├── LICENSE ├── README.md ├── Supplemental-Information.pdf ├── base.py ├── models.py ├── mydata ├── cancer_type.txt ├── deg.csv └── sga.txt ├── prepare_dataset.py ├── test_run.py └── utils.py /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2019 Yifeng Tao 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. 22 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Genomic Impact Transformer (GIT) 2 | 3 | 4 | ## Introduction 5 | 6 | This repository contains PyTorch implementation of GIT model (and its variants) in the following paper: 7 | Yifeng Tao, Chunhui Cai, William W. Cohen, Xinghua Lu. [From genome to phenome: Predicting multiple cancer phenotypes based on somatic genomic alterations via the genomic impact transformer](https://arxiv.org/abs/1902.00078). Proceedings of the Pacific Symposium on Biocomputing (PSB). 2020. 8 | 9 | The preprocessed TCGA dataset, and gene embeddings mentioned in the paper are also released below. 10 | 11 | 12 | ## Prerequisites 13 | 14 | The code runs on `Python 2.7`, and following packages are used: 15 | * `PyTorch 0.1.12_2`, `pandas`, `pickle`, `numpy`, `random`, `argparse`, `os`. 16 | 17 | 18 | ## Data 19 | 20 | All the required data can be downloaded to the directory: 21 | ``` 22 | cd genome-transformer 23 | wget www.cs.cmu.edu/~yifengt/paper/2019a/data.tar.gz -O data.tar.gz 24 | tar -zxvf data.tar.gz 25 | ``` 26 | 27 | ### TCGA dataset 28 | 29 | The preprocessed SGA-DEG TCGA dataset is available at `data/dataset.pkl`, which contains SGAs, DEGs, cancer types, and barcodes of 4,468 cancer samples from TCGA: 30 | ``` 31 | data = pickle.load( open("data/dataset.pkl", "rb") ) 32 | ``` 33 | 34 | ### Gene embeddings 35 | 36 | Two types of gene embeddings are available: 37 | * Gene2Vec-pretrained gene embeddings: `data/gene_emb_pretrain.npy`; 38 | * Gene2Vec-pretrained + GIT-finetuned gene embeddings: `data/gene_emb_finetune.npy`. 39 | 40 | You may extract the gene embeddings in Python: 41 | ``` 42 | gene_emb_matrix = numpy.load("data/gene_emb_finetune.npy") 43 | ``` 44 | 45 | It is a 19782 by 512 matrix, where the index of each row can be mapped to a gene name through `data/idx2gene.txt`. 46 | 47 | ### Trained GIT model 48 | 49 | The parameters of trained GIT model are stored at `data/trained_model.pth`. 50 | 51 | 52 | ## Replicate experiments 53 | 54 | ### Train GIT model 55 | 56 | You can train the GIT from scratch and then evaluate its performance on test set: 57 | ``` 58 | python test_run.py 59 | ``` 60 | 61 | ### Evaluate GIT model 62 | 63 | If you have already downloaded the trained parameters of GIT, you can directly evaluate its performance: 64 | ``` 65 | python test_run.py --train_model False 66 | ``` 67 | 68 | ### GIT variants 69 | 70 | You may run more GIT variants, e.g., `GIT-init`, `GIT-attn`, `GIT-can` etc., by checking the code and comments of `test_run.py`, or: 71 | ``` 72 | python test_run.py --help 73 | ``` 74 | 75 | ## FAQ 76 | 77 | * Q: How to prepare my own input data and run experiments on the new data? 78 | 79 | * A: You can directly prepare the `dataset.pkl` pickle file. However, we also provide pieces of code (`prepare_dataset.py`) to faciliate you to convert more general data types (`csv` and `txt`) into the `dataset.pkl`. 80 | 81 | First, you need to format your own data into three files: `mydata/cancer_type.txt` (each line a cancer type), `deg.csv` (each row a sample, each column an over/under-expressed gene), `sga.txt` (each row a sample name and mutated genes, comma separated). 82 | 83 | Second, run the following code to convert and merge these files into a single `mydata/dataset.pkl` file: 84 | ``` 85 | python prepare_dataset.py --input_dir mydata --output_dir mydata 86 | ``` 87 | 88 | Third, run experiments on the new data: 89 | ``` 90 | python test_run.py --initializtion False --input_dir mydata --output_dir mydata 91 | ``` 92 | 93 | ## Citation 94 | 95 | If you find the data or code from this repository helpful, please cite this [paper](https://arxiv.org/abs/1902.00078): 96 | ``` 97 | @inproceedings{tao2020git, 98 | title = {From Genome to Phenome: Predicting Multiple Cancer Phenotypes based on Somatic Genomic Alterations via the Genomic Impact Transformer}, 99 | author = {Tao, Yifeng and 100 | Cai, Chunhui and 101 | Cohen, William W. and 102 | Lu, Xinghua}, 103 | booktitle={Pacific Symposium on Biocomputing}, 104 | year = {2020}, 105 | } 106 | ``` 107 | 108 | -------------------------------------------------------------------------------- /Supplemental-Information.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/yifengtao/genome-transformer/0296cc65037bc3a81832dfc4813f9df625337c7f/Supplemental-Information.pdf -------------------------------------------------------------------------------- /base.py: -------------------------------------------------------------------------------- 1 | """ Base model of GIT and its variants. 2 | 3 | """ 4 | import torch 5 | import torch.nn as nn 6 | 7 | __author__ = "Yifeng Tao" 8 | 9 | 10 | class ModelBase(nn.Module): 11 | """ Base models for all models. 12 | 13 | """ 14 | 15 | def __init__(self, args): 16 | """ Initialize the hyperparameters of model. 17 | 18 | Parameters 19 | ---------- 20 | args: arguments for initializing the model. 21 | 22 | """ 23 | 24 | super(ModelBase, self).__init__() 25 | 26 | self.epsilon = 1e-4 27 | 28 | self.input_dir = args.input_dir 29 | self.output_dir = args.output_dir 30 | 31 | self.sga_size = args.sga_size 32 | self.deg_size = args.deg_size 33 | self.can_size = args.can_size 34 | 35 | self.num_max_sga = args.num_max_sga 36 | 37 | self.embedding_size = args.embedding_size 38 | self.hidden_size = args.hidden_size 39 | self.attention_size = args.attention_size 40 | self.attention_head = args.attention_head 41 | 42 | self.learning_rate = args.learning_rate 43 | self.dropout_rate = args.dropout_rate 44 | self.weight_decay = args.weight_decay 45 | 46 | self.initializtion = args.initializtion 47 | self.attention = args.attention 48 | self.cancer_type = args.cancer_type 49 | self.deg_shuffle = args.deg_shuffle 50 | 51 | 52 | def build(self): 53 | """ Define modules of the model. 