation 185 | between biological samples. 186 | - A negative binomial model requires estimation of an additional 187 | parameter ('dispersion'), which is estimated poorly in small 188 | samples. 189 | - Basic strategy is to moderate per-gene estimates with more robust 190 | local estimates derived from genes with similar expression values (a 191 | little more on borrowing information is provided below). 192 | 193 | Pre-filtering 194 | 195 | - Naively, a statistical test (e.g., t-test) could be applied to each 196 | row of a counts table. However, we have relatively few samples 197 | (10's) and very many comparisons (10,000's) so a naive approach is 198 | likely to be very underpowered, resulting in a very high _false 199 | discovery rate_ 200 | - A simple approach is perform fewer tests by removing regions that 201 | could not possibly result in statistical significance, regardless of 202 | hypothesis under consideration. 203 | - Example: a region with 0 counts in all samples could not possibly be 204 | significant regradless of hypothesis, so exclude from further 205 | analysis. 206 | - Basic approaches: 'K over A'-style filter -- require a minimum of A 207 | (normalized) read counts in at least K samples. Variance filter, 208 | e.g., IQR (inter-quartile range) provides a robust estimate of 209 | variability; can be used to rank and discard least-varying regions. 210 | - More nuanced approaches: `r Biocpkg("edgeR")` vignette; work flow 211 | today. 212 | 213 | Borrowing information 214 | 215 | - Why does low statistical power elevate false discovery rate? 216 | - One way of developing intuition is to recognize a t-test (for 217 | example) as a ratio of variances. The numerator is 218 | treatment-specific, but the denominator is a measure of overall 219 | variability. 220 | - Variances are measured with uncertainty; over- or under-estimating 221 | the denominator variance has an asymmetric effect on a t-statistic 222 | or similar ratio, with an underestimate _inflating_ the statistic 223 | more dramatically than an overestimate deflates the statistic. Hence 224 | elevated false discovery rate. 225 | - Under the typical null hypothesis used in microarray or RNA-seq 226 | experiments, each gene may respond differently to the treatment 227 | (numerator variance) but the overall variability of a gene is 228 | the same, at least for genes with similar average expression 229 | - The strategy is to estimate the denominator variance as the 230 | between-group variance for the gene, _moderated_ by the average 231 | between-group variance across all genes. 232 | - This strategy exploits the fact that the same experimental design 233 | has been applied to all genes assayed, and is effective at 234 | moderating false discovery rate. 235 | 236 | ### Step 7. Comprehension 237 | 238 | Placing differentially expressed regions in context 239 | 240 | - Gene names associated with genomic ranges 241 | - Gene set enrichment and similar analysis 242 | - Proximity to regulatory marks 243 | - Integrate with other analyses, e.g., methylation, copy number, 244 | variants, ... 245 | 246 | ![Copy number / expression QC](our_figures/copy_number_QC_2.png) 247 | Correlation between genomic copy number and mRNA expression 248 | identified 38 mis-labeled samples in the TCGA ovarian cancer 249 | Affymetrix microarray dataset. 250 | 251 | ## Lab 252 | 253 | [The lab](B02.1.1_RNASeqLab.html) is based on a modified version of the 254 | RNA-seq work flow developed by Michael Love, Simon Anders, Wolfgang 255 | Huber. 256 | 257 | 258 | [Bowtie2]: http://bowtie-bio.sourceforge.net/bowtie2/index.shtml 259 | [tophat]: http://ccb.jhu.edu/software/tophat/index.shtml 260 | [Cufflinks]: http://cufflinks.cbcb.umd.edu/ 261 | 262 | [RSEM]: http://deweylab.biostat.wisc.edu/rsem/ 263 | [STAR]: https://github.com/alexdobin/STAR 264 | [GMAP/GSNAP]: http://research-pub.gene.com/gmap/ 265 | -------------------------------------------------------------------------------- /vignettes/B02.2.3_CopyNumber.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | options(error=traceback) 9 | library(LearnBioconductor) 10 | set.seed(123L) 11 | stopifnot(BiocInstaller::biocVersion() == "3.0") 12 | ``` 13 | 14 | ```{r style, echo = FALSE, results = 'asis'} 15 | BiocStyle::markdown() 16 | knitr::opts_chunk$set(tidy=FALSE) 17 | ``` 18 | 19 | # Copy number work flow 20 | 21 | Sonali Arora, Martin Morgan
22 | October 28, 2014 23 | 24 | Copy Number Variation (CNV's) refers to the duplication or deletion of 25 | DNA segments larger than 1 kb. CNV's are structural variations in the 26 | genome which range in length between 50 bp and 1 Mbp. They are 27 | widespread among humans - on an average 12 CNVs exist per individual 28 | in comparison to the reference genome. They have also been shown to 29 | play a role in diseases such as autism, breast cancer, obesity, 30 | Alzheimer’s disease and schizophrenia among other diseases. 31 | 32 | ## Experimental design 33 | 34 | Like any other genomic analysis, before we start a copy Number 35 | analysis, we need to consider experimental design. Here, we highlight 36 | two specific pointers that one needs to keep in mind while designing a 37 | copy number analysis. 38 | 39 | **Tumor and normal samples** Are we planning to just find copy number 40 | profile in individuals? For example, how does the copy number profile 41 | for a region evolve over a certain period of time? (Here we are 42 | comparing the copy number profile to a reference genome) 43 | 44 | Are we planning to compare copy number profiles from tumor vs normal 45 | profiles? For example, we may be trying to find out if copy number 46 | changes are responsible for a certain form of cancer and want to find 47 | the exact genomic region against which a treatment can be developed. 48 | 49 | There are different packages and different functions inside the same 50 | package which handle CNV for tumor and normal samples or CNV in 51 | samples. 52 | 53 | **Germ line versus somatic CNV** Germ line CNV are relatively short 54 | (a few bp to a few Mbp) copy number changes that the individual 55 | inherits from one of the two parental gametes and thus are typically 56 | present in 100% of cells. 57 | 58 | Somatic CNV (often called CNA where A stands for alterations or aberration) are 59 | copy number changes of any size and amount (from a few bases to whole 60 | chromosomes) that happen (and often carry on happening) in cancer cells. 61 | Cancer cells can be aneuploid (that means they are largely triploid, 62 | tetraploid or even aploid) and can have high focal amplifications 63 | (some regions could have many copies: it is not unusual to have 8-12 copies for 64 | some regions). Furthermore, because tumor samples are typically an admixture 65 | of normal and cancer cells, the tumor purity in unknown and variable. 66 | 67 | Different algorithms make different assumptions while handling somatic 68 | or germ line CNV. Typically, germ line cnv caller can assume: 69 | 70 | - The genome is largely diploid. 71 | - The sample is pure and homogeneous. 72 | - Any gain or loss should be 50% move or 50% less coverage. 73 | 74 | For these reasons, the algorithms can focus more on associating 75 | _p_-values for each call; it is possible to estimate false positive 76 | and false negative rates. 77 | 78 | Somatic CNA callers cannot make any of the assumption above, or if 79 | they do, they have limited scope. 80 | 81 | ## Sequencing technology 82 | 83 | Some key questions when thinking about sequencing technology to use 84 | include: 85 | 86 | + What kind of sequencing data are we working with? 87 | 88 | Is it array CGH data, SNP array or next generation sequencing 89 | data? For example, the `r Biocpkg("CGHbase")` detects CNV's in 90 | array CGH data whereas the `r Biocpkg("seqCNA")` among others 91 | detects CNV's in high-throughput sequencing data. 92 | 93 | + Amount of genome being sequenced - whole genome vs exome? 94 | 95 | Are we looking for copy number across the whole genome or are we 96 | looking for copy number only in exomes? Again different packages 97 | handle different kind of sequencing data. For example, The 98 | _Bioconductor_ package `r Biocpkg("exomeCopy")` detects CNV in 99 | exome sequencing data whereas the _Bioconductor_ package 100 | `r Biocpkg("cn.mops")` detects CNV in whole genome sequencing data. 101 | 102 | + Which platform was the sequencing done on? 103 | 104 | A lot of packages detecting CNV on platform-specific data. 105 | For example, the `r Biocpkg("crlmm")` detects CNV's in Affymetrix SNP 5.0 106 | and 6.0 and Illumina arrays whereas the `r Biocpkg("CopyNumber450K")` 107 | detects them in Illumina 450k methylation microarrays. 108 | 109 | + What is the coverage of sequenced data? 110 | 111 | Most packages work well with high coverage sequencing data, but 112 | some packages are designed to work well with low coverage data. It 113 | is best to recognize how coverage of our data will affect the 114 | choice of the package we use for our analysis at an early stage. 115 | 116 | ## Copy number analysis algorithm? 117 | 118 | Since statistics plays a huge role in copy number analysis, we should 119 | also spend some time in thoroughly understanding the underlying 120 | algorithm of the _R_ package being used. A few questions to consider 121 | while choosing a package would be - 122 | 123 | 1. How is our chosen package binning and counting reads? 124 | 125 | 2. Is any pre-processing required from our end? Is it trimming aligned 126 | reads internally? 127 | 128 | 3. What segmentation algorithm is being used ? For example, does the 129 | package use Circular Binary Segmentation, HMM based methods etc. 130 | 131 | 4. How efficiently can it handle big data? Do I need additional 132 | computational resources to run the analysis? Does the function run 133 | in parallel? 134 | 135 | ## Available resources in _Bioconductor_ 136 | 137 | _Bioconductor_ currently has about 41 packages for Copy Number 138 | Analysis. To find these, one can visit the 139 | [biocViews](http://bioconductor.org/packages/devel/BiocViews.html#___Software.) 140 | page and type "CopyNumberVariation" in the "Autocomplete biocViews 141 | search" 142 | 143 | ## Workflow using _cn.mops_ 144 | 145 | Lets work through a small example to illustrate how straight-forward a 146 | copy number analysis can be once you've figured out all the 147 | logistics. We will also find the genes that lie within the detected 148 | copy number regions. 149 | 150 | For this analysis, I chose the `r Biocpkg("cn.mops")` package as it 151 | helps us with 152 | 153 | - Detecting germ-line CNVs 154 | - Works well with low coverage data 155 | - Handles both single copy number analysis and tumor vs normal copy 156 | number analysis 157 | - Uses parallel processing internally so we get fast computation 158 | - Handles whole genome sequencing data 159 | - Supports _GenomicRanges_ infrastructure which allows easy workflows 160 | with other Bioconductor packages. 161 | 162 | We start by downloading relevant files, if necessary 163 | 164 | ```{r cvn-setup-1, echo=FALSE} 165 | destdir <- "~/bigdata" 166 | if (!file.exists(destdir)) 167 | dir.create(destdir) 168 | ``` 169 | 170 | ```{r cnv-setup-2, message=FALSE} 171 | ## set path/to/download/directory, e.g., 172 | ## destdir <- "~/bam/copynumber" 173 | stopifnot(file.exists(destdir)) 174 | 175 | bamFiles <- file.path(destdir, 176 | c("tumorA.chr4.bam", "normalA.chr4.bam")) 177 | urls <- paste0("http://s3.amazonaws.com/copy-number-analysis/", 178 | basename(bamFiles)) 179 | for (i in seq_along(bamFiles)) 180 | if (!file.exists(bamFiles[i])) { 181 | download.file(urls[i], bamFiles[i]) 182 | download.file(paste0(urls[i], ".bai"), paste0(bamFiles[i], ".bai")) 183 | } 184 | ``` 185 | 186 | The main work flow 1) loads the library; 2) counts reads; 3) 187 | normalizes counts; 4) detects CNVs; and 5) visualizes results. 188 | 189 | ```{r cnv-workflow, message=FALSE} 190 | ## 1. Load the cn.mops package 191 | suppressPackageStartupMessages({ 192 | library(cn.mops) 193 | }) 194 | 195 | ## 2. We can bin and count the reads 196 | reads_gr <- getReadCountsFromBAM(BAMFiles = bamFiles, 197 | sampleNames = c("tumor", "normal"), 198 | refSeqName = "chr4", WL = 10000, mode = "unpaired") 199 | 200 | ## 3. Noramlization 201 | ## We need a special normalization because the tumor has many large CNVs 202 | X <- normalizeGenome(reads_gr, normType="poisson") 203 | 204 | ## 4. Detect cnv's 205 | ref_analysis <- referencecn.mops(X[,1], X[,2], 206 | norm=0, 207 | I = c(0.025, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 8, 16, 32, 64), 208 | classes = paste0("CN", c(0:8, 16, 32, 64, 128)), 209 | segAlgorithm="DNAcopy") 210 | resCNMOPS <- calcIntegerCopyNumbers(ref_analysis) 211 | 212 | ## 5. Visualize the cnv's 213 | segplot(resCNMOPS) 214 | ``` 215 | 216 | Here the x-axis represents the genomic position and the y-axis 217 | represents the log ratio of read counts and copy number call of each 218 | segment (red) 219 | 220 | ```{r cnv-regions} 221 | human_cn <- cnvr(resCNMOPS) 222 | human_cn 223 | ``` 224 | 225 | To find the genes that lie in these copy number regions, we will use 226 | the _TranscriptDb_ object for hg19 227 | 228 | ```{r cnv-annotate-txdb, message=FALSE} 229 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 230 | ## subset to work with only chr4 231 | txdb <- keepSeqlevels(TxDb.Hsapiens.UCSC.hg19.knownGene, "chr4") 232 | genes0 <- genes(txdb) 233 | ## 'unlist' so that each range is associated with a single gene identifier 234 | idx <- rep(seq_along(genes0), elementLengths(genes0$gene_id)) 235 | genes <- granges(genes0)[idx] 236 | genes$gene_id = unlist(genes0$gene_id) 237 | ``` 238 | 239 | Next we will use find overlaps to assign gene identifiers to cnv 240 | regions. 241 | 242 | ```{r cnv-annotate} 243 | olaps <- findOverlaps(genes, human_cn, type="within") 244 | idx <- factor(subjectHits(olaps), levels=seq_len(subjectLength(olaps))) 245 | human_cn$gene_ids <- splitAsList(genes$gene_id[queryHits(olaps)], idx) 246 | human_cn 247 | ``` 248 | 249 | ## Session info 250 | 251 | The packages and versions used in this work flow are as follows: 252 | 253 | ```{r cvn-session-info} 254 | restoreSeqlevels(TxDb.Hsapiens.UCSC.hg19.knownGene) 255 | sessionInfo() 256 | ``` 257 | -------------------------------------------------------------------------------- /vignettes/B02.2_CommonWorkFlows.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | knitr::opts_chunk$set(tidy=FALSE) 15 | ``` 16 | 17 | # Common Sequence Analysis Work Flows 18 | 19 | Martin Morgan, Sonali Arora
20 | October 28, 2014 21 | 22 | ## RNA-seq 23 | 24 | See the [lecture notes](B02.1_RNASeq.html) and [lab](B02.1_RNASeqLab.html). 25 | 26 | RNA-seq differential expression of known _genes_ 27 | 28 | - Simplest scenario 29 | - Experimental design: simple, replicated; track covariates and be 30 | aware of batch effects 31 | - Sequencing: moderate length and number of reads; single or 32 | paired-end (though probably paired-end). 33 | - Alignment: basic splice-aware aligner, e.g., _Bowtie2_, 34 | _STAR_. Viable _Bioconductor_ approaches: `r Biocpkg("Rsubread")`, 35 | `r Biocpkg("Rbowtie")` (especially via the `r Biocpkg("QuasR")` 36 | package). 37 | - Reduction: `GenomicRanges::summarizeOverlaps()` or external tools, 38 | using gene model from `TxDb.*` package or GFF / GTF files. End 39 | result: matrix of counts. 40 | - Analysis: `r Biocpkg("DESeq2")`, `r Biocpkg("edgeR")`, and 41 | additional software. 42 | 43 | RNA-seq differential expression of known and novel _transcripts_ 44 | 45 | - Popular non-_R_ work flow: _Rbowtie2_, _tophat_, _cufflinks_, _cuffdiff_. 46 | - _Biocondutor_ options 47 | 48 | - `r Biocpkg("DEXSeq")`: differential _exon_ use. 49 | - `Rsubread::subjunc()` for aligning without requiring known gene models. 50 | - `r Biocpkg("cummeRbund")`: working with _cufflinks_ output. 51 | 52 | Single-cell expression 53 | 54 | - `r Biocpkg("monocle")` 55 | 56 | ## ChIP-seq 57 | 58 | See my recent 59 | [slides](http://bioconductor.org/help/course-materials/2014/CSAMA2014/4_Thursday/lectures/ChIPSeq_slides.pdf) 60 | outlining ChIP-seq and relevant _Bioconductor_ software. 61 | 62 | - Experimental design / wet lab: important to effectively enrich 63 | genomic DNA via ChIP, otherwise hard to distinguish signal peaks from background 64 | - Sequencing: moderate length and number of single-end reads very adequate. 65 | - Alignment: Basic aligners sufficient 66 | - Reduction 67 | 68 | - External software; many tools depending on application, e.g., _MACS_. 69 | - Product: BED and / or WIG files of called peaks 70 | 71 | - Analysis & Comprehension 72 | 73 | - `r Biocpkg("ChIPQC")` for quality control. 74 | - `r Biocpkg("rtracklayer")` to input BED and WIG files to 75 | standard _Bioconductor_ data structures. 76 | - `r Biocpkg("ChIPpeakAnno")`, `r Biocpkg("ChIPXpres")` for 77 | annotating peaks in relation to genes. 78 | - `r Biocpkg("DiffBind")` to assess differential representation of 79 | peaks in a designed experiment. 80 | - `r Biocpkg("AnnotationHub")` for accessing (some) 81 | consortium-level summary data. 82 | 83 | ## Copy Number 84 | 85 | See the [Copy Number Workflow](./B02.2.3_CopyNumber.html) document. 86 | 87 | ## Variants 88 | 89 | See Michael Lawrence's variant calling with 90 | [VariantTools](http://bioconductor.org/help/course-materials/2014/BioC2014/Lawrence_Tutorial.pdf). 91 | and Val Obenchain's manipulation and annotation of called variants with 92 | [VariantAnnotation](http://bioconductor.org/help/workflows/variants/). 93 | 94 | - Sequencing: requires high-quality reads with high per-nucleotide 95 | depth of coverage -- longer, paired-end sequencing. 96 | - Alignment: requires effective aligners; _BWA_, _GMAP_, ... 97 | 98 | - `r Biocpkg("gmapR")` wraps the GMAP aligner in _R_. 99 | 100 | - Reduction: typically to VCF files summarizing variants and / or 101 | population-level variation. _GATK_ and other non-_R_ tools commonly 102 | used. 103 | 104 | - `r Biocpkg("VariantTools")` includes facilities for calling 105 | variants. 106 | - `r Biocpkg("h5vc")` targets a different intermediate step: 107 | summarize base counts at each position in the genome; use this 108 | as a starting point for calling variants, and to evaluate false 109 | positives, etc. 110 | 111 | - Analysis & comprehension 112 | 113 | - `r Biocpkg("VariantAnnotation")`, `r Biocpkg("ensemblVEP")` for 114 | querying / inputting VCF files, and for annotation of variants 115 | ("is this a coding variant?", etc.). 116 | - `r Biocpkg("SomaticSignatures")` for working with somatic 117 | signatures of single-nucleotide variants. 118 | 119 | ## Epigenomics 120 | 121 | See the short 122 | [introduction](http://bioconductor.org/help/course-materials/2014/Epigenomics/MethylationArrays.html) 123 | and 124 | [lab](http://bioconductor.org/help/course-materials/2014/Epigenomics/MethylationArrays-lab.html) 125 | centered around Illumina 450k methylation arrays and the `r Biocpkg("minfi")` package. 126 | 127 | - Analysis & comprehension: `r Biocpkg("bsseq")`, `r Biocpkg("BiSeq")` 128 | for processing and analysis; `r Biocpkg("bumphunter")` as basic tool 129 | for identifying CpG features. 130 | 131 | ## Microbiome 132 | 133 | - Experimental design: typically population-level surveys with 134 | moderate (10's-100's) of samples. 135 | - Wet lab & sequencing: often target phylogenetically-informative 136 | genes, requiring longer (overlapping) paired-end reads. Many 137 | existing studies used 454 technology, which has a different 138 | sequencing error model than Illumina (e.g., homopolymers are a 139 | common error, instead of trailing nucleotide quality deterioration). 140 | - Reduction: Pre-processing (e.g., knitting together overlapping 141 | paired-end reads) and taxonomic classification / placement in 142 | third-party software, e.g., _QIIME_, _pplacer_. End result: count 143 | table summarizing represenation of distinct taxa in each sample. 144 | 145 | - `r Biocpkg("rRDP")` provides an _R_ / _Bioconductor_ interface 146 | to the RDP classifiere. 147 | 148 | - Analysis: _R_ / _Bioconductor_ and many insights from microarray / 149 | RNA-seq analysis well suited to count table, but common pipelines 150 | have re- or dis-invented the wheel. 151 | 152 | - `r Biocpkg("phyloseq")` provides very nice tools for general 153 | analysis. 154 | -------------------------------------------------------------------------------- /vignettes/B02.3_MachineLearning.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r style, echo = FALSE, results = 'asis'} 8 | BiocStyle::markdown() 9 | knitr::opts_chunk$set(tidy=FALSE) 10 | ``` 11 | 12 | ```{r setup, echo=FALSE} 13 | library(LearnBioconductor) 14 | library(xtable) 15 | stopifnot(BiocInstaller::biocVersion() == "3.0") 16 | ``` 17 | 18 | # Machine Learning 19 | 20 | Sonali Arora