54 | 55 | """ 56 | 57 | raise NotImplementedError 58 | 59 | 60 | def forward(self): 61 | """ Define the data flow across modules of the model. 62 | 63 | """ 64 | 65 | raise NotImplementedError 66 | 67 | 68 | def train(self): 69 | """ Train the model using training set. 70 | 71 | """ 72 | 73 | raise NotImplementedError 74 | 75 | 76 | def test(self): 77 | """ Test the model using test set. 78 | 79 | """ 80 | raise NotImplementedError 81 | 82 | 83 | def load_model(self, path="data/trained_model.pth"): 84 | """ Load trained parameters of the model. 85 | 86 | """ 87 | 88 | print("Loading model from "+path) 89 | self.load_state_dict(torch.load(path)) 90 | 91 | 92 | def save_model(self, path="data/trained_model.pth"): 93 | """ Save learnable parameters of the trained model. 94 | 95 | """ 96 | 97 | print("Saving model to "+path) 98 | torch.save(self.state_dict(), path) 99 | 100 | -------------------------------------------------------------------------------- /models.py: -------------------------------------------------------------------------------- 1 | """ Implementation of GIT model and its variants. 2 | 3 | """ 4 | import os 5 | import numpy as np 6 | 7 | import torch 8 | import torch.nn as nn 9 | import torch.nn.functional as F 10 | import torch.optim as optim 11 | 12 | from utils import get_minibatch, evaluate 13 | from base import ModelBase 14 | 15 | __author__ = "Yifeng Tao" 16 | 17 | 18 | class GIT(ModelBase): 19 | """ GIT model and its variants. 20 | 21 | """ 22 | 23 | def __init__(self, args, **kwargs): 24 | """ Initialize the model. 25 | 26 | Parameters 27 | ---------- 28 | args: arguments for initializing the model. 29 | 30 | """ 31 | 32 | super(GIT, self).__init__(args, **kwargs) 33 | 34 | 35 | def build(self): 36 | """ Define modules of the model. 37 | 38 | """ 39 | 40 | self.layer_sga_emb = nn.Embedding( 41 | num_embeddings=self.sga_size+1, 42 | embedding_dim=self.embedding_size, 43 | padding_idx=0) 44 | 45 | self.layer_can_emb = nn.Embedding( 46 | num_embeddings=self.can_size+1, 47 | embedding_dim=self.embedding_size, 48 | padding_idx=0) 49 | 50 | self.layer_w_0 = nn.Linear( 51 | in_features=self.embedding_size, 52 | out_features=self.attention_size, 53 | bias=True) 54 | 55 | self.layer_beta = nn.Linear( 56 | in_features=self.attention_size, 57 | out_features=self.attention_head, 58 | bias=True) 59 | 60 | self.layer_dropout_1 = nn.Dropout(p=self.dropout_rate) 61 | 62 | self.layer_w_1 = nn.Linear( 63 | in_features=self.embedding_size, 64 | out_features=self.hidden_size, 65 | bias=True) 66 | 67 | self.layer_dropout_2 = nn.Dropout(p=self.dropout_rate) 68 | 69 | self.layer_w_2 = nn.Linear( 70 | in_features=self.hidden_size, 71 | out_features=self.deg_size, 72 | bias=True) 73 | 74 | if self.initializtion: 75 | gene_emb_pretrain = np.load(os.path.join(self.input_dir, "gene_emb_pretrain.npy")) 76 | self.layer_sga_emb.weight.data.copy_(torch.from_numpy(gene_emb_pretrain)) 77 | 78 | self.optimizer = optim.Adam( 79 | self.parameters(), 80 | lr=self.learning_rate, 81 | weight_decay=self.weight_decay) 82 | 83 | 84 | def forward(self, sga_index, can_index): 85 | """ Forward process. 86 | 87 | Parameters 88 | ---------- 89 | sga_index: list of SGA index vectors. 90 | can_index: list of cancer type indices. 91 | 92 | Returns 93 | ------- 94 | preds: 2D array of float in [0, 1] 95 | predicted gene expression 96 | hid_tmr: 2D array of float 97 | hidden layer of MLP 98 | emb_tmr: 2D array of float 99 | tumor embedding 100 | emb_sga: 2D array of float 101 | stratified tumor embedding 102 | attn_wt: 2D array of float 103 | attention weights of SGAs 104 | 105 | """ 106 | 107 | # cancer type embedding 108 | emb_can = self.layer_can_emb(can_index) 109 | emb_can = emb_can.view(-1,self.embedding_size) 110 | 111 | # gene embedings 112 | E_t = self.layer_sga_emb(sga_index) 113 | 114 | # squeeze and tanh-curve the gene embeddings 115 | E_t_flatten = E_t.view(-1, self.embedding_size) 116 | E_t1_flatten = torch.tanh( self.layer_w_0(E_t_flatten) ) 117 | 118 | # multiplied by attention heads 119 | E_t2_flatten = self.layer_beta(E_t1_flatten) 120 | E_t2 = E_t2_flatten.view(-1, self.num_max_sga, self.attention_head) 121 | 122 | # normalize by softmax 123 | E_t2 = E_t2.permute(1,0,2) 124 | A = F.softmax(E_t2) 125 | A = A.permute(1,0,2) 126 | 127 | if self.attention: 128 | # multi-head attention weighted sga embedding: 129 | emb_sga = torch.sum( torch.bmm( A.permute(0,2,1), E_t ), dim=1) 130 | emb_sga = emb_sga.view(-1,self.embedding_size) 131 | else: 132 | # if not using attention, simply sum up SGA embeddings 133 | emb_sga = torch.sum(E_t, dim=1) 134 | emb_sga = emb_sga.view(-1, self.embedding_size) 135 | 136 | # if use cancer type input, add cancer type embedding 137 | if self.cancer_type: 138 | emb_tmr = emb_can+emb_sga 139 | else: 140 | emb_tmr = emb_sga 141 | 142 | # MLP decoder 143 | emb_tmr_relu = self.layer_dropout_1(F.relu(emb_tmr)) 144 | hid_tmr = self.layer_w_1(emb_tmr_relu) 145 | hid_tmr_relu = self.layer_dropout_2(F.relu(hid_tmr)) 146 | preds = F.sigmoid(self.layer_w_2(hid_tmr_relu)) 147 | 148 | # attention weights 149 | attn_wt = torch.sum(A, dim=2) 150 | attn_wt = attn_wt.view(-1, self.num_max_sga) 151 | 152 | return preds, hid_tmr, emb_tmr, emb_sga, attn_wt 153 | 154 | 155 | def train(self, train_set, test_set, 156 | batch_size=None, test_batch_size=None, 157 | max_iter=None, max_fscore=None, 158 | test_inc_size=None, **kwargs): 159 | """ Train the model until max_iter or max_fscore reached. 