21 | 28 October, 2014 22 | 23 | ## Introduction 24 | 25 | Statistical Learning plays an important role in Genomics. A few example of 26 | learning problems would be - 27 | * Identify the risk factors(genes) for prostrate cancer based on gene 28 | expression data 29 | * Predict the chances of breast cancer survival in a patient. 30 | * Identify patterns of gene expression among different sub types of 31 | breast cancer 32 | 33 | Typically, we have an outcome measurement, usually quantitative (such as gene 34 | expression) or categorical (such as breast cancer /no breast cancer), that we 35 | wish to predict based on a set of features (such as diet and clinical 36 | measurements). We have a training set of data, in which we observe the outcome 37 | and feature measurements for a set of objects (such as people). With this data 38 | we build a prediction model, or a learner, which will enable us to predict 39 | the outcome for new unseen objects. A good learner is one that accurately 40 | predicts such an outcome. This is called _supervised_ learning because the 41 | outcome variable guides the learning process. In the _unsupervised_ learning 42 | problem, have no measurements of the outcome and observe only the features. 43 | Our task is to describe how the data are organized or clustered. 44 | 45 | ## Machine Learning with Bioconductor : Resources Overview 46 | 47 | Bioconductor houses a lot of R packages which provide machine learning tools for 48 | different kinds of Genomic Data analysis. It has had many releases since its 49 | inception. The latest cutting edge development happens in the "devel" version. 50 | One can find R packages specially catered to Machine Learning for Genomic 51 | Analysis by visiting the following link and using 52 | [biocViews](http://www.bioconductor.org/packages/devel/BiocViews.html#___Software). 53 | 54 | For example, one can type "Clustering" in the "Autocomplete biocViews search" 55 | box and find that there are 89 packages in Bioconductor which provide 56 | "Clustering" functionality for Genomic Data. On the same page, under 57 | "StatisticalMethod" one can find R packages for other machine learning 58 | techniques - 59 | 60 | ```{r echo=FALSE, results='asis'} 61 | biocView_df <- data.frame( technique=c("Bayesian","Classification","Clustering" 62 | ,"DecisionTree","NeuralNetwork","SupportVectorMachines","DimensionReduction", 63 | "HiddenMarkovModel","Regression","PrincipalComponent"), packages = 64 | as.integer(c(15, 64, 89, 7, 1, 1, 2, 4, 7, 4))) 65 | print(xtable(biocView_df), type="html", comment=FALSE) 66 | ``` 67 | 68 | ## Data for this Lab 69 | 70 | We will use two data sets in this lab - one for supervised learning and one for 71 | unsupervised learning. 72 | 73 | For **Unsupervised learning**, we will use the NCI60 cancer cell 74 | microarray data which contains 6830 gene expression measurements of 64 cancer 75 | cell lines. We dropped the sub types which contain only 1 sample ( there were 76 | about 6 of them) and created a SummarizedExperiment object. 77 | 78 | The *SummarizedExperiment* class is a matrix-like container where rows 79 | represent ranges of interest (as a GRanges or GRangesList-class) and columns 80 | represent samples (with sample data summarized as a DataFrame-class). A 81 | SummarizedExperiment contains one or more assays, each represented by a 82 | matrix-like object of numeric or other mode. 83 | 84 | ![Summarized Experiment](our_figures/SummarizedExperiment.png) 85 | 86 | This object has been made available to you, you can simply read it in using 87 | 88 | ```{r message=FALSE} 89 | library(GenomicRanges) 90 | sefile <- system.file("extdata", "NCI60.Rda", package="LearnBioconductor") 91 | load(sefile) 92 | nci60data <- t(assay(NCI60)) 93 | ncilabels <- colData(NCI60) 94 | ``` 95 | The gene expression data is stored in "assay" whereas the labels are stored in 96 | colData. 97 | 98 | For **Supervised learning**, we will use cervical count data from the 99 | Biocoductor package, `r Biocpkg("MLSeq")`. This data set contains 100 | expressions of 714 miRNA's of human samples. There are 29 tumor and 29 101 | non-tumor cervical samples. For learning purposes, we can treat these 102 | as two separate groups and run various classification algorithms. 103 | 104 | ```{r message=FALSE} 105 | library(MLSeq) 106 | filepath = system.file("extdata/cervical.txt", package = "MLSeq") 107 | cervical = read.table(filepath, header = TRUE) 108 | 109 | ``` 110 | 111 | ### Exercise 112 | 113 | Download the NCI60 [data](http://statweb.stanford.edu/~tibs/ElemStatLearn/) 114 | from the given url and create a SummarizedExperiment 115 | object filtering out subtypes which contain only 1 sample. 116 | 117 | ## Unsupervised Learning 118 | 119 | Unsupervised Learning is a set of statistical tools intended for the setting 120 | in which we have only a set of features $X_1$, $X_2$, ....,$X_p$ measured 121 | on 'n' observations. We are primarily interested in discovering interesting 122 | things on the measurement $X_1$, $X_2$, ....,$X_p$ 123 | 124 | Unsupervised Learning is often performed as a part of Exploratory Data Analysis. 125 | These tools help us to get a good idea about the data set. Unlike a supervised 126 | learning problem, where we can use prediction to gain some confidence about our 127 | learning algorithm, there is no way to check our model. The learning algorithm 128 | is thus, aptly named "unsupervised". 129 | 130 | Nevertheless, it offers great insights for our given problem. For Example, in 131 | case of our NCI60 data set, we may want to look for subgroups among our cancer 132 | samples or among genes to better understand how the disease works. 133 | 134 | ### PCA 135 | 136 | One of the classic things that we are told as soon as we get our hands on is to 137 | make a scatter plot and visualize the data. With big data, this is not always 138 | very easy to do. For example, in our NCI60 data set we would have about 139 | ( 6830 C 2) = 6830(6830-1)/2= 23321035 scatter plots ! It is certainly not 140 | possible to look at all of them and since they each contain only a small amount 141 | of the total information present in the data set, none of them will be 142 | informative. Thus, we need a low dimensional representation of data which 143 | captures as much variation as possible. 144 | 145 | *Principal components* allow us to summarize the data with a smaller number of 146 | variables that explain the variability in the data set. The process by which 147 | these principal components are computed and the usage of these components in 148 | understanding the data is known as *Principal Componenet Analysis*. 149 | 150 | Let us use PCA to visualize the NCI60 data set. We first perform PCA on the 151 | data after scaling the variables (genes) to have standard deviation one and 152 | plot the first few principal components in order to visualize the data. As you 153 | can see from the following figure, with big data, biplot() is not very 154 | informative. 155 | 156 | ```{r} 157 | pcaRes <- prcomp(nci60data, scale=TRUE) 158 | biplot(pcaRes) 159 | ``` 160 | So lets look at the first 3 components and see if we can gain some interesting 161 | insights from them. 162 | 163 | ```{r fig.width=12} 164 | # make colors as factors. 165 | labs <- as.character(unlist(as.list(ncilabels))) 166 | 167 | cellColor <- function(vec) 168 | { 169 | uvec <- unique(vec) 170 | cols = rainbow(length(uvec)) 171 | colvec <- cols[as.numeric(as.factor(vec))] 172 | list(colvec=colvec, cols=cols, labels= uvec) 173 | } 174 | 175 | par(mfrow=c(1,2)) 176 | 177 | colres <- cellColor(labs) 178 | 179 | plot(pcaRes$x[,1:2],col=colres$colvec, xlab = "z1", ylab="z2", pch=19) 180 | legend("bottomright", legend = colres$labels, text.col = colres$cols, 181 | bty="n", cex=0.80) 182 | plot(pcaRes$x[,c(1,3)], col=colres$colvec, xlab="z1", ylab="z3", pch=19) 183 | legend("topright", legend = colres$labels,text.col = colres$cols, 184 | bty ="n" , cex=0.80) 185 | 186 | par(mfrow=c(1,1)) 187 | 188 | ``` 189 | 190 | Overall, we can conclude that the cell lines corresponding to a single cancer 191 | type tend to have similar values on the first few principal component score 192 | vectors. This indicates that cell lines from the same cancer type tend to have 193 | similar gene expression levels. 