160 | 161 | Parameters 162 | ---------- 163 | train_set: dict 164 | dict of lists, including SGAs, cancer types, DEGs, patient barcodes 165 | test_set: dict 166 | batch_size: int 167 | test_batch_size: int 168 | max_iter: int 169 | max number of iterations that the training will run 170 | max_fscore: float 171 | max test F1 score that the model will continue to train itself 172 | test_inc_size: int 173 | interval of running a test/evaluation 174 | 175 | """ 176 | 177 | for iter_train in range(0, max_iter+1, batch_size): 178 | batch_set = get_minibatch(train_set, iter_train, batch_size,batch_type="train") 179 | preds, _, _, _, _ = self.forward(batch_set["sga"], batch_set["can"]) 180 | labels = batch_set["deg"] 181 | 182 | self.optimizer.zero_grad() 183 | loss = -torch.log( self.epsilon +1-torch.abs(preds-labels) ).mean() 184 | loss.backward() 185 | self.optimizer.step() 186 | 187 | if test_inc_size and (iter_train % test_inc_size == 0): 188 | labels, preds, _, _, _, _, _ = self.test(test_set, test_batch_size) 189 | precision, recall, f1score, accuracy = evaluate( 190 | labels, preds, epsilon=self.epsilon) 191 | print("[%d,%d], f1_score: %.3f, acc: %.3f"% (iter_train//len(train_set["can"]), 192 | iter_train%len(train_set["can"]), f1score, accuracy)) 193 | 194 | if f1score >= max_fscore: 195 | break 196 | 197 | self.save_model(os.path.join(self.output_dir, "trained_model.pth")) 198 | 199 | 200 | def test(self, test_set, test_batch_size, **kwargs): 201 | """ Run forward process over the given whole test set. 202 | 203 | Parameters 204 | ---------- 205 | test_set: dict 206 | dict of lists, including SGAs, cancer types, DEGs, patient barcodes 207 | test_batch_size: int 208 | 209 | Returns 210 | ------- 211 | labels: 2D array of 0/1 212 | groud truth of gene expression 213 | preds: 2D array of float in [0, 1] 214 | predicted gene expression 215 | hid_tmr: 2D array of float 216 | hidden layer of MLP 217 | emb_tmr: 2D array of float 218 | tumor embedding 219 | emb_sga: 2D array of float 220 | stratified tumor embedding 221 | attn_wt: 2D array of float 222 | attention weights of SGAs 223 | tmr: list of str 224 | barcodes of patients/tumors 225 | 226 | """ 227 | 228 | labels, preds, hid_tmr, emb_tmr, emb_sga, attn_wt, tmr = [], [], [], [], [], [], [] 229 | 230 | for iter_test in range(0, len(test_set["can"]), test_batch_size): 231 | batch_set = get_minibatch(test_set, iter_test, test_batch_size, batch_type="test") 232 | batch_preds, batch_hid_tmr, batch_emb_tmr, batch_emb_sga, batch_attn_wt = self.forward( 233 | batch_set["sga"], batch_set["can"]) 234 | batch_labels = batch_set["deg"] 235 | 236 | labels.append(batch_labels.data.numpy()) 237 | preds.append(batch_preds.data.numpy()) 238 | hid_tmr.append(batch_hid_tmr.data.numpy()) 239 | emb_tmr.append(batch_emb_tmr.data.numpy()) 240 | emb_sga.append(batch_emb_sga.data.numpy()) 241 | attn_wt.append(batch_attn_wt.data.numpy()) 242 | tmr = tmr + batch_set["tmr"] 243 | 244 | labels = np.concatenate(labels,axis=0) 245 | preds = np.concatenate(preds,axis=0) 246 | hid_tmr = np.concatenate(hid_tmr,axis=0) 247 | emb_tmr = np.concatenate(emb_tmr,axis=0) 248 | emb_sga = np.concatenate(emb_sga,axis=0) 249 | attn_wt = np.concatenate(attn_wt,axis=0) 250 | 251 | return labels, preds, hid_tmr, emb_tmr, emb_sga, attn_wt, tmr 252 | 253 | -------------------------------------------------------------------------------- /mydata/cancer_type.txt: -------------------------------------------------------------------------------- 1 | LUSC 2 | KIRP 3 | PRAD 4 | PRAD 5 | COAD 6 | STAD 7 | BRCA 8 | ESCA 9 | LUAD 10 | BLCA 11 | COAD 12 | BRCA 13 | KIRC 14 | LUAD 15 | GBM 16 | KIRP 17 | LIHC 18 | BRCA 19 | LUSC 20 | BRCA 21 | OV 22 | HNSC 23 | BRCA 24 | GBM 25 | BRCA 26 | HNSC 27 | BRCA 28 | LUAD 29 | STAD 30 | PRAD 31 | KIRC 32 | LIHC 33 | KIRC 34 | BRCA 35 | LUAD 36 | PRAD 37 | OV 38 | BRCA 39 | BLCA 40 | HNSC 41 | COAD 42 | HNSC 43 | OV 44 | BRCA 45 | BRCA 46 | PRAD 47 | PRAD 48 | HNSC 49 | LUAD 50 | OV 51 | GBM 52 | KIRP 53 | KIRC 54 | BRCA 55 | BRCA 56 | LIHC 57 | LUAD 58 | PRAD 59 | KIRC 60 | KIRC 61 | HNSC 62 | COAD 63 | STAD 64 | KIRC 65 | STAD 66 | BRCA 67 | LUAD 68 | PRAD 69 | LIHC 70 | ESCA 71 | LUAD 72 | BRCA 73 | LUAD 74 | KIRC 75 | KIRP 76 | OV 77 | OV 78 | HNSC 79 | HNSC 80 | UCEC 81 | PRAD 82 | LUAD 83 | PRAD 84 | OV 85 | LUAD 86 | BLCA 87 | BRCA 88 | BRCA 89 | BRCA 90 | OV 91 | STAD 92 | KIRC 93 | UCEC 94 | LUAD 95 | BRCA 96 | BRCA 97 | KIRC 98 | PRAD 99 | OV 100 | STAD 101 | HNSC 102 | BRCA 103 | LUAD 104 | LIHC 105 | LUAD 106 | KIRC 107 | BRCA 108 | STAD 109 | KIRP 110 | GBM 111 | ESCA 112 | BRCA 113 | BRCA 114 | BRCA 115 | HNSC 116 | HNSC 117 | KIRC 118 | KIRC 119 | LUAD 120 | KIRC 121 | BLCA 122 | GBM 123 | LUSC 124 | KIRC 125 | ESCA 126 | UCEC 127 | BRCA 128 | KIRP 129 | HNSC 130 | HNSC 131 | KIRC 132 | ESCA 133 | COAD 134 | OV 135 | BRCA 136 | LUAD 137 | LUAD 138 | PRAD 139 | BRCA 140 | HNSC 141 | KIRC 142 | PRAD 143 | BLCA 144 | LUAD 145 | BLCA 146 | COAD 147 | OV 148 | HNSC 149 | COAD 150 | UCEC 151 | LUAD 152 | KIRC 153 | COAD 154 | LUSC 155 | OV 156 | BRCA 157 | KIRC 158 | BRCA 159 | BLCA 160 | BLCA 161 | PRAD 162 | GBM 163 | BRCA 164 | UCEC 165 | BRCA 166 | BRCA 167 | HNSC 168 | BRCA 169 | COAD 170 | LIHC 171 | HNSC 172 | BRCA 173 | OV 174 | BRCA 175 | BLCA 176 | HNSC 177 | BRCA 178 | STAD 179 | LUAD 180 | GBM 181 | STAD 182 | KIRP 183 | PRAD 184 | BRCA 185 | OV 186 | LUAD 187 | KIRC 188 | UCEC 189 | PRAD 190 | STAD 191 | HNSC 192 | HNSC 193 | ESCA 194 | LUSC 195 | BLCA 196 | BLCA 197 | HNSC 198 | BRCA 199 | GBM 200 | BRCA 201 | BRCA 202 | BLCA 203 | LUAD 204 | UCEC 205 | HNSC 206 | BRCA 207 | READ 208 | BRCA 209 | HNSC 210 | KIRP 211 | BRCA 212 | STAD 213 | BLCA 214 | BRCA 215 | HNSC 216 | HNSC 217 | PRAD 218 | KIRC 219 | PRAD 220 | UCEC 221 | LUAD 222 | LUAD 223 | LUAD 224 | BRCA 225 | BRCA 226 | OV 227 | KIRC 228 | ESCA 229 | UCEC 230 | PRAD 231 | READ 232 | KIRC 233 | UCEC 234 | OV 235 | BRCA 236 | BRCA 237 | KIRC 238 | COAD 239 | LUAD 240 | STAD 241 | LUAD 242 | OV 243 | KIRC 244 | UCEC 245 | ESCA 246 | PRAD 247 | LUAD 248 | OV 249 | ESCA 250 | HNSC 251 | BRCA 252 | BRCA 253 | READ 254 | OV 255 | PRAD 256 | LUAD 257 | LUAD 258 | LUAD 259 | PRAD 260 | UCEC 261 | BLCA 262 | KIRC 263 | BRCA 264 | COAD 265 | BRCA 266 | LUSC 267 | OV 268 | HNSC 269 | PRAD 270 | HNSC 271 | LIHC 272 | PRAD 273 | BRCA 274 | GBM 275 | BRCA 276 | GBM 277 | LUAD 278 | BRCA 279 | KIRP 280 | OV 281 | HNSC 282 | OV 283 | LUAD 284 | READ 285 | HNSC 286 | ESCA 287 | GBM 288 | BRCA 289 | KIRC 290 | OV 291 | HNSC 292 | LIHC 293 | LUAD 294 | KIRP 295 | BRCA 296 | OV 297 | KIRP 298 | LUAD 299 | KIRC 300 | UCEC 301 | BRCA 302 | BLCA 303 | BRCA 304 | KIRC 305 | PRAD 306 | UCEC 307 | LUAD 308 | PRAD 309 | PRAD 310 | BRCA 311 | COAD 312 | BLCA 313 | KIRP 314 | HNSC 315 | HNSC 316 | PRAD 317 | HNSC 318 | GBM 319 | HNSC 320 | BRCA 321 | COAD 322 | BLCA 323 | BRCA 324 | KIRC 325 | KIRP 326 | OV 327 | BRCA 328 | KIRC 329 | UCEC 330 | PRAD 331 | BLCA 332 | PRAD 333 | UCEC 334 | KIRP 335 | BRCA 336 | OV 337 | PRAD 338 | KIRP 339 | PRAD 340 | HNSC 341 | COAD 342 | STAD 343 | COAD 344 | GBM 345 | BRCA 346 | COAD 347 | COAD 348 | BRCA 349 | PRAD 350 | LUSC 351 | BRCA 352 | KIRP 353 | HNSC 354 | KIRC 355 | OV 356 | ESCA 357 | LUAD 358 | GBM 359 | OV 360 | LUAD 361 | OV 362 | ESCA 363 | BLCA 364 | PRAD 365 | COAD 366 | BRCA 367 | COAD 368 | BRCA 369 | KIRC 370 | KIRC 371 | BRCA 372 | HNSC 373 | GBM 374 | BRCA 375 | BRCA 376 | KIRP 377 | OV 378 | OV 379 | KIRC 380 | UCEC 381 | GBM 382 | STAD 383 | GBM 384 | BLCA 385 | KIRP 386 | BRCA 387 | ESCA 388 | LUAD 389 | PRAD 390 | HNSC 391 | STAD 392 | LUSC 393 | HNSC 394 | KIRC 395 | BRCA 396 | BRCA 397 | BRCA 398 | GBM 399 | HNSC 400 | BRCA 401 | COAD 402 | HNSC 403 | KIRP 404 | PRAD 405 | KIRP 406 | BRCA 407 | BRCA 408 | OV 409 | BLCA 410 | PRAD 411 | BRCA 412 | LUSC 413 | OV 414 | OV 415 | LUAD 416 | KIRC 417 | KIRC 418 | KIRC 419 | KIRC 420 | KIRC 421 | HNSC 422 | READ 423 | BRCA 424 | READ 425 | UCEC 426 | BRCA 427 | BLCA 428 | KIRC 429 | LUAD 430 | PRAD 431 | KIRC 432 | STAD 433 | HNSC 434 | PRAD 435 | KIRC 436 | KIRP 437 | BRCA 438 | BRCA 439 | BRCA 440 | OV 441 | HNSC 442 | LUSC 443 | HNSC 444 | STAD 445 | OV 446 | STAD 447 | KIRC 448 | STAD 449 | LUAD 450 | HNSC 451 | KIRP 452 | BRCA 453 | READ 454 | BRCA 455 | BRCA 456 | OV 457 | PRAD 458 | KIRC 459 | BRCA 460 | LUAD 461 | LUAD 462 | HNSC 463 | BRCA 464 | UCEC 465 | UCEC 466 | HNSC 467 | PRAD 468 | STAD 469 | PRAD 470 | OV 471 | COAD 472 | KIRC 473 | KIRC 474 | BRCA 475 | BLCA 476 | READ 477 | COAD 478 | LUAD 479 | ESCA 480 | HNSC 481 | STAD 482 | KIRP 483 | LIHC 484 | KIRC 485 | BRCA 486 | LUSC 487 | PRAD 488 | LUAD 489 | LUAD 490 | OV 491 | COAD 492 | KIRP 493 | BLCA 494 | BRCA 495 | BRCA 496 | BRCA 497 | PRAD 498 | STAD 499 | HNSC 500 | OV 501 | LUSC 502 | OV 503 | GBM 504 | PRAD 505 | LUAD 506 | BRCA 507 | STAD 508 | GBM 509 | UCEC 510 | HNSC 511 | LUAD 512 | HNSC 513 | LUAD 514 | UCEC 515 | LUAD 516 | BLCA 517 | HNSC 518 | KIRC 519 | KIRC 520 | LUAD 521 | KIRP 522 | ESCA 523 | COAD 524 | HNSC 525 | HNSC 526 | LUAD 527 | OV 528 | BLCA 529 | READ 530 | KIRC 531 | GBM 532 | KIRC 533 | READ 534 | LUSC 535 | LIHC 536 | KIRC 537 | BRCA 538 | BLCA 539 | BRCA 540 | LUSC 541 | LUAD 542 | PRAD 543 | STAD 544 | PRAD 545 | KIRC 546 | BRCA 547 | UCEC 548 | OV 549 | KIRC 550 | UCEC 551 | BRCA 552 | UCEC 553 | BRCA 554 | HNSC 555 | KIRC 556 | ESCA 557 | COAD 558 | LUSC 559 | READ 560 | BLCA 561 | BRCA 562 | KIRC 563 | LIHC 564 | KIRC 565 | UCEC 566 | HNSC 567 | BRCA 568 | OV 569 | BRCA 570 | BRCA 571 | BRCA 572 | HNSC 573 | KIRP 574 | PRAD 575 | BRCA 576 | KIRC 577 | LUAD 578 | PRAD 579 | BRCA 580 | PRAD 581 | HNSC 582 | KIRC 583 | BRCA 584 | PRAD 585 | PRAD 586 | STAD 587 | PRAD 588 | KIRC 589 | KIRC 590 | OV 591 | BRCA 592 | KIRC 593 | COAD 594 | LUAD 595 | HNSC 596 | BRCA 597 | GBM 598 | BRCA 599 | BRCA 600 | OV 601 | LUAD 602 | BRCA 603 | ESCA 604 | BRCA 605 | BRCA 606 | LIHC 607 | LUAD 608 | HNSC 609 | HNSC 610 | UCEC 611 | GBM 612 | BLCA 613 | KIRC 614 | UCEC 615 | LIHC 616 | LUAD 617 | COAD 618 | PRAD 619 | OV 620 | OV 621 | GBM 622 | COAD 623 | HNSC 624 | LUAD 625 | COAD 626 | HNSC 627 | HNSC 628 | BRCA 629 | KIRC 630 | COAD 631 | BRCA 632 | OV 633 | UCEC 634 | GBM 635 | OV 636 | OV 637 | UCEC 638 | PRAD 639 | OV 640 | BRCA 641 | BRCA 642 | GBM 643 | OV 644 | HNSC 645 | BRCA 646 | KIRC 647 | KIRC 648 | BRCA 649 | STAD 650 | KIRP 651 | UCEC 652 | OV 653 | BLCA 654 | GBM 655 | HNSC 656 | KIRP 657 | BRCA 658 | LUAD 659 | PRAD 660 | HNSC 661 | HNSC 662 | KIRP 663 | KIRC 664 | OV 665 | LIHC 666 | BRCA 667 | KIRC 668 | GBM 669 | STAD 670 | OV 671 | BLCA 672 | ESCA 673 | ESCA 674 | KIRP 675 | PRAD 676 | KIRC 677 | PRAD 678 | LUSC 679 | KIRC 680 | BRCA 681 | UCEC 682 | HNSC 683 | KIRP 684 | LUSC 685 | COAD 686 | BRCA 687 | READ 688 | ESCA 689 | PRAD 690 | LUSC 691 | UCEC 692 | LUAD 693 | ESCA 694 | KIRP 695 | STAD 696 | READ 697 | OV 698 | PRAD 699 | HNSC 700 | PRAD 701 | HNSC 702 | LUAD 703 | GBM 704 | HNSC 705 | COAD 706 | ESCA 707 | COAD 708 | BRCA 709 | KIRC 710 | HNSC 711 | READ 712 | KIRC 713 | BRCA 714 | COAD 715 | LIHC 716 | HNSC 717 | KIRC 718 | STAD 719 | UCEC 720 | HNSC 721 | BLCA 722 | HNSC 723 | PRAD 724 | KIRC 725 | BRCA 726 | HNSC 727 | HNSC 728 | ESCA 729 | PRAD 730 | KIRP 731 | BRCA 732 | BRCA 733 | PRAD 734 | KIRC 735 | STAD 736 | GBM 737 | PRAD 738 | ESCA 739 | PRAD 740 | COAD 741 | STAD 742 | BRCA 743 | GBM 744 | UCEC 745 | KIRC 746 | BRCA 747 | GBM 748 | BRCA 749 | LUAD 750 | READ 751 | LUSC 752 | PRAD 753 | KIRC 754 | BRCA 755 | OV 756 | GBM 757 | PRAD 758 | LIHC 759 | STAD 760 | PRAD 761 | LUSC 762 | KIRC 763 | OV 764 | KIRP 765 | LIHC 766 | KIRP 767 | BRCA 768 | LUAD 769 | HNSC 770 | BRCA 771 | OV 772 | COAD 773 | COAD 774 | LUSC 775 | ESCA 776 | PRAD 777 | LUAD 778 | BRCA 779 | HNSC 780 | BRCA 781 | LUAD 782 | BRCA 783 | LUAD 784 | HNSC 785 | LUAD 786 | LUAD 787 | OV 788 | COAD 789 | OV 790 | LUAD 791 | LUAD 792 | BRCA 793 | ESCA 794 | BRCA 795 | HNSC 796 | PRAD 797 | UCEC 798 | LUAD 799 | UCEC 800 | LUAD 801 | HNSC 802 | KIRC 803 | BRCA 804 | STAD 805 | BRCA 806 | BLCA 807 | HNSC 808 | LUAD 809 | KIRC 810 | HNSC 811 | HNSC 812 | LUAD 813 | BRCA 814 | BRCA 815 | KIRC 816 | LUAD 817 | BRCA 818 | PRAD 819 | HNSC 820 | BLCA 821 | ESCA 822 | BRCA 823 | BRCA 824 | OV 825 | BRCA 826 | PRAD 827 | GBM 828 | LUAD 829 | BRCA 830 | STAD 831 | LUSC 832 | KIRP 833 | BRCA 834 | HNSC 835 | READ 836 | KIRC 837 | LUSC 838 | LUSC 839 | LIHC 840 | BLCA 841 | BLCA 842 | OV 843 | LUSC 844 | ESCA 845 | LIHC 846 | PRAD 847 | STAD 848 | PRAD 849 | BRCA 850 | PRAD 851 | HNSC 852 | BRCA 853 | HNSC 854 | OV 855 | LUSC 856 | BLCA 857 | KIRP 858 | KIRC 859 | KIRC 860 | GBM 861 | COAD 862 | LUAD 863 | OV 864 | UCEC 865 | BRCA 866 | KIRC 867 | BRCA 868 | GBM 869 | LUAD 870 | UCEC 871 | BRCA 872 | BRCA 873 | BRCA 874 | LUSC 875 | HNSC 876 | LUSC 877 | BRCA 878 | HNSC 879 | ESCA 880 | GBM 881 | LUSC 882 | BLCA 883 | HNSC 884 | COAD 885 | PRAD 886 | UCEC 887 | KIRP 888 | KIRC 889 | HNSC 890 | KIRC 891 | PRAD 892 | HNSC 893 | KIRC 894 | UCEC 895 | OV 896 | OV 897 | COAD 898 | HNSC 899 | BRCA 900 | LUAD 901 | BRCA 902 | STAD 903 | BRCA 904 | BRCA 905 | HNSC 906 | BLCA 907 | UCEC 908 | ESCA 909 | STAD 910 | KIRP 911 | HNSC 912 | LIHC 913 | BRCA 914 | STAD 915 | LUAD 916 | STAD 917 | KIRP 918 | LUAD 919 | BRCA 920 | BRCA 921 | BRCA 922 | LUAD 923 | HNSC 924 | HNSC 925 | LUAD 926 | STAD 927 | UCEC 928 | KIRC 929 | LUAD 930 | KIRC 931 | KIRC 932 | STAD 933 | BRCA 934 | LUAD 935 | LIHC 936 | BRCA 937 | OV 938 | KIRP 939 | HNSC 940 | READ 941 | LUAD 942 | KIRC 943 | GBM 944 | KIRP 945 | HNSC 946 | STAD 947 | LUSC 948 | OV 949 | BRCA 950 | KIRP 951 | BLCA 952 | KIRC 953 | KIRP 954 | PRAD 955 | OV 956 | KIRC 957 | BRCA 958 | STAD 959 | GBM 960 | LUAD 961 | BRCA 962 | BRCA 963 | BRCA 964 | BRCA 965 | COAD 966 | BRCA 967 | COAD 968 | PRAD 969 | BRCA 970 | HNSC 971 | PRAD 972 | PRAD 973 | GBM 974 | BLCA 975 | HNSC 976 | BRCA 977 | BRCA 978 | KIRC 979 | BRCA 980 | HNSC 981 | BRCA 982 | BRCA 983 | KIRP 984 | BRCA 985 | BRCA 986 | UCEC 987 | PRAD 988 | KIRC 989 | PRAD 990 | HNSC 991 | OV 992 | BRCA 993 | BRCA 994 | LUAD 995 | PRAD 996 | PRAD 997 | BRCA 998 | KIRC 999 | STAD 1000 | KIRP 1001 | BRCA 1002 | PRAD 1003 | BRCA 1004 | BRCA 1005 | LUAD 1006 | ESCA 1007 | ESCA 1008 | OV 1009 | LUAD 1010 | LIHC 1011 | LUAD 1012 | PRAD 1013 | HNSC 1014 | ESCA 1015 | BRCA 1016 | LUAD 1017 | LUAD 1018 | BRCA 1019 | OV 1020 | LUAD 1021 | HNSC 1022 | KIRC 1023 | BRCA 1024 | KIRP 1025 | BRCA 1026 | UCEC 1027 | LUAD 1028 | BLCA 1029 | HNSC 1030 | PRAD 1031 | OV 1032 | BRCA 1033 | BRCA 1034 | BRCA 1035 | GBM 1036 | HNSC 1037 | LUAD 1038 | PRAD 1039 | PRAD 1040 | BRCA 1041 | PRAD 1042 | PRAD 1043 | BRCA 1044 | GBM 1045 | READ 1046 | ESCA 1047 | UCEC 1048 | COAD 1049 | HNSC 1050 | PRAD 1051 | OV 1052 | GBM 1053 | GBM 1054 | UCEC 1055 | BRCA 1056 | KIRC 1057 | OV 1058 | ESCA 1059 | KIRP 1060 | KIRC 1061 | PRAD 1062 | UCEC 1063 | ESCA 1064 | BLCA 1065 | PRAD 1066 | HNSC 1067 | LIHC 1068 | OV 1069 | HNSC 1070 | UCEC 1071 | PRAD 1072 | HNSC 1073 | BLCA 1074 | READ 1075 | BLCA 1076 | LUAD 1077 | KIRC 1078 | HNSC 1079 | UCEC 1080 | LUAD 1081 | KIRC 1082 | HNSC 1083 | GBM 1084 | BRCA 1085 | BLCA 1086 | HNSC 1087 | LIHC 1088 | LUAD 1089 | LUAD 1090 | READ 1091 | LIHC 1092 | HNSC 1093 | HNSC 1094 | OV 1095 | BRCA 1096 | UCEC 1097 | PRAD 1098 | UCEC 1099 | KIRC 1100 | BLCA 1101 | BRCA 1102 | KIRC 1103 | STAD 1104 | HNSC 1105 | HNSC 1106 | COAD 1107 | HNSC 1108 | LUAD 1109 | ESCA 1110 | BRCA 1111 | OV 1112 | OV 1113 | ESCA 1114 | HNSC 1115 | KIRC 1116 | KIRC 1117 | LUSC 1118 | GBM 1119 | BRCA 1120 | OV 1121 | ESCA 1122 | HNSC 1123 | PRAD 1124 | COAD 1125 | OV 1126 | LUAD 1127 | GBM 1128 | GBM 1129 | BRCA 1130 | KIRP 1131 | BLCA 1132 | UCEC 1133 | GBM 1134 | HNSC 1135 | KIRC 1136 | BRCA 1137 | BLCA 1138 | PRAD 1139 | KIRP 1140 | BRCA 1141 | COAD 1142 | KIRC 1143 | PRAD 1144 | LIHC 1145 | KIRP 1146 | LUSC 1147 | BRCA 1148 | GBM 1149 | LUSC 1150 | KIRC 1151 | BLCA 1152 | BRCA 1153 | OV 1154 | LUSC 1155 | STAD 1156 | KIRC 1157 | KIRC 1158 | BRCA 1159 | PRAD 1160 | HNSC 1161 | BRCA 1162 | UCEC 1163 | PRAD 1164 | COAD 1165 | BRCA 1166 | BRCA 1167 | STAD 1168 | GBM 1169 | OV 1170 | STAD 1171 | UCEC 1172 | OV 1173 | BRCA 1174 | OV 1175 | OV 1176 | HNSC 1177 | BRCA 1178 | KIRC 1179 | OV 1180 | OV 1181 | HNSC 1182 | HNSC 1183 | HNSC 1184 | BLCA 1185 | HNSC 1186 | HNSC 1187 | STAD 1188 | COAD 1189 | OV 1190 | KIRP 1191 | LUAD 1192 | KIRC 1193 | STAD 1194 | BRCA 1195 | BRCA 1196 | BLCA 1197 | BRCA 1198 | BRCA 1199 | READ 1200 | KIRC 1201 | READ 1202 | STAD 1203 | LUAD 1204 | ESCA 1205 | BRCA 1206 | HNSC 1207 | COAD 1208 | LUSC 1209 | OV 1210 | OV 1211 | LUSC 1212 | BRCA 1213 | KIRP 1214 | BRCA 1215 | LUAD 1216 | BRCA 1217 | BRCA 1218 | KIRC 1219 | HNSC 1220 | LUAD 1221 | HNSC 1222 | PRAD 1223 | UCEC 1224 | BRCA 1225 | KIRP 1226 | PRAD 1227 | KIRC 1228 | ESCA 1229 | LUAD 1230 | PRAD 1231 | BRCA 1232 | KIRC 1233 | PRAD 1234 | HNSC 1235 | BLCA 1236 | OV 1237 | BRCA 1238 | COAD 1239 | BRCA 1240 | OV 1241 | LUAD 1242 | READ 1243 | BRCA 1244 | BRCA 1245 | OV 1246 | HNSC 1247 | KIRP 1248 | UCEC 1249 | GBM 1250 | UCEC 1251 | LUAD 1252 | LUAD 1253 | LUAD 1254 | UCEC 1255 | LUAD 1256 | BRCA 1257 | COAD 1258 | KIRC 1259 | HNSC 1260 | OV 1261 | KIRC 1262 | GBM 1263 | STAD 1264 | LUSC 1265 | LIHC 1266 | PRAD 1267 | BRCA 1268 | LIHC 1269 | GBM 1270 | HNSC 1271 | PRAD 1272 | LUAD 1273 | HNSC 1274 | OV 1275 | STAD 1276 | KIRC 1277 | LUAD 1278 | KIRP 1279 | BRCA 1280 | BLCA 1281 | BRCA 1282 | BLCA 1283 | HNSC 1284 | LUAD 1285 | KIRC 1286 | OV 1287 | HNSC 1288 | LIHC 1289 | STAD 1290 | KIRC 1291 | PRAD 1292 | BRCA 1293 | HNSC 1294 | PRAD 1295 | BRCA 1296 | BLCA 1297 | PRAD 1298 | STAD 1299 | LUAD 1300 | GBM 1301 | KIRC 1302 | HNSC 1303 | HNSC 1304 | PRAD 1305 | PRAD 1306 | BLCA 1307 | BLCA 1308 | HNSC 1309 | ESCA 1310 | LUAD 1311 | KIRC 1312 | BRCA 1313 | PRAD 1314 | LUAD 1315 | OV 1316 | KIRP 1317 | GBM 1318 | OV 1319 | BRCA 1320 | BRCA 1321 | BLCA 1322 | LUSC 1323 | LUAD 1324 | LUAD 1325 | OV 1326 | LUSC 1327 | LUAD 1328 | BLCA 1329 | BRCA 1330 | COAD 1331 | UCEC 1332 | HNSC 1333 | LUAD 1334 | HNSC 1335 | OV 1336 | HNSC 1337 | OV 1338 | LIHC 1339 | KIRC 1340 | STAD 1341 | LUSC 1342 | ESCA 1343 | LUAD 1344 | OV 1345 | LUAD 1346 | HNSC 1347 | COAD 1348 | ESCA 1349 | COAD 1350 | PRAD 1351 | ESCA 1352 | HNSC 1353 | COAD 1354 | LUSC 1355 | KIRC 1356 | BRCA 1357 | OV 1358 | BRCA 1359 | ESCA 1360 | STAD 1361 | PRAD 1362 | OV 1363 | OV 1364 | KIRC 1365 | KIRP 1366 | LIHC 1367 | KIRC 1368 | KIRC 1369 | READ 1370 | KIRP 1371 | KIRP 1372 | LIHC 1373 | UCEC 1374 | LUSC 1375 | BRCA 1376 | KIRC 1377 | KIRC 1378 | KIRC 1379 | KIRC 1380 | GBM 1381 | BRCA 1382 | BRCA 1383 | BRCA 1384 | BLCA 1385 | BRCA 1386 | KIRC 1387 | READ 1388 | KIRC 1389 | KIRC 1390 | HNSC 1391 | PRAD 1392 | KIRP 1393 | STAD 1394 | LIHC 1395 | PRAD 1396 | KIRC 1397 | ESCA 1398 | BLCA 1399 | PRAD 1400 | BRCA 1401 | COAD 1402 | HNSC 1403 | READ 1404 | COAD 1405 | GBM 1406 | BRCA 1407 | OV 1408 | KIRP 1409 | LUSC 1410 | BRCA 1411 | HNSC 1412 | UCEC 1413 | KIRC 1414 | HNSC 1415 | BLCA 1416 | LUAD 1417 | COAD 1418 | LUAD 1419 | BRCA 1420 | STAD 1421 | OV 1422 | LUAD 1423 | PRAD 1424 | BLCA 1425 | GBM 1426 | UCEC 1427 | HNSC 1428 | BRCA 1429 | KIRC 1430 | BRCA 1431 | BRCA 1432 | OV 1433 | BLCA 1434 | BRCA 1435 | LUSC 1436 | STAD 1437 | PRAD 1438 | GBM 1439 | BRCA 1440 | KIRP 1441 | PRAD 1442 | PRAD 1443 | HNSC 1444 | HNSC 1445 | LIHC 1446 | KIRC 1447 | LUSC 1448 | BRCA 1449 | OV 1450 | READ 1451 | UCEC 1452 | LUAD 1453 | COAD 1454 | OV 1455 | LUAD 1456 | BLCA 1457 | UCEC 1458 | HNSC 1459 | KIRC 1460 | BRCA 1461 | PRAD 1462 | HNSC 1463 | LUAD 1464 | KIRP 1465 | HNSC 1466 | BRCA 1467 | PRAD 1468 | KIRC 1469 | LIHC 1470 | READ 1471 | COAD 1472 | BRCA 1473 | KIRC 1474 | ESCA 1475 | BRCA 1476 | OV 1477 | COAD 1478 | BRCA 1479 | BRCA 1480 | LUSC 1481 | GBM 1482 | ESCA 1483 | LIHC 1484 | PRAD 1485 | LUSC 1486 | KIRC 1487 | LUSC 1488 | LUAD 1489 | OV 1490 | GBM 1491 | PRAD 1492 | KIRC 1493 | HNSC 1494 | STAD 1495 | HNSC 1496 | LUSC 1497 | KIRC 1498 | PRAD 1499 | ESCA 1500 | LUSC 1501 | LUAD 1502 | KIRC 1503 | STAD 1504 | LIHC 1505 | HNSC 1506 | LUSC 1507 | BRCA 1508 | PRAD 1509 | LUAD 1510 | OV 1511 | KIRC 1512 | HNSC 1513 | BRCA 1514 | OV 1515 | BRCA 1516 | BLCA 1517 | LUAD 1518 | LUAD 1519 | COAD 1520 | KIRC 1521 | UCEC 1522 | KIRC 1523 | COAD 1524 | LUAD 1525 | BRCA 1526 | READ 1527 | BLCA 1528 | KIRC 1529 | KIRC 1530 | KIRC 1531 | BLCA 1532 | OV 1533 | STAD 1534 | GBM 1535 | HNSC 1536 | COAD 1537 | PRAD 1538 | PRAD 1539 | STAD 1540 | KIRC 1541 | OV 1542 | OV 1543 | OV 1544 | OV 1545 | GBM 1546 | HNSC 1547 | GBM 1548 | KIRC 1549 | BLCA 1550 | PRAD 1551 | LUAD 1552 | BRCA 1553 | STAD 1554 | GBM 1555 | PRAD 1556 | BRCA 1557 | KIRC 1558 | KIRC 1559 | KIRP 1560 | OV 1561 | PRAD 1562 | LUAD 1563 | BRCA 1564 | OV 1565 | GBM 1566 | STAD 1567 | ESCA 1568 | BRCA 1569 | PRAD 1570 | BRCA 1571 | LUSC 1572 | BRCA 1573 | HNSC 1574 | PRAD 1575 | OV 1576 | BRCA 1577 | COAD 1578 | BRCA 1579 | PRAD 1580 | KIRC 1581 | KIRC 1582 | OV 1583 | HNSC 1584 | BRCA 1585 | LUAD 1586 | KIRC 1587 | OV 1588 | HNSC 1589 | KIRP 1590 | HNSC 1591 | ESCA 1592 | LUAD 1593 | COAD 1594 | KIRC 1595 | UCEC 1596 | UCEC 1597 | KIRC 1598 | HNSC 1599 | BRCA 1600 | BLCA 1601 | BRCA 1602 | KIRC 1603 | PRAD 1604 | KIRC 1605 | BLCA 1606 | STAD 1607 | LIHC 1608 | COAD 1609 | KIRP 1610 | PRAD 1611 | PRAD 1612 | KIRC 1613 | LUAD 1614 | LUAD 1615 | HNSC 1616 | BRCA 1617 | LUAD 1618 | LUAD 1619 | LUAD 1620 | HNSC 1621 | KIRC 1622 | KIRC 1623 | BRCA 1624 | KIRC 1625 | BRCA 1626 | BRCA 1627 | KIRC 1628 | LUSC 1629 | OV 1630 | KIRC 1631 | READ 1632 | BRCA 1633 | STAD 1634 | BRCA 1635 | BLCA 1636 | KIRP 1637 | BRCA 1638 | COAD 1639 | LUAD 1640 | HNSC 1641 | BRCA 1642 | OV 1643 | LUAD 1644 | BRCA 1645 | HNSC 1646 | KIRP 1647 | LUAD 1648 | GBM 1649 | PRAD 1650 | BRCA 1651 | OV 1652 | PRAD 1653 | KIRP 1654 | GBM 1655 | HNSC 1656 | BRCA 1657 | COAD 1658 | LUAD 1659 | BRCA 1660 | PRAD 1661 | GBM 1662 | ESCA 1663 | BRCA 1664 | PRAD 1665 | GBM 1666 | BRCA 1667 | LUAD 1668 | ESCA 1669 | UCEC 1670 | LUAD 1671 | KIRP 1672 | LUSC 1673 | BRCA 1674 | GBM 1675 | PRAD 1676 | LUAD 1677 | STAD 1678 | HNSC 1679 | HNSC 1680 | BLCA 1681 | COAD 1682 | KIRC 1683 | GBM 1684 | BRCA 1685 | PRAD 1686 | GBM 1687 | BLCA 1688 | OV 1689 | BRCA 1690 | BRCA 1691 | KIRP 1692 | PRAD 1693 | ESCA 1694 | BRCA 1695 | UCEC 1696 | OV 1697 | PRAD 1698 | LIHC 1699 | PRAD 1700 | KIRP 1701 | HNSC 1702 | LUAD 1703 | GBM 1704 | UCEC 1705 | KIRP 1706 | BRCA 1707 | LUAD 1708 | OV 1709 | BRCA 1710 | PRAD 1711 | PRAD 1712 | PRAD 1713 | LUAD 1714 | STAD 1715 | UCEC 1716 | STAD 1717 | UCEC 1718 | BLCA 1719 | KIRC 1720 | BRCA 1721 | KIRC 1722 | HNSC 1723 | LUSC 1724 | LUAD 1725 | PRAD 1726 | HNSC 1727 | ESCA 1728 | PRAD 1729 | READ 1730 | UCEC 1731 | PRAD 1732 | LUSC 1733 | PRAD 1734 | BRCA 1735 | BRCA 1736 | BLCA 1737 | LUAD 1738 | ESCA 1739 | ESCA 1740 | BRCA 1741 | PRAD 1742 | KIRP 1743 | LUAD 1744 | BRCA 1745 | KIRC 1746 | KIRP 1747 | BRCA 1748 | READ 1749 | BRCA 1750 | LUAD 1751 | OV 1752 | BRCA 1753 | READ 1754 | PRAD 1755 | LIHC 1756 | READ 1757 | PRAD 1758 | GBM 1759 | READ 1760 | BRCA 1761 | PRAD 1762 | UCEC 1763 | UCEC 1764 | GBM 1765 | PRAD 1766 | OV 1767 | GBM 1768 | HNSC 1769 | PRAD 1770 | KIRC 1771 | BRCA 1772 | PRAD 1773 | BRCA 1774 | BRCA 1775 | COAD 1776 | COAD 1777 | LUAD 1778 | HNSC 1779 | OV 1780 | BLCA 1781 | BRCA 1782 | BRCA 1783 | KIRC 1784 | PRAD 1785 | BLCA 1786 | BRCA 1787 | HNSC 1788 | LUAD 1789 | PRAD 1790 | LIHC 1791 | GBM 1792 | HNSC 1793 | BLCA 1794 | BRCA 1795 | BLCA 1796 | LUAD 1797 | BRCA 1798 | GBM 1799 | PRAD 1800 | UCEC 1801 | LUAD 1802 | BRCA 1803 | BRCA 1804 | READ 1805 | BRCA 1806 | BRCA 1807 | UCEC 1808 | PRAD 1809 | KIRP 1810 | GBM 1811 | ESCA 1812 | ESCA 1813 | PRAD 1814 | LUSC 1815 | COAD 1816 | HNSC 1817 | BRCA 1818 | LIHC 1819 | BRCA 1820 | HNSC 1821 | HNSC 1822 | GBM 1823 | BRCA 1824 | BLCA 1825 | KIRC 1826 | BLCA 1827 | LUAD 1828 | STAD 1829 | STAD 1830 | KIRC 1831 | BRCA 1832 | HNSC 1833 | BRCA 1834 | BRCA 1835 | COAD 1836 | BLCA 1837 | BRCA 1838 | PRAD 1839 | BRCA 1840 | LUAD 1841 | KIRC 1842 | BLCA 1843 | COAD 1844 | READ 1845 | KIRC 1846 | LUAD 1847 | LUAD 1848 | HNSC 1849 | LUAD 1850 | LUAD 1851 | ESCA 1852 | BRCA 1853 | HNSC 1854 | STAD 1855 | LIHC 1856 | HNSC 1857 | READ 1858 | BRCA 1859 | UCEC 1860 | HNSC 1861 | BRCA 1862 | ESCA 1863 | HNSC 1864 | STAD 1865 | GBM 1866 | LUAD 1867 | PRAD 1868 | LUAD 1869 | READ 1870 | ESCA 1871 | ESCA 1872 | GBM 1873 | KIRP 1874 | GBM 1875 | BRCA 1876 | PRAD 1877 | BRCA 1878 | PRAD 1879 | PRAD 1880 | UCEC 1881 | READ 1882 | OV 1883 | BRCA 1884 | HNSC 1885 | KIRP 1886 | KIRC 1887 | BRCA 1888 | BLCA 1889 | COAD 1890 | KIRC 1891 | HNSC 1892 | PRAD 1893 | BRCA 1894 | BLCA 1895 | OV 1896 | KIRC 1897 | OV 1898 | LUSC 1899 | OV 1900 | ESCA 1901 | READ 1902 | BRCA 1903 | KIRC 1904 | UCEC 1905 | LUAD 1906 | GBM 1907 | KIRP 1908 | KIRC 1909 | BRCA 1910 | KIRC 1911 | OV 1912 | LUAD 1913 | COAD 1914 | BRCA 1915 | OV 1916 | BRCA 1917 | HNSC 1918 | BRCA 1919 | BLCA 1920 | BLCA 1921 | OV 1922 | COAD 1923 | PRAD 1924 | HNSC 1925 | LIHC 1926 | BRCA 1927 | BRCA 1928 | BLCA 1929 | BRCA 1930 | OV 1931 | ESCA 1932 | GBM 1933 | BLCA 1934 | COAD 1935 | LUAD 1936 | COAD 1937 | UCEC 1938 | BRCA 1939 | BRCA 1940 | PRAD 1941 | BRCA 1942 | GBM 1943 | LUAD 1944 | BRCA 1945 | KIRC 1946 | PRAD 1947 | BRCA 1948 | BLCA 1949 | BRCA 1950 | OV 1951 | BRCA 