194 | 195 | ### Clustering observations 196 | 197 | In hierarchical clustering, we start by defining some dissimilarity measure 198 | between each pair of observations (like Euclidean distance). We start at the 199 | bottom of the dendrogram, each of the n observations are considered as a 200 | cluster, the two clusters that are most similar to each other are fused together 201 | so now we have n-1 clusters. Next the two clusters that are similar together are 202 | fused so that there are n-2 clusters. This algorithm proceeds iteractively until 203 | all samples belong to one single cluster and the dendrogram is complete. 204 | 205 | We define the dissimilarity between two clusters or two groups of observations 206 | using Linkage. There are four common types of linkage - "average", "complete", 207 | "single", "centroid". Assuming we have two clusters A and B, then - 208 | 209 | 1. _Complete_ refers to recording the *largest* of all pairwise dissimilarities 210 | between observations in cluster A and observations in Cluster B. 211 | 2. _Single_ refers to recording the *smallest* of all pairwise dissimilarities 212 | between observations in cluster A and observations in Cluster B. 213 | 3. _Single_ refers to recording the *average* of all pairwise dissimilarities 214 | between observations in cluster A and observations in Cluster B. 215 | 4. _Centroid_ refers to the dissimilarity between the centroid for cluster A and 216 | the centroid for cluster B. 217 | 218 | Usually, the dissimilarity measure is the Euclidean Distance. 219 | 220 | Lets cluster the observations using complete linkage with Euclidean distance 221 | as the dissimilarity measure in "complete linkage". 222 | 223 | ```{r fig.width=12 , message=FALSE} 224 | library(dendextend) 225 | 226 | sdata <- scale(nci60data) 227 | d <- dist(sdata) 228 | labs <- as.character(unlist(as.list(ncilabels))) 229 | comp_clust <- hclust(d) 230 | dend <- as.dendrogram(comp_clust) 231 | leaves <- labs[order.dendrogram(dend)] 232 | labels_colors(dend, labels=TRUE) <- cellColor(leaves)$colvec 233 | labels(dend) <- leaves 234 | plot(dend, main ="Clustering using Complete Linkage") 235 | ``` 236 | 237 | ### Exercise 238 | 239 | 1. Perform hierarchical clustering using average and single linkage. 240 | 2. Interpret the difference in the dendrograms. 241 | 3. Can you observe some patterns from these dendrograms? (hint: use cutree) 242 | 243 | **Solutions:** 244 | 245 | 1. The plots can be made with the following code - 246 | ```{r fig.width=12, fig.height=6} 247 | plot(hclust(d, method="average"), labels= labs, 248 | main ="Clustering using Average Linkage" , xlab="", ylab="" ) 249 | plot(hclust(d, method="single"), labels= labs, 250 | main ="Clusteringg using Single Linkage" , xlab="", ylab="" ) 251 | ``` 252 | 253 | 254 | 2. Briefly, complete linkage provides maximal inter cluster dissimilarity, single 255 | linkage results in minimal inter cluster dissimilarity and average results in 256 | mean inter cluster dissimilarity. We see that the results are affected by the 257 | choice of the linkage. Single linkage tend to yield trailing clusters while 258 | complete and average linkage leads to more balanced and attractive clusters. 259 | 260 | 3. For our example, we see that the cell lines within a single cancer cell type do 261 | not cluster together. But if we consider complete linkage and cut the tree at 262 | height=4 ( we tried different heights to observe patterns) we observe some clear 263 | patterns like the leukemia cell lines fall in cluster 2 and the breast cancer 264 | cell lines spread out. 265 | ```{r} 266 | hc <- cutree(comp_clust, 4) 267 | table(hc, labs) 268 | ``` 269 | 270 | ## Supervised Learning 271 | 272 | In supervised learning, along with the features $X_1$, $X_2$, ....,$X_p$, we 273 | also have the a response Y measured on the same n observations. The goal is then 274 | to predict Y using $X_1$, $X_2$, ....,$X_p$ for new observations. 275 | 276 | ### A simple example using knn 277 | 278 | For the cervical data, we know that the first 29 are non-Tumor samples whereas 279 | the last 29 are Tumor samples. We will code these as 0 and 1 respectively. 280 | 281 | ```{r} 282 | class = data.frame(condition = factor(rep(c(0, 1), c(29, 29)))) 283 | ``` 284 | 285 | Lets look at one of the most basic supervised learning techniques 286 | **k-Nearest Neighbor** and see what all goes into building a simple model with 287 | it. For the sake of simplicity, we will use only 2 predictors (so that we can 288 | represent the data in 2 dimensional space) 289 | 290 | ```{r} 291 | data <- t(cervical) 292 | data <- data[,1:2] 293 | df <- cbind(data, class) 294 | colnames(df) <- c("x1","x2","y") 295 | rownames(df) <- NULL 296 | head(df) 297 | ``` 298 | 299 | This is how the data looks - 300 | ```{r fig.width=12} 301 | plot(df[,"x1"], df[,"x2"], xlab="x1", ylab="x2", 302 | main="data representation for knn", 303 | col=ifelse(as.character(df[,"y"])==1, "red","blue")) 304 | ``` 305 | 306 | Given a observation $x_0$ and a positive integer, K, the KNN classifier first 307 | identifies K points in the training data that are closest to $x_0$, represented 308 | by $N_0$. It estimates the conditional probability for class j as a fraction of 309 | $N_0$ and applies Bayes rule to classify the test observation to the class 310 | with the largest probability. 311 | More concretely, if k=3 and there are 2 observation belonging to class 1 and 1 312 | observation belonging to class 2, then we the test observation is assigned to 313 | class1. 314 | 315 | ![knn_figure](our_figures/knn.png) 316 | 317 | For all supervised experiments its a good idea to hold out some data as 318 | _Training Data_ and build a model with this data. We can then test the built 319 | model using the left over data (_Test Data_) to gain confidence in our model. 320 | We will randomly sample 30% of our data and use that as a test set. The 321 | remaining 70% of the data will be used as training data 322 | 323 | ```{r } 324 | set.seed(9) 325 | nTest = ceiling(ncol(cervical) * 0.2) 326 | ind = sample(ncol(cervical), nTest, FALSE) 327 | 328 | cervical.train = cervical[, -ind] 329 | cervical.train = as.matrix(cervical.train + 1) 330 | classtr = data.frame(condition = class[-ind, ]) 331 | 332 | cervical.test = cervical[, ind] 333 | cervical.test = as.matrix(cervical.test + 1) 334 | classts = data.frame(condition = class[ind, ]) 335 | ``` 336 | 337 | Training set error is the proportion of mistakes made if we apply our model to 338 | the training data and Test set error is the proportion of mistakes made when 339 | we apply our model to test data. 340 | 341 | For different neighbors , let us calculate the training error and test error 342 | using KNN. 343 | 344 | ```{r message=FALSE} 345 | library(class) 346 | 347 | newknn <- function( testset, trainset, testclass, trainclass, k) 348 | { 349 | pred.train <- knn.cv(trainset, trainclass, k=k) 350 | pred.test <- knn(trainset, testset, trainclass, k=k) 351 | 352 | test_fit <- length(which(mapply(identical, as.character(pred.test), 353 | testclass)==FALSE))/length(testclass) 354 | 355 | train_fit <- length(which(mapply(identical, as.character(pred.train), 356 | trainclass)==FALSE))/length(trainclass) 357 | 358 | c(train_fit=train_fit, test_fit= test_fit) 359 | } 360 | 361 | trainset <- t(cervical.train) 362 | testset <- t(cervical.test) 363 | testclass <- t(classts) 364 | trainclass <- t(classtr) 365 | klist <- 1:15 366 | ans <- lapply(klist, function(x) 367 | newknn(testset, trainset, testclass, trainclass,k =x)) 368 | 369 | resdf <- t(as.data.frame(ans)) 370 | rownames(resdf) <- NULL 371 | plot(klist, resdf[,"train_fit"], col="blue", type="b",ylim=c(range(resdf)), 372 | main="k Nearest Neighbors for Cervical Data", xlab="No of neighbors", 373 | ylab ="Training and Test Error") 374 | points(klist, resdf[,"test_fit"], col="red", type="b") 375 | legend("bottomright", legend=c("Training error","Test error"), 376 | text.col=c("blue","red"), bty="n") 377 | 378 | ``` 379 | 380 | Another important concept in machine learning is **cross validation**. Since 381 | samples are often scarse, it is often not possible to set aside a validation set 382 | ans then use it to assess the performance of our prediction model. So we use 383 | cross validation to train a better model. We start by dividing the training data 384 | randomly into n equal parts. The learning method is fit to n-1 parts of the 385 | data, and the prediction error is computed on the remaining part. This is done 386 | in turn for each 1/n parts of the data, and finally the n prediction error 387 | estimates are averaged. 388 | 389 | For example, K-fold cross validation where K=5 390 | 391 | ![cv_figure](our_figures/cross_validation.png) 392 | 393 | As you can see, computation for this very simple learner can become quite 394 | complicated. 395 | 396 | ### Fast classification using Bioconductor. 397 | 398 | MLSeq aims to make computation less complicated for a user and 399 | allows one to learn a model using various classifier's with one single function. 400 | 401 | The main function of this package is classify which requires data in the form of 402 | a DESeqDataSet instance. The DESeqDataSet is a subclass of SummarizedExperiment, 403 | used to store the input values, intermediate calculations and results of an 404 | analysis of differential expression. 405 | 406 | So lets create DESeqDataSet object for both the training and test set, and run 407 | DESeq on it. 408 | 409 | ```{r} 410 | cervical.trainS4 = DESeqDataSetFromMatrix(countData = cervical.train, 411 | colData = classtr, formula(~condition)) 412 | cervical.trainS4 = DESeq(cervical.trainS4, fitType = "local") 413 | 414 | cervical.testS4 = DESeqDataSetFromMatrix(countData = cervical.test, colData = classts, 415 | formula(~condition)) 416 | cervical.testS4 = DESeq(cervical.testS4, fitType = "local") 417 | 418 | ``` 419 | Classify using Support Vector Machines. 