1952 | BRCA 1953 | HNSC 1954 | KIRC 1955 | BRCA 1956 | OV 1957 | BRCA 1958 | LUAD 1959 | LUAD 1960 | PRAD 1961 | OV 1962 | PRAD 1963 | BRCA 1964 | KIRC 1965 | PRAD 1966 | STAD 1967 | LUAD 1968 | HNSC 1969 | PRAD 1970 | READ 1971 | BRCA 1972 | PRAD 1973 | KIRC 1974 | STAD 1975 | COAD 1976 | BRCA 1977 | HNSC 1978 | LIHC 1979 | GBM 1980 | STAD 1981 | LUSC 1982 | OV 1983 | KIRP 1984 | COAD 1985 | GBM 1986 | BRCA 1987 | PRAD 1988 | BRCA 1989 | PRAD 1990 | KIRC 1991 | PRAD 1992 | PRAD 1993 | KIRC 1994 | BRCA 1995 | COAD 1996 | READ 1997 | LIHC 1998 | HNSC 1999 | STAD 2000 | LUAD 2001 | BRCA 2002 | BLCA 2003 | STAD 2004 | LUSC 2005 | LUAD 2006 | LUAD 2007 | GBM 2008 | ESCA 2009 | KIRC 2010 | OV 2011 | BRCA 2012 | OV 2013 | HNSC 2014 | HNSC 2015 | PRAD 2016 | GBM 2017 | ESCA 2018 | GBM 2019 | LUAD 2020 | BRCA 2021 | KIRP 2022 | BRCA 2023 | BRCA 2024 | STAD 2025 | BRCA 2026 | PRAD 2027 | BRCA 2028 | LUAD 2029 | BRCA 2030 | STAD 2031 | LUAD 2032 | BLCA 2033 | BRCA 2034 | LUAD 2035 | UCEC 2036 | OV 2037 | LUSC 2038 | LUAD 2039 | UCEC 2040 | LIHC 2041 | PRAD 2042 | STAD 2043 | KIRC 2044 | KIRC 2045 | BRCA 2046 | READ 2047 | STAD 2048 | LUAD 2049 | HNSC 2050 | KIRP 2051 | OV 2052 | KIRP 2053 | LUSC 2054 | LIHC 2055 | BRCA 2056 | BLCA 2057 | LUAD 2058 | BRCA 2059 | KIRC 2060 | KIRC 2061 | HNSC 2062 | KIRC 2063 | BLCA 2064 | BRCA 2065 | READ 2066 | KIRC 2067 | ESCA 2068 | BLCA 2069 | LIHC 2070 | BRCA 2071 | LUSC 2072 | KIRC 2073 | GBM 2074 | BRCA 2075 | UCEC 2076 | GBM 2077 | BRCA 2078 | STAD 2079 | BRCA 2080 | KIRC 2081 | KIRC 2082 | KIRP 2083 | LUAD 2084 | HNSC 2085 | KIRP 2086 | OV 2087 | BLCA 2088 | ESCA 2089 | HNSC 2090 | COAD 2091 | HNSC 2092 | OV 2093 | PRAD 2094 | LUAD 2095 | COAD 2096 | ESCA 2097 | GBM 2098 | PRAD 2099 | BRCA 2100 | PRAD 2101 | LIHC 2102 | GBM 2103 | LUAD 2104 | UCEC 2105 | OV 2106 | OV 2107 | BRCA 2108 | GBM 2109 | LIHC 2110 | HNSC 2111 | BRCA 2112 | HNSC 2113 | READ 2114 | GBM 2115 | GBM 2116 | BRCA 2117 | LIHC 2118 | COAD 2119 | UCEC 2120 | BRCA 2121 | HNSC 2122 | LUSC 2123 | UCEC 2124 | COAD 2125 | BRCA 2126 | LUAD 2127 | COAD 2128 | PRAD 2129 | PRAD 2130 | LUAD 2131 | READ 2132 | KIRC 2133 | OV 2134 | HNSC 2135 | HNSC 2136 | BRCA 2137 | STAD 2138 | KIRC 2139 | LIHC 2140 | UCEC 2141 | STAD 2142 | BRCA 2143 | STAD 2144 | KIRC 2145 | BRCA 2146 | LIHC 2147 | HNSC 2148 | LUSC 2149 | BRCA 2150 | OV 2151 | HNSC 2152 | PRAD 2153 | HNSC 2154 | BRCA 2155 | HNSC 2156 | KIRP 2157 | OV 2158 | LUSC 2159 | LUAD 2160 | KIRP 2161 | PRAD 2162 | KIRC 2163 | BRCA 2164 | BRCA 2165 | KIRC 2166 | PRAD 2167 | PRAD 2168 | KIRP 2169 | KIRC 2170 | BLCA 2171 | KIRC 2172 | UCEC 2173 | OV 2174 | OV 2175 | BRCA 2176 | GBM 2177 | LIHC 2178 | UCEC 2179 | HNSC 2180 | UCEC 2181 | LUSC 2182 | BRCA 2183 | PRAD 2184 | HNSC 2185 | ESCA 2186 | HNSC 2187 | LUAD 2188 | HNSC 2189 | KIRC 2190 | KIRP 2191 | LUAD 2192 | KIRP 2193 | PRAD 2194 | PRAD 2195 | UCEC 2196 | OV 2197 | COAD 2198 | KIRC 2199 | BRCA 2200 | BRCA 2201 | KIRC 2202 | BRCA 2203 | HNSC 2204 | KIRP 2205 | COAD 2206 | GBM 2207 | LUAD 2208 | BLCA 2209 | LUSC 2210 | HNSC 2211 | GBM 2212 | COAD 2213 | UCEC 2214 | LUAD 2215 | PRAD 2216 | BRCA 2217 | HNSC 2218 | UCEC 2219 | OV 2220 | HNSC 2221 | LUAD 2222 | LIHC 2223 | PRAD 2224 | BRCA 2225 | HNSC 2226 | BRCA 2227 | HNSC 2228 | PRAD 2229 | PRAD 2230 | LUSC 2231 | LUAD 2232 | COAD 2233 | LUAD 2234 | UCEC 2235 | STAD 2236 | KIRC 2237 | OV 2238 | COAD 2239 | OV 2240 | HNSC 2241 | HNSC 2242 | OV 2243 | BLCA 2244 | PRAD 2245 | LUAD 2246 | PRAD 2247 | KIRC 2248 | COAD 2249 | BRCA 2250 | BRCA 2251 | BRCA 2252 | PRAD 2253 | KIRC 2254 | STAD 2255 | OV 2256 | BRCA 2257 | COAD 2258 | BRCA 2259 | HNSC 2260 | KIRC 2261 | ESCA 2262 | KIRP 2263 | BLCA 2264 | OV 2265 | BRCA 2266 | OV 2267 | LUAD 2268 | UCEC 2269 | GBM 2270 | BRCA 2271 | HNSC 2272 | KIRC 2273 | STAD 2274 | LUAD 2275 | UCEC 2276 | BRCA 2277 | BRCA 2278 | COAD 2279 | BRCA 2280 | KIRP 2281 | PRAD 2282 | UCEC 2283 | BRCA 2284 | PRAD 2285 | BRCA 2286 | ESCA 2287 | PRAD 2288 | OV 2289 | UCEC 2290 | BRCA 2291 | STAD 2292 | COAD 2293 | BRCA 2294 | BRCA 2295 | ESCA 2296 | PRAD 2297 | PRAD 2298 | LUSC 2299 | PRAD 2300 | BRCA 2301 | STAD 2302 | OV 2303 | STAD 2304 | READ 2305 | HNSC 2306 | BRCA 2307 | LIHC 2308 | ESCA 2309 | KIRP 2310 | HNSC 2311 | BLCA 2312 | BRCA 2313 | COAD 2314 | HNSC 2315 | HNSC 2316 | BLCA 2317 | PRAD 2318 | BRCA 2319 | KIRC 2320 | ESCA 2321 | LUAD 2322 | PRAD 2323 | UCEC 2324 | BRCA 2325 | OV 2326 | LIHC 2327 | STAD 2328 | HNSC 2329 | PRAD 2330 | GBM 2331 | LUSC 2332 | HNSC 2333 | ESCA 2334 | PRAD 2335 | LUSC 2336 | GBM 2337 | KIRC 2338 | ESCA 2339 | KIRP 2340 | ESCA 2341 | BRCA 2342 | HNSC 2343 | OV 2344 | HNSC 2345 | LIHC 2346 | BRCA 2347 | BRCA 2348 | BLCA 2349 | COAD 2350 | HNSC 2351 | BRCA 2352 | HNSC 2353 | BRCA 2354 | HNSC 2355 | BRCA 2356 | HNSC 2357 | KIRC 2358 | BRCA 2359 | COAD 2360 | BRCA 2361 | PRAD 2362 | BRCA 2363 | PRAD 2364 | PRAD 2365 | BRCA 2366 | GBM 2367 | OV 2368 | BRCA 2369 | OV 2370 | PRAD 2371 | COAD 2372 | BRCA 2373 | LUSC 2374 | OV 2375 | HNSC 2376 | GBM 2377 | HNSC 2378 | KIRP 2379 | LUAD 2380 | STAD 2381 | KIRC 2382 | OV 2383 | LUAD 2384 | BRCA 2385 | KIRC 2386 | BRCA 2387 | KIRP 2388 | HNSC 2389 | KIRC 2390 | STAD 2391 | ESCA 2392 | BRCA 2393 | PRAD 2394 | KIRC 2395 | HNSC 2396 | PRAD 2397 | BRCA 2398 | HNSC 2399 | BRCA 2400 | BRCA 2401 | OV 2402 | BRCA 2403 | COAD 2404 | HNSC 2405 | BRCA 2406 | LIHC 2407 | KIRP 2408 | OV 2409 | LUAD 2410 | BRCA 2411 | OV 2412 | BRCA 2413 | LUAD 2414 | BLCA 2415 | HNSC 2416 | KIRP 2417 | LUSC 2418 | HNSC 2419 | ESCA 2420 | GBM 2421 | LUSC 2422 | BRCA 2423 | BRCA 2424 | LUAD 2425 | LIHC 2426 | STAD 2427 | LUAD 2428 | READ 2429 | BRCA 2430 | OV 2431 | COAD 2432 | HNSC 2433 | BLCA 2434 | LIHC 2435 | ESCA 2436 | ESCA 2437 | LUAD 2438 | LUSC 2439 | READ 2440 | KIRP 2441 | HNSC 2442 | HNSC 2443 | OV 2444 | LUAD 2445 | PRAD 2446 | OV 2447 | BRCA 2448 | LUAD 2449 | BRCA 2450 | LIHC 2451 | STAD 2452 | BRCA 2453 | STAD 2454 | BRCA 2455 | GBM 2456 | COAD 2457 | BRCA 2458 | OV 2459 | PRAD 2460 | LUAD 2461 | PRAD 2462 | HNSC 2463 | LUSC 2464 | UCEC 2465 | HNSC 2466 | COAD 2467 | KIRC 2468 | BRCA 2469 | BLCA 2470 | BRCA 2471 | BRCA 2472 | KIRP 2473 | COAD 2474 | ESCA 