420 | 421 | ```{r} 422 | svm = classify(data = cervical.trainS4, method = "svm", normalize = "deseq", 423 | deseqTransform = "vst", cv = 5, rpt = 3, ref = "1") 424 | svm 425 | ``` 426 | 427 | It returns an object of class 'MLseq' and we observe that it successfully 428 | fitted a model with 97.8% accuracy. We can access the slots of this S4 object by 429 | ```{r} 430 | getSlots("MLSeq") 431 | ``` 432 | And also, ask about the model trained. 433 | 434 | ```{r} 435 | trained(svm) 436 | ``` 437 | 438 | We can predict the class labels of our test data using "predict" 439 | 440 | ```{r} 441 | pred.svm = predictClassify(svm, cervical.testS4) 442 | table(pred.svm, relevel(cervical.testS4$condition, 2)) 443 | ``` 444 | 445 | The other classification methods available are 'randomforest', 'cart' and 446 | 'bagsvm'. 447 | 448 | ### Exercise: 449 | 450 | Train the same training data and test data using randomForest. 451 | 452 | **Solutions:** 453 | 454 | ```{r} 455 | rf = classify(data = cervical.trainS4, method = "randomforest", 456 | normalize = "deseq", deseqTransform = "vst", cv = 5, rpt = 3, ref = "1") 457 | trained(rf) 458 | pred.rf = predictClassify(rf, cervical.testS4) 459 | table(pred.rf, relevel(cervical.testS4$condition, 2)) 460 | ``` 461 | 462 | ## SessionInfo 463 | 464 | ```{r} 465 | sessionInfo() 466 | ``` 467 | 468 | 469 | -------------------------------------------------------------------------------- /vignettes/B02.4_Annotation.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Annotating Genes, Genomes, and Variants 17 | 18 | Martin Morgan
19 | October 28, 2014 20 | 21 | ```{r pkgs, eval=TRUE, echo=FALSE, warning=FALSE, message=FALSE} 22 | options(max.print=1000) 23 | suppressPackageStartupMessages({ 24 | library(org.Hs.eg.db) 25 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 26 | library(BSgenome.Hsapiens.UCSC.hg19) 27 | library(GenomicRanges) 28 | library(biomaRt) 29 | library(rtracklayer) 30 | }) 31 | ``` 32 | 33 | ## Gene annotation 34 | 35 | ### Data packages 36 | 37 | Organism-level ('org') packages contain mappings between a central 38 | identifier (e.g., Entrez gene ids) and other identifiers (e.g. GenBank 39 | or Uniprot accession number, RefSeq id, etc.). The name of an org 40 | package is always of the form `org...db` 41 | (e.g. [org.Sc.sgd.db][]) where `` is a 2-letter abbreviation of 42 | the organism (e.g. `Sc` for *Saccharomyces cerevisiae*) and `` is 43 | an abbreviation (in lower-case) describing the type of central 44 | identifier (e.g. `sgd` for gene identifiers assigned by the 45 | *Saccharomyces* Genome Database, or `eg` for Entrez gene ids). The 46 | "How to use the '.db' annotation packages" vignette in the 47 | [AnnotationDbi][] package (org packages are only one type of ".db" 48 | annotation packages) is a key reference. The '.db' and most other 49 | Bioconductor annotation packages are updated every 6 months. 50 | 51 | Annotation packages usually contain an object named after the package 52 | itself. These objects are collectively called `AnnotationDb` objects, 53 | with more specific classes named `OrgDb`, `ChipDb` or `TranscriptDb` 54 | objects. Methods that can be applied to these objects include 55 | `cols()`, `keys()`, `keytypes()` and `select()`. Common operations 56 | for retrieving annotations are summarized in the table. 57 | 58 | | Category | Function | Description | 59 | |------------|---------------------------------------|------------------------------------------------------------------| 60 | | Discover | `columns()` | List the kinds of columns that can be returned | 61 | | | `keytypes()` | List columns that can be used as keys | 62 | | | `keys()` | List values that can be expected for a given keytype | 63 | | | `select()` | Retrieve annotations matching `keys`, `keytype` and `columns` | 64 | | Manipulate | `setdiff()`, `union()`, `intersect()` | Operations on sets | 65 | | | `duplicated()`, `unique()` | Mark or remove duplicates | 66 | | | `%in%`, `match()` | Find matches | 67 | | | `any()`, `all()` | Are any `TRUE`? Are all? | 68 | | | `merge()` | Combine two different \Robject{data.frames} based on shared keys | 69 | | `GRanges*` | `transcripts()`, `exons()`, `cds()` | Features (transcripts, exons, coding sequence) as `GRanges`. | 70 | | | `transcriptsBy()` , `exonsBy()` | Features group by gene, transcript, etc., as `GRangesList`. | 71 | | | `cdsBy()` | | 72 | 73 | **Exercise**: This exercise illustrates basic use of the `select' 74 | interface to annotation packages. 75 | 76 | 1. What is the name of the org package for *Homo sapiens*? Load it. 77 | Display the `OrgDb` object for the [org.Hs.eg.db][] package. Use 78 | the `columns()` method to discover which sorts of annotations can 79 | be extracted from it. 80 | 2. Use the `keys()` method to extract ENSEMBL identifiers and then 81 | pass those keys in to the `select()` method in such a way that you 82 | extract the SYMBOL (gene symbol) and GENENAME information for 83 | each. Use the following ENSEMBL ids. 84 | 85 | ```{r select-setup} 86 | ensids <- c("ENSG00000130720", "ENSG00000103257", "ENSG00000156414", 87 | "ENSG00000144644", "ENSG00000159307", "ENSG00000144485") 88 | ``` 89 | 90 | **Solution** The `OrgDb` object is named `org.Hs.eg.db`. 91 | ```{r select} 92 | library(org.Hs.eg.db) 93 | keytypes(org.Hs.eg.db) 94 | columns(org.Hs.eg.db) 95 | cols <- c("SYMBOL", "GENENAME") 96 | select(org.Hs.eg.db, keys=ensids, columns=cols, keytype="ENSEMBL") 97 | ``` 98 | 99 | ### Internet resources 100 | 101 | A short summary of select Bioconductor packages enabling web-based 102 | queries is in following Table. 103 | 104 | | Package | Description | 105 | |-----------------------------------------------------|-------------------------------------------| 106 | | [AnnotationHub][] | Ensembl, Encode, dbSNP, UCSC data objects | 107 | | [biomaRt](http://biomart.org) | Ensembl and other annotations | 108 | | [PSICQUIC](https://code.google.com/p/psicquic) | Protein interactions | 109 | | [uniprot.ws](http://uniprot.org) | Protein annotations | 110 | | [KEGGREST](http://www.genome.jp/kegg) | KEGG pathways | 111 | | [SRAdb](http://www.ncbi.nlm.nih.gov/sra) | Sequencing experiments. | 112 | | [rtracklayer](http://genome.ucsc.edu) | genome tracks. | 113 | | [GEOquery](http://www.ncbi.nlm.nih.gov/geo/) | Array and other data | 114 | | [ArrayExpress](http://www.ebi.ac.uk/arrayexpress/) | Array and other data | 115 | 116 | *Using biomaRt* 117 | 118 | The [biomaRt][] package offers access to the online 119 | [biomart](http://www.biomart.org) resource. this consists of several 120 | data base resources, referred to as 'marts'. Each mart allows access 121 | to multiple data sets; the [biomaRt][] package provides methods for 122 | mart and data set discovery, and a standard method `getBM()` to 123 | retrieve data. 124 | 125 | *Exercise* 126 | 127 | 1. Load the [biomaRt][] package and list the available marts. Choose 128 | the *ensembl* mart and list the datasets for that mart. Set up a 129 | mart to use the *ensembl* mart and the *hsapiens gene ensembl* 130 | dataset. 131 | 2. A [biomaRt][] dataset can be accessed via `getBM()`. In addition to 132 | the mart to be accessed, this function takes filters and attributes 133 | as arguments. Use `filterOptions()` and `listAttributes()` to 134 | discover values for these arguments. Call `getBM()` using filters 135 | and attributes of your choosing. 136 | 137 | *Solution* 138 | ```{r biomaRt1, eval=FALSE, results="hide"} 139 | ## NEEDS INTERNET ACCESS !! 140 | library(biomaRt) 141 | head(listMarts(), 3) ## list the marts 142 | head(listDatasets(useMart("ensembl")), 3) ## mart datasets 143 | ensembl <- ## fully specified mart 144 | useMart("ensembl", dataset = "hsapiens_gene_ensembl") 145 | 146 | head(listFilters(ensembl), 3) ## filters 147 | myFilter <- "chromosome_name" 148 | substr(filterOptions(myFilter, ensembl), 1, 50) ## return values 149 | myValues <- c("21", "22") 150 | head(listAttributes(ensembl), 3) ## attributes 151 | myAttributes <- c("ensembl_gene_id","chromosome_name") 152 | 153 | ## assemble and query the mart 154 | res <- getBM(attributes = myAttributes, filters = myFilter, 155 | values = myValues, mart = ensembl) 156 | ``` 157 | 158 | *Exercise* 159 | 160 | As an optional exercise, annotate the genes that are differentially 161 | expressed in the DESeq2 laboratory, e.g., find the \texttt{GENENAME} 162 | associated with the five most differentially expressed genes. Do these 163 | make biological sense? Can you `merge()` the annotation results with 164 | the `top table' results to provide a statistically and biologically 165 | informative summary? 166 | 167 | ## Genome annotation 168 | 169 | There are a diversity of packages and classes available for 170 | representing large genomes. Several include: 171 | 172 | - 'TxDb.*' For transcript and other genome / coordinate annotation. 173 | - [BSgenome][] For whole-genome representation. See 174 | `available.packages()` for pre-packaged genomes, and the vignette 175 | 'How to forge a BSgenome data package' in the 176 | - [Homo.sapiens][] For integrating 'TxDb*' and 'org.*' packages. 177 | - 'SNPlocs.*' For model organism SNP locations derived from dbSNP. 178 | - `FaFile()` ([Rsamtools][]) for accessing indexed FASTA files. 179 | - 'SIFT.*', 'PolyPhen', 'ensemblVEP' Variant effect scores. 180 | 181 | ### Transcript annotation packages 182 | 183 | Genome-centric packages are very useful for annotations involving 184 | genomic coordinates. It is straight-forward, for instance, to discover 185 | the coordinates of coding sequences in regions of interest, and from 186 | these retrieve corresponding DNA or protein coding sequences. Other 187 | examples of the types of operations that are easy to perform with 188 | genome-centric annotations include defining regions of interest for 189 | counting aligned reads in RNA-seq experiments and retrieving DNA 190 | sequences underlying regions of interest in ChIP-seq analysis, e.g., 191 | for motif characterization. 