2475 | BRCA 2476 | PRAD 2477 | BRCA 2478 | HNSC 2479 | KIRC 2480 | KIRC 2481 | STAD 2482 | GBM 2483 | BRCA 2484 | STAD 2485 | PRAD 2486 | OV 2487 | BRCA 2488 | KIRC 2489 | LUAD 2490 | COAD 2491 | LIHC 2492 | ESCA 2493 | LUSC 2494 | PRAD 2495 | LUSC 2496 | LUSC 2497 | GBM 2498 | BLCA 2499 | HNSC 2500 | HNSC 2501 | GBM 2502 | BRCA 2503 | LUAD 2504 | LUAD 2505 | KIRC 2506 | OV 2507 | BRCA 2508 | BRCA 2509 | HNSC 2510 | ESCA 2511 | BRCA 2512 | BRCA 2513 | HNSC 2514 | STAD 2515 | OV 2516 | KIRP 2517 | HNSC 2518 | BRCA 2519 | HNSC 2520 | PRAD 2521 | LUSC 2522 | STAD 2523 | COAD 2524 | LUSC 2525 | COAD 2526 | READ 2527 | HNSC 2528 | STAD 2529 | HNSC 2530 | LUAD 2531 | UCEC 2532 | PRAD 2533 | BRCA 2534 | COAD 2535 | KIRP 2536 | GBM 2537 | LUAD 2538 | LIHC 2539 | PRAD 2540 | PRAD 2541 | BRCA 2542 | OV 2543 | PRAD 2544 | BRCA 2545 | OV 2546 | HNSC 2547 | OV 2548 | BRCA 2549 | KIRC 2550 | KIRC 2551 | GBM 2552 | HNSC 2553 | BRCA 2554 | BRCA 2555 | HNSC 2556 | GBM 2557 | BRCA 2558 | KIRP 2559 | BRCA 2560 | BRCA 2561 | LUAD 2562 | BRCA 2563 | KIRC 2564 | LIHC 2565 | KIRC 2566 | KIRP 2567 | HNSC 2568 | BLCA 2569 | BRCA 2570 | PRAD 2571 | LUAD 2572 | OV 2573 | LUSC 2574 | BRCA 2575 | BRCA 2576 | KIRP 2577 | UCEC 2578 | KIRC 2579 | GBM 2580 | GBM 2581 | BRCA 2582 | BRCA 2583 | LUAD 2584 | ESCA 2585 | PRAD 2586 | BLCA 2587 | STAD 2588 | HNSC 2589 | BRCA 2590 | COAD 2591 | LUAD 2592 | BRCA 2593 | BRCA 2594 | HNSC 2595 | BRCA 2596 | BRCA 2597 | LUAD 2598 | LUAD 2599 | LUAD 2600 | PRAD 2601 | PRAD 2602 | ESCA 2603 | HNSC 2604 | BRCA 2605 | PRAD 2606 | KIRP 2607 | BRCA 2608 | GBM 2609 | COAD 2610 | BRCA 2611 | BRCA 2612 | KIRP 2613 | HNSC 2614 | PRAD 2615 | KIRP 2616 | KIRC 2617 | LIHC 2618 | COAD 2619 | UCEC 2620 | GBM 2621 | KIRC 2622 | KIRC 2623 | BLCA 2624 | KIRP 2625 | OV 2626 | BRCA 2627 | PRAD 2628 | GBM 2629 | GBM 2630 | BRCA 2631 | COAD 2632 | LUAD 2633 | GBM 2634 | KIRP 2635 | BRCA 2636 | HNSC 2637 | UCEC 2638 | BRCA 2639 | KIRC 2640 | COAD 2641 | BLCA 2642 | PRAD 2643 | UCEC 2644 | GBM 2645 | OV 2646 | KIRP 2647 | OV 2648 | HNSC 2649 | LUAD 2650 | LIHC 2651 | HNSC 2652 | COAD 2653 | HNSC 2654 | LIHC 2655 | STAD 2656 | KIRP 2657 | LIHC 2658 | OV 2659 | BRCA 2660 | LIHC 2661 | UCEC 2662 | STAD 2663 | BRCA 2664 | STAD 2665 | BLCA 2666 | BRCA 2667 | HNSC 2668 | BRCA 2669 | PRAD 2670 | STAD 2671 | BRCA 2672 | KIRC 2673 | BRCA 2674 | LUSC 2675 | OV 2676 | LUSC 2677 | PRAD 2678 | BRCA 2679 | KIRP 2680 | BRCA 2681 | LUSC 2682 | LUAD 2683 | BRCA 2684 | BRCA 2685 | LUAD 2686 | KIRC 2687 | BLCA 2688 | BRCA 2689 | LIHC 2690 | BRCA 2691 | OV 2692 | STAD 2693 | PRAD 2694 | LUAD 2695 | LUAD 2696 | LUSC 2697 | BRCA 2698 | ESCA 2699 | LUAD 2700 | OV 2701 | COAD 2702 | HNSC 2703 | UCEC 2704 | BRCA 2705 | HNSC 2706 | OV 2707 | COAD 2708 | BRCA 2709 | ESCA 2710 | BRCA 2711 | STAD 2712 | LUAD 2713 | OV 2714 | PRAD 2715 | KIRC 2716 | BRCA 2717 | READ 2718 | ESCA 2719 | KIRC 2720 | LUAD 2721 | COAD 2722 | LUAD 2723 | GBM 2724 | LIHC 2725 | HNSC 2726 | PRAD 2727 | BRCA 2728 | HNSC 2729 | BRCA 2730 | STAD 2731 | STAD 2732 | BLCA 2733 | LUAD 2734 | OV 2735 | ESCA 2736 | BRCA 2737 | PRAD 2738 | BRCA 2739 | PRAD 2740 | STAD 2741 | LUAD 2742 | OV 2743 | BRCA 2744 | BRCA 2745 | BRCA 2746 | BRCA 2747 | BLCA 2748 | BRCA 2749 | ESCA 2750 | KIRC 2751 | HNSC 2752 | KIRC 2753 | BRCA 2754 | LUAD 2755 | HNSC 2756 | KIRP 2757 | ESCA 2758 | BLCA 2759 | HNSC 2760 | BRCA 2761 | KIRC 2762 | HNSC 2763 | GBM 2764 | KIRC 2765 | KIRC 2766 | BRCA 2767 | HNSC 2768 | GBM 2769 | OV 2770 | BRCA 2771 | PRAD 2772 | HNSC 2773 | READ 2774 | BRCA 2775 | KIRP 2776 | LIHC 2777 | KIRC 2778 | HNSC 2779 | BLCA 2780 | BRCA 2781 | PRAD 2782 | KIRC 2783 | BRCA 2784 | OV 2785 | PRAD 2786 | BRCA 2787 | COAD 2788 | BLCA 2789 | ESCA 2790 | BRCA 2791 | HNSC 2792 | PRAD 2793 | BRCA 2794 | UCEC 2795 | HNSC 2796 | STAD 2797 | KIRC 2798 | COAD 2799 | KIRC 2800 | OV 2801 | ESCA 2802 | READ 2803 | BRCA 2804 | BLCA 2805 | PRAD 2806 | BRCA 2807 | COAD 2808 | HNSC 2809 | KIRC 2810 | LUAD 2811 | ESCA 2812 | BRCA 2813 | LIHC 2814 | BRCA 2815 | BRCA 2816 | OV 2817 | KIRC 2818 | KIRC 2819 | GBM 2820 | LUSC 2821 | OV 2822 | OV 2823 | LUAD 2824 | ESCA 2825 | KIRC 2826 | BRCA 2827 | BLCA 2828 | GBM 2829 | KIRC 2830 | BLCA 2831 | OV 2832 | BRCA 2833 | COAD 2834 | KIRC 2835 | BRCA 2836 | OV 2837 | OV 2838 | STAD 2839 | PRAD 2840 | LUAD 2841 | HNSC 2842 | KIRC 2843 | HNSC 2844 | HNSC 2845 | KIRC 2846 | UCEC 2847 | COAD 2848 | BRCA 2849 | LUAD 2850 | UCEC 2851 | OV 2852 | KIRP 2853 | PRAD 2854 | PRAD 2855 | HNSC 2856 | LUAD 2857 | BRCA 2858 | BRCA 2859 | PRAD 2860 | GBM 2861 | HNSC 2862 | LIHC 2863 | BRCA 2864 | KIRC 2865 | COAD 2866 | BRCA 2867 | PRAD 2868 | BRCA 2869 | BLCA 2870 | COAD 2871 | PRAD 2872 | READ 2873 | LIHC 2874 | KIRC 2875 | LUAD 2876 | LIHC 2877 | KIRC 2878 | KIRC 2879 | LUAD 2880 | PRAD 2881 | ESCA 2882 | LUAD 2883 | BRCA 2884 | PRAD 2885 | OV 2886 | HNSC 2887 | BRCA 2888 | LIHC 2889 | BLCA 2890 | BRCA 2891 | BLCA 2892 | KIRC 2893 | READ 2894 | BRCA 2895 | BRCA 2896 | GBM 2897 | GBM 2898 | BLCA 2899 | OV 2900 | BRCA 2901 | HNSC 2902 | KIRC 2903 | BRCA 2904 | COAD 2905 | PRAD 2906 | BRCA 2907 | BRCA 2908 | OV 2909 | KIRP 2910 | BLCA 2911 | LIHC 2912 | HNSC 2913 | BLCA 2914 | HNSC 2915 | BRCA 2916 | BRCA 2917 | BRCA 2918 | PRAD 2919 | BRCA 2920 | KIRC 2921 | ESCA 2922 | BRCA 2923 | BRCA 2924 | BRCA 2925 | LUAD 2926 | PRAD 2927 | BRCA 2928 | PRAD 2929 | PRAD 2930 | HNSC 2931 | BRCA 2932 | PRAD 2933 | LUAD 2934 | BRCA 2935 | READ 2936 | OV 2937 | HNSC 2938 | PRAD 2939 | GBM 2940 | COAD 2941 | OV 2942 | OV 2943 | HNSC 2944 | LUAD 2945 | ESCA 2946 | OV 2947 | LUAD 2948 | GBM 2949 | UCEC 2950 | UCEC 2951 | PRAD 2952 | KIRC 2953 | BRCA 2954 | BRCA 2955 | BLCA 2956 | PRAD 2957 | BLCA 2958 | UCEC 2959 | HNSC 2960 | OV 2961 | BRCA 2962 | OV 2963 | KIRC 2964 | LUAD 2965 | UCEC 2966 | LUSC 2967 | HNSC 2968 | HNSC 2969 | BLCA 2970 | GBM 2971 | KIRC 2972 | UCEC 2973 | LIHC 2974 | BRCA 2975 | LIHC 2976 | GBM 2977 | PRAD 2978 | BRCA 2979 | LUAD 2980 | STAD 2981 | HNSC 2982 | OV 2983 | UCEC 2984 | HNSC 2985 | KIRC 2986 | OV 2987 | HNSC 2988 | PRAD 2989 | BRCA 2990 | HNSC 2991 | HNSC 2992 | BLCA 2993 | OV 2994 | COAD 2995 | BRCA 2996 | KIRC 2997 | PRAD 2998