192 | 193 | *Exercise* 194 | 195 | This exercise uses annotation resources to go from a gene symbol 196 | 'BRCA1' through to the genomic coordinates of each transcript 197 | associated with the gene, and finally to the DNA sequences of the 198 | transcripts. 199 | 200 | 1. Use the [org.Hs.eg.db][] package to map from the gene symbol 201 | 'BRCA1' to its Entrez identifier. Do this using the `select` 202 | command. 203 | 2. Use the [TxDb.Hsapiens.UCSC.hg19.knownGene][] package to retrieve 204 | the transcript names (`TXNAME`) corresponding to the BRCA1 Entrez 205 | identifier. (The 'org\*' packages are based on information from 206 | NCBI, where Entrez identifiers are labeled ENTREZID; the 'TxDb*' 207 | package we are using is from UCSC, where Entrez identifiers are 208 | labeled GENEID). 209 | 3. Use the `cdsBy()` function to retrieve the genomic coordinates of 210 | all coding sequences grouped by transcript, and select the 211 | transcripts corresponding to the identifiers we're interested 212 | in. The coding sequences are returned as an `GRangesList`, where 213 | each element of the list is a `GRanges` object representing the 214 | exons in the coding sequence. As a sanity check, ensure that the 215 | sum of the widths of the exons in each coding sequence is evenly 216 | divisible by 3 (the R 'modulus' operator `%%` returns the remainder 217 | of the division of one number by another, and might be helpful in 218 | this case). 219 | 220 | 4. Visualize the transcripts in genomic coordinates using the [Gviz][] 221 | package to construct an `AnnotationTrack`, and plotting it using 222 | `plotTracks()`. 223 | 224 | 5. Use the [Bsgenome.Hsapiens.UCSC.hg19][] package and 225 | `extractTranscriptSeqs()` function to extract the DNA sequence of 226 | each transcript. 227 | 228 | 229 | *Solution* 230 | 231 | Retrieve the Entrez identifier corresponding to the BRCA1 gene symbol 232 | 233 | ```{r symbol-to-entrez} 234 | library(org.Hs.eg.db) 235 | eid <- select(org.Hs.eg.db, "BRCA1", "ENTREZID", "SYMBOL")[["ENTREZID"]] 236 | ``` 237 | 238 | Map from Entrez gene identifier to transcript name 239 | 240 | ```{r entrez-to-tx, messages=FALSE} 241 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 242 | txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene 243 | txid <- select(txdb, eid, "TXNAME", "GENEID")[["TXNAME"]] 244 | ``` 245 | 246 | Retrieve all coding sequences grouped by transcript, and select those 247 | matching the transcript ids of interest, verifying that each coding 248 | sequence width is a multiple of 3 249 | 250 | ```{r tx-to-cds-coords} 251 | cds <- cdsBy(txdb, by="tx", use.names=TRUE) 252 | brca1cds <- cds[names(cds) %in% txid] 253 | class(brca1cds) 254 | length(brca1cds) 255 | brca1cds[[1]] # exons in cds 256 | cdswidth <- width(brca1cds) # width of each exon 257 | all((sum(cdswidth) %% 3) == 0) # sum within cds, modulus 3 258 | ``` 259 | 260 | Visualize the BRCA1 transcripts using [Gviz][] (this package has an 261 | excellent vignette, `vignette("Gviz")`) 262 | 263 | ```{r Gviz, message=FALSE} 264 | require(Gviz) 265 | anno <- AnnotationTrack(brca1cds) 266 | plotTracks(list(GenomeAxisTrack(), anno)) 267 | ``` 268 | 269 | Extract the coding sequences of each transcript 270 | 271 | ```{r cds-to-seq} 272 | library(BSgenome.Hsapiens.UCSC.hg19) 273 | genome <- BSgenome.Hsapiens.UCSC.hg19 274 | tx_seq <- extractTranscriptSeqs(genome, brca1cds) 275 | tx_seq 276 | ``` 277 | 278 | Intron coordinates can be identified by first calculating the range of 279 | the genome (from the start of the first exon to the end of the last 280 | exon) covered by each transcript, and then taking the (algebraic) set 281 | difference between this and the genomic coordinates covered by each 282 | exon 283 | 284 | ```{r introns} 285 | introns <- psetdiff(range(brca1cds), brca1cds) 286 | ``` 287 | 288 | Retrieve the intronic sequences with `getSeq()` (these are *not* 289 | assembled, the way that `extractTranscriptSeqs()` assembles exon 290 | sequences into mature transcripts); note that introns start and end 291 | with the appropriate acceptor and donor site sequences. 292 | 293 | ```{r intron-seqs} 294 | seq <- getSeq(genome, introns) 295 | names(seq) 296 | seq[["uc010whl.2"]] # 21 introns 297 | ``` 298 | 299 | ### [rtracklayer][] 300 | 301 | The [rtracklayer][] package allows us to query the UCSC genome 302 | browser, as well as providing `import()` and 303 | `export()` functions for common annotation file formats like 304 | GFF, GTF, and BED. 305 | 306 | *Exercise* 307 | 308 | Here we use [rtracklayer][] to retrieve estrogen receptor 309 | binding sites identified across cell lines in the ENCODE project. We 310 | focus on binding sites in the vicinity of a particularly interesting 311 | region of interest. 312 | 313 | 1. Define our region of interest by creating a `GRanges` instance with 314 | appropriate genomic coordinates. Our region corresponds to 10Mb up- 315 | and down-stream of a particular gene. 316 | 2. Create a session for the UCSC genome browser 317 | 3. Query the UCSC genome browser for ENCODE estrogen receptor 318 | ERalpha$_a$ transcription marks; identifying the appropriate track, 319 | table, and transcription factor requires biological knowledge and 320 | detective work. 321 | 4. Visualize the location of the binding sites and their scores; 322 | annotate the mid-point of the region of interest. 323 | 324 | *Solution* 325 | 326 | Define the region of interest 327 | 328 | ```{r rtracklayer-roi} 329 | library(GenomicRanges) 330 | roi <- GRanges("chr10", IRanges(92106877, 112106876, names="ENSG00000099194")) 331 | ``` 332 | 333 | Create a session 334 | 335 | ```{r rtracklayer-session} 336 | library(rtracklayer) 337 | session <- browserSession() 338 | ``` 339 | 340 | Query the UCSC for a particular track, table, and transcription 341 | factor, in our region of interest 342 | 343 | ```{r rtracklayer-marks} 344 | trackName <- "wgEncodeRegTfbsClusteredV2" 345 | tableName <- "wgEncodeRegTfbsClusteredV2" 346 | trFactor <- "ERalpha_a" 347 | ucscTable <- getTable(ucscTableQuery(session, track=trackName, 348 | range=roi, table=tableName, name=trFactor)) 349 | ``` 350 | 351 | Visualize the result 352 | 353 | ```{r rtracklayer-plot, fig.height=3} 354 | plot(score ~ chromStart, ucscTable, pch="+") 355 | abline(v=start(roi) + (end(roi) - start(roi) + 1) / 2, col="blue") 356 | ``` 357 | 358 | ## Variants 359 | 360 | Follow the 361 | [Variants](http://bioconductor.org/help/workflows/variants/) work 362 | flow. 363 | 364 | [AnnotationHub]: http://bioconductor.org/packages/release/bioc/html/AnnotationHub.html 365 | [BSgenome]: http://bioconductor.org/packages/release/bioc/html/BSgenome.html 366 | [Bsgenome.Hsapiens.UCSC.hg19]: http://bioconductor.org/packages/release/data/annotation/html/Bsgenome.Hsapiens.UCSC.hg19.html 367 | [Homo.sapiens]: http://bioconductor.org/packages/release/data/annotation/html/Homo.sapiens.html 368 | [Rsamtools]: http://bioconductor.org/packages/release/bioc/html/Rsamtools.html 369 | [TxDb.Hsapiens.UCSC.hg19.knownGene]: http://bioconductor.org/packages/release/data/annotation/html/TxDb.Hsapiens.UCSC.hg19.knownGene.html 370 | [biomaRt]: http://bioconductor.org/packages/release/bioc/html/biomaRt.html 371 | [org.Hs.eg.db]: http://bioconductor.org/packages/release/data/annotation/html/org.Hs.eg.db.html 372 | [org.Sc.sgd.db]: http://bioconductor.org/packages/release/data/annotation/html/org.Sc.sgd.db.html 373 | [rtracklayer]: http://bioconductor.org/packages/release/bioc/html/rtracklayer.html 374 | -------------------------------------------------------------------------------- /vignettes/B02.5_GeneSetEnrichment.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Gene Set Enrichment and Pathway Analysis 17 | 18 | Martin Morgan
19 | October 28, 2014 20 | 21 | My overview of approaches to [gene set 22 | enrichment](http://bioconductor.org/help/course-materials/2013/EMBOBGI/GeneSetEnrichment.pdf), 23 | with a sketch of relevant _Bioconductor_ packages. -------------------------------------------------------------------------------- /vignettes/C03.1_LargeData.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Working with Large Data 17 | 18 | Martin Morgan
19 | October 29, 2014 20 | 21 | ## Scalable computing 22 | 23 | 1. Efficient _R_ code 24 | - Vectorize! 25 | - Reuse others' work -- `r Biocpkg("DESeq2")`, 26 | `r Biocpkg("GenomicRanges")`, `r Biocpkg("Biostrings")`, ..., 27 | `r CRANpkg("dplyr")`, `r CRANpkg("data.table")`, `r CRANpkg("Rcpp")` 28 | - Useful tools: `system.time()`, `Rprof()`, `r CRANpkg("microbenchmark")` 29 | - More detail in 30 | [deadly sins](http://bioconductor.org/help/course-materials/2014/CSAMA2014/1_Monday/labs/IntermediateR.html#efficient-code) 31 | of a previous course. 32 | 2. Iteration 33 | - Chunk-wise 34 | - `open()`, read chunk(s), `close()`. 35 | - e.g., `yieldSize` argument to `Rsamtools::BamFile()` 36 | 3. Restriction 37 | - Limit to columns and / or rows of interest 38 | - Exploit domain-specific formats, e.g., BAM files and 39 | `Rsamtools::ScanBamParam()` 40 | - Use a data base 41 | 4. Sampling 42 | - Iterate through large data, retaining a manageable sample, e.g., 43 | `ShortRead::FastqSampler()` 44 | 5. Parallel evaluation 45 | - **After** writing efficient code 46 | - Typically, `lapply()`-like operations 47 | - Cores on a single machine ('easy'); clusters (more tedious); 48 | clouds 49 | 50 | ## Parallel evaluation in _Bioconductor_ 51 | 52 | - [BiocParallel][] -- `bplapply()` for `lapply()`-like functions, 53 | increasingly used by package developers to provide easy, standard 54 | way of gaining parallel evaluation. 55 | - [GenomicFiles][] -- Framework for working on groups of files, 56 | ranges, or ranges x files 57 | - Bioconductor 58 | [AMI](http://bioconductor.org/help/bioconductor-cloud-ami/) (Amazon 59 | Machine Instance) including pre-configured StarCluster. 60 | 61 | ## Lab 62 | 63 | ### Efficient code 64 | 65 | Write the following as a function. Use `system.time()` to explore how 66 | long this takes to execute as `n` increases from 100 to 10000. Use 67 | `identical()` and `r CRANpkg("microbenchmark")` to compare 68 | alternatives `f1()`, `f2()`, and `f3()` for both correctness and performance of 69 | these three different functions. What strategies are these functions 70 | using? 71 | 72 | ```{r benchmark} 73 | f0 <- function(n) { 74 | ## inefficient! 75 | ans <- numeric() 76 | for (i in seq_len(n)) 77 | ans <- c(ans, exp(i)) 78 | ans 79 | } 80 | 81 | f1 <- function(n) { 82 | ans <- numeric(n) 83 | for (i in seq_len(n)) 84 | ans[[i]] <- exp(i) 85 | ans 86 | } 87 | 88 | f2 <- function(n) 89 | sapply(seq_len(n), exp) 90 | 91 | f3 <- function(n) 92 | exp(seq_len(n)) 93 | ``` 94 | 95 | ### Sleeping serially and in parallel 96 | 97 | Go to sleep for 1 second, then return `i`. This takes 8 seconds. 98 | 99 | ```{r parallel-sleep} 100 | library(BiocParallel) 101 | 102 | fun <- function(i) { 103 | Sys.sleep(1) 104 | i 105 | } 106 | 107 | ## serial 108 | f0 <- function(n) 109 | lapply(seq_len(n), fun) 110 | 111 | ## parallel 112 | f1 <- function(n) 113 | bplapply(seq_len(n), fun) 114 | ``` 115 | 116 | ### Counting overlaps -- our own version 117 | 118 | Regions of interest, named like the chromosomes in the bam file. 119 | 120 | ```{r count-overlaps-roi, eval=FALSE} 121 | library(TxDb.Hsapiens.UCSC.hg19.knownGene) 122 | exByTx <- exonsBy(TxDb.Hsapiens.UCSC.hg19.knownGene, "tx") 123 | 124 | map0 <- read.delim("~/igv/genomes/hg19_alias.tab", header=FALSE, 125 | stringsAsFactors=FALSE) 126 | map <- setNames(map0$V1, map0$V2) 127 | seqlevels(exByTx, force=TRUE) <- map 128 | ``` 129 | 130 | A function to iterate through a bam file 131 | 132 | ```{r count-overlaps, eval=FALSE} 133 | count1 <- function(filename, roi) { 134 | ## Create and open BAM file 135 | bf <- BamFile(filename, yieldSize=1000000) 136 | open(bf) 137 | 138 | ## initialize variables 139 | n <- 0 # number of reads examined 140 | count <- integer(length(roi)) # running count of reads overlapping roi 141 | names(counts) <- names(roi) 142 | 143 | ## read in and count chunks of data, until done 144 | repeat { 145 | ## input 146 | aln <- readGAlignments(bf) # input next chunk 147 | if (length(aln) == 0) # stopping condition 148 | break 149 | n <- n + length(aln) # how are we doing? 150 | message(n) 151 | 152 | ## overlaps 153 | olaps <- findOverlaps(aln, roi, type="within", ignore.strand=TRUE) 154 | count <- count + tabulate(subjectHits(olaps), subjectLength(olaps)) 155 | } 156 | 157 | ## finish and return result 158 | close(bf) 159 | count 160 | } 161 | ``` 162 | 163 | In action 164 | 165 | ```{r count-overlaps-doit, eval=FALSE} 166 | filename <- "~/bam/SRR1039508_sorted.bam" 167 | count <- count1(filename, exByTx) 168 | ``` 169 | 170 | Parallelize 171 | 172 | ```{r count-overlaps-parallel, eval=FALSE} 173 | library(BiocParallel) 174 | 175 | ## all bam files 176 | filenames <- dir("~/bam", pattern="bam$", full=TRUE) 177 | names(filenames) <- sub("_sorted.bam", "", basename(filenames)) 178 | 179 | ## iterate 180 | counts <- bplapply(filenames, count1, exByTx) 181 | counts <- simplify2array(counts) 182 | head(counts) 183 | ``` 184 | 185 | ## Resources 186 | 187 | - Lawrence, M, and Morgan, M. 2014. Scalable Genomics with R and 188 | Bioconductor. Statistical Science 2014, Vol. 29, No. 2, 189 | 214-226. http://arxiv.org/abs/1409.2864v1 190 | 191 | [BiocParallel]: http://bioconductor.org/packages/release/bioc/html/BiocParallel.html 192 | [GenomicFiles]: http://bioconductor.org/packages/release/bioc/html/GenomicFiles.html 193 | 194 | -------------------------------------------------------------------------------- /vignettes/C03.2_CodeToPackages.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Organizing Code in Functions, Files ,and Packages 17 | 18 | Martin Morgan
19 | October 29, 2014 20 | 21 | ## Programming 22 | 23 | _R_ is a programming language 24 | 25 | - Your own functions 26 | 27 | ```{r fun} 28 | fun <- function(n) { 29 | x <- rnorm(n) 30 | y <- x + rnorm(n, sd=.5) 31 | lm(y ~ x) 32 | } 33 | ``` 34 | 35 | - Check out the _RStudio_ function wizard! 36 | 37 | - Iteration 38 | 39 | ```{r iteration, eval=FALSE} 40 | ## 'side effects' 41 | for (i in 1:10) 42 | message(i) 43 | 44 | ## result for subsequent computation 45 | sapply(1:10, function(i) { 46 | sum(rnorm(1000)) 47 | }) 48 | ``` 49 | 50 | - Other iteration paradigms: `lapply()`, `apply()`, `mapply() / Map()`. 51 | 52 | - Conditional execution 53 | 54 | ```{r conditional, eval=FALSE} 55 | sapply(1:10, function(i)) { 56 | if (sum(rnorm(1000)) > 0) { 57 | "hi!" 58 | } else { 59 | "low!" 60 | } 61 | }) 62 | ``` 63 | 64 | ## From script to package 65 | 66 | Scripts contain a combination of data and transformations. Often the 67 | transformations are idiosyncratic, and rely heavily on functions 68 | provided by various packages. Sometimes the script contains a useful 69 | chunk of code that could be reused in different places. Examples we've 70 | encountered in this course might include GC-content of DNA sequences 71 | (or is there a function for that already? check out `r 72 | Biocpkg("Biostrings")`!) and creating a simple 'map' from one type of 73 | annotation to another. 74 | 75 | It is easy and beneficial to create a package. 76 | 77 | - Only need to think carefully about how to implement the function once. 78 | - Code in the package is reused -- if you correct an error, then all 79 | your scripts automatically benefit. 80 | - Share with your lab mates / colleagues for consistent results, and 81 | with the wider community for fame and glory. 82 | 83 | What is an _R_ package? 84 | 85 | - Simple directory structure of required files. Minimal: 86 | 87 | MyPackage 88 | |-- DESCRIPTION (title, author, version, etc.) 89 | |-- NAMESPACE (imports / exports) 90 | |-- R (R function definitions) 91 | |-- gc.R 92 | |-- annoHelper.R 93 | |-- man (help pages) 94 | |-- MyPackage.Rd 95 | |-- gc.Rd 96 | |-- annoHelper.Rd 97 | 98 | Check out the _Rstudio_ package wizard! 99 | 100 | ## Lab 101 | 102 | ### GC-content 103 | 104 | Write a function that takes a `DNAStringSet` and returns the GC content. 105 | 106 | Modify the function using a conditional statement to work whether 107 | provided a `DNAString` or a `DNAStringSet`. Test the function. 108 | 109 | Save the function in a file on your AMI. 110 | 111 | ### Annotation-helper 112 | 113 | Write a function that takes as its argument Ensembl gene identifiers 114 | (like the `rownames()` of the _SummarizedExperiment_ object in the 115 | RNASeq vignette yesterday) and uses the `select()` method and 116 | annotation package `r Biocannopkg("org.Hs.eg.db")` to return a named 117 | character vector, where the names of the vector are the Ensembl 118 | identifiers and the values are the corresponding gene SYMBOLs. Adopt 119 | some simple-to-implement policy for handling Ensembl identifiers that 120 | map to more than one gene symbol. Save this function to another file 121 | 122 | ### A package 123 | 124 | Use the _RStudio_ wizard to create a package from the files containing 125 | your GC-content and annotation-helper functions. 126 | -------------------------------------------------------------------------------- /vignettes/C03.3_ReproducibleResearch.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Reproducible Research 17 | 18 | Martin Morgan
19 | October 29, 2014 20 | 21 | ## Importance of reproducibility 22 | 23 | A [cautionary tale](http://bioconductor.org/help/course-materials/2013/EMBOBGI/reproducible-research.pdf) 24 | 25 | ## Facilities enabling reproducible analysis in _R_ 26 | 27 | Basics -- quickly becomes unsatisfactory 28 | 29 | - 'Old-school' text scripts 30 | - Comments describing approaches 31 | - Package versions 32 | 33 | Vignettes -- 'literate programming' 34 | 35 | - Markdown 36 | - _R_ markdown; weaving and tangling 37 | - _LaTeX_ 38 | 39 | ## More advanced considerations 40 | 41 | Version control 42 | 43 | - Alternative (better!) to using cryptic file names to indicate 44 | different versions 45 | - Create a 'repository', add / edit files, 'commit' changes 46 | - Easily review 'diff'erences, restore previous versions, ... 47 | - _RStudio_ 48 | 49 | - File -> New Project --> New Directory --> Empty Project. Click 50 | 'Create git repository. 51 | - File --> New File --> R Script. Edit. 52 | - 'Git' icon on menu bar --> Commit 53 | 54 | Packages 55 | 56 | - The Rstudio 'wizard' 57 | - Benefits 58 | 59 | - Code re-use standardizes analysis 60 | - No need to copy / paste code 61 | - Easy to share with colleagues (work group, company, world) 62 | 63 | ## Lab 64 | 65 | Produce a short 'vignette' summarizing your RNA-seq work 66 | yesterday. The code chunks might be along the lines of 67 | 68 | ```{r workflow-code-chunks, eval=FALSE} 69 | library(airway) 70 | library(DESeq2) 71 | 72 | data(airway) 73 | 74 | dds <- DESeqDataSet(airway, design = ~ cell + dex) 75 | dds$dex <- relevel(dds$dex, "untrt") 76 | dds <- DESeq(dds) 77 | res <- results(dds) 78 | ``` 79 | 80 | Embed this in textual description with relevant descriptive 81 | information (title, author, date) as well as text describing nuances 82 | of each step, plus figures or tables of your choosing to illustrate 83 | relevant aspects of the experimental design or results. 84 | 85 | Render this as an HTML document to share with your colleagues at home. 86 | -------------------------------------------------------------------------------- /vignettes/C03.4_Visualization.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | library(LearnBioconductor) 9 | stopifnot(BiocInstaller::biocVersion() == "3.0") 10 | ``` 11 | 12 | ```{r style, echo = FALSE, results = 'asis'} 13 | BiocStyle::markdown() 14 | ``` 15 | 16 | # Visualization 17 | 18 | Martin Morgan
19 | October 29, 2014 20 | 21 | [Three 22 | approaches](http://bioconductor.org/help/course-materials/2014/summerx/Visualization-slides.pdf) 23 | 24 | - 'base' graphics 25 | - _lattice_ 26 | - `r CRANpkg("ggplot2")` 27 | 28 | - A neat _ggplot2_ example: 29 | [slide 19+](http://bioconductor.org/help/course-materials/2014/CSAMA2014/2_Tuesday/lectures/Visualization_in_Statistical_Genomics-Carey.pdf) 30 | from Vince Carey's presentation. 31 | 32 | 33 | Genome-centric 34 | 35 | - `r Biocpkg("Gviz")` for visualizing genomic regions 36 | - `r Biocpkg("ggbio")` provides extensive options 37 | 38 | Interactive visualization 39 | 40 | - `r Biocpkg("rtracklayer")` for visualizing genomic ranges and 41 | managing UCSC genome browser sessions. 42 | - `r Biocpkg('epivizr')` for interactive display 43 | - `r CRANpkg("shiny")` [web](http://shiny.rstudio.com) for interactive 44 | apps. 45 | 46 | Reports 47 | 48 | - `r Biocpkg("ReportingTools")` for easy report templates 49 | 50 | And... 51 | 52 | - `r CRANpkg("RColorBrewer")` and 53 | [web site](http://colorbrewer2.org/): helps choose sensible color 54 | schemes. 55 | - [Accidental aRt](http://accidental-art.tumblr.com/archive) 56 | - [stick figure](https://github.com/EconometricsBySimulation/R-Graphics/blob/master/Stick-Figures/draw.stick.R) 57 | 58 | ## Lab 59 | 60 | ### Gviz / ggbio 61 | 62 | Work through 63 | [vignette section 2](http://bioconductor.org/packages/release/bioc/vignettes/Gviz/inst/doc/Gviz.pdf) 64 | the `r Biocpkg('Gviz')` package, and then the 'Plotting fold changes 65 | in genomic space' in the [RNASeq lab](B02.1.1_RNASeqLab.html). 66 | 67 | Peruse the 68 | [vignette](http://bioconductor.org/packages/release/bioc/vignettes/ggbio/inst/doc/ggbio.pdf) of the `r Biocpkg('ggbio')` package. 69 | 70 | Run the following _DESeq2_ work flow to arrive at a top-table; coerce 71 | the result to a `data.frame`. 72 | 73 | ```{r ggplot-setup, eval=FALSE} 74 | library(DESeq2) 75 | library(airway) 76 | data(airway) 77 | se = airway 78 | dds <- DESeqDataSet(se, design = ~ cell + dex) 79 | dds$dex <- relevel(dds$dex, "untrt") 80 | dds <- DESeq(dds) 81 | res <- results(dds) 82 | resdf <- as.data.frame(res) 83 | ``` 84 | 85 | A 'volcano plot' shows the relationship between P-value and log fold 86 | change. Here's a basic volcano plot using base graphics; create a 87 | volcano plot using _ggplot2_ and / or _lattice_. 88 | 89 | ```{r volcano, eval=FALSE} 90 | plot(-log10(padj) ~ log2FoldChange, resdf) 91 | ``` 92 | 93 | -------------------------------------------------------------------------------- /vignettes/D04.1_InstallIGV.Rmd: -------------------------------------------------------------------------------- 1 | 6 | 7 | ```{r setup, echo=FALSE} 8 | stopifnot(BiocInstaller::biocVersion() == "3.0") 9 | ``` 10 | 11 | ```{r style, echo = FALSE, results = 'asis'} 12 | BiocStyle::markdown() 13 | ``` 14 | ## Appendix 1: Installing and using IGV 15 | 16 | _NOTE_: All instructions in this document should be performed 17 | **on your laptop**, not on the RStudio Server AMI. 18 | 19 | * We'll first create a directory called "igv" in your 20 | home directory. You can determine your home directory by issuing 21 | the following command in a Terminal window on Linux or Mac: 22 | 23 | echo $HOME 24 | 25 | On Windows, open a command window by clicking Start, then Run, then 26 | typing `cmd` and pressing Enter. In the window, type 27 | 28 | echo %USERPROFILE% 29 | 30 | * So go to the indicated directory, and then issue the command 31 | 32 | mkdir igv 33 | 34 | That will create an `igv` directory in your home directory. 35 | 36 | * Copy the file [hg19_alias.tab](hg19_alias.tab) to this 37 | directory. 38 | This is a simple tab-delimited file that maps between the sequence 39 | names used by the alignment, and the sequence names known to IGV. 40 | 41 | * Install Java if you haven't already. Go to 42 | [https://www.java.com/en/](https://www.java.com/en/) and 43 | click on Free Java Download. Download and install. 44 | 45 | * Download IGV from [https://www.broadinstitute.org/igv/download](https://www.broadinstitute.org/igv/download). There are several 46 | ways to do this, the easiest is to click on one of the Java Web Start 47 | links. Make sure you have more memory available than what is listed below the Launch button. You may need to change the security settings on your 48 | computer to allow IGV to launch. 49 | 50 | * With IGV running, let's change the default genome. It's currently 51 | set to Human hg18, so click on the dropdown in the upper left 52 | that says "Human hg18". Click on "More..." and then in the 53 | `Filter` box, start typing `hg19`. `Human hg19` will then show 54 | up in the box and you can double-click on it. We're now using 55 | Human hg19 as our default genome. 56 | 57 | * You can now open bam files in IGV. Here are the URLs of 58 | BAM files for this class. You can open them by clicking on 59 | IGV's `File` menu, then clicking on `Load From URL...`. 60 | You can paste one of these URLs into the box and click OK: 61 | 62 | 
63 | 
64 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039508_sorted.bam
65 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039509_sorted.bam
66 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039512_sorted.bam
67 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039513_sorted.bam
68 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039516_sorted.bam
69 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039517_sorted.bam
70 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039520_sorted.bam
71 |     http://s3-us-west-2.amazonaws.com/oct2014bamfiles/SRR1039521_sorted.bam
72 |   
73 |
74 | 75 | You may see a warning that the bam file does not contain any sequence names which match the current genome. You can ignore this (click OK). 76 | 77 | * Zoom in to a particular gene, e.g., SPARCL1, by entering the gene 78 | symbol in the box toward the top center of the browser window. Then click `Go`. Adjust 79 | the zoom until reads come in to view, and interpret the result. 80 | 81 | -------------------------------------------------------------------------------- /vignettes/hg19_alias.tab: -------------------------------------------------------------------------------- 1 | gi|224384768|gb|CM000663.1| chr1 2 | gi|224384767|gb|CM000664.1| chr2 3 | gi|224384766|gb|CM000665.1| chr3 4 | gi|224384765|gb|CM000666.1| chr4 5 | gi|224384764|gb|CM000667.1| chr5 6 | gi|224384763|gb|CM000668.1| chr6 7 | gi|224384762|gb|CM000669.1| chr7 8 | gi|224384761|gb|CM000670.1| chr8 9 | gi|224384760|gb|CM000671.1| chr9 10 | gi|224384759|gb|CM000672.1| chr10 11 | gi|224384758|gb|CM000673.1| chr11 12 | gi|224384757|gb|CM000674.1| chr12 13 | gi|224384756|gb|CM000675.1| chr13 14 | gi|224384755|gb|CM000676.1| chr14 15 | gi|224384754|gb|CM000677.1| chr15 16 | gi|224384753|gb|CM000678.1| chr16 17 | gi|224384752|gb|CM000679.1| chr17 18 | gi|224384751|gb|CM000680.1| chr18 19 | gi|224384750|gb|CM000681.1| chr19 20 | gi|224384749|gb|CM000682.1| chr20 21 | gi|224384748|gb|CM000683.1| chr21 22 | gi|224384747|gb|CM000684.1| chr22 23 | -------------------------------------------------------------------------------- /vignettes/our_figures/GRanges.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/GRanges.png -------------------------------------------------------------------------------- /vignettes/our_figures/GRangesList.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/GRangesList.png -------------------------------------------------------------------------------- /vignettes/our_figures/RangeOperations.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/RangeOperations.png -------------------------------------------------------------------------------- /vignettes/our_figures/SequencingEcosystem.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/SequencingEcosystem.png -------------------------------------------------------------------------------- /vignettes/our_figures/SequencingEcosystem_no_bioc_pkgs.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/SequencingEcosystem_no_bioc_pkgs.png -------------------------------------------------------------------------------- /vignettes/our_figures/Solexa-bridge-pcr.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/Solexa-bridge-pcr.jpg -------------------------------------------------------------------------------- /vignettes/our_figures/SummarizedExperiment.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/SummarizedExperiment.png -------------------------------------------------------------------------------- /vignettes/our_figures/copy_number_QC_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/copy_number_QC_2.png -------------------------------------------------------------------------------- /vignettes/our_figures/cross_validation.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/cross_validation.png -------------------------------------------------------------------------------- /vignettes/our_figures/knn.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/knn.png -------------------------------------------------------------------------------- /vignettes/our_figures/nrg2825-f2.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Bioconductor/LearnBioconductor/f1f94b661f33cc911ac06bfffc823c6690fc84f7/vignettes/our_figures/nrg2825-f2.jpg --------------------------------------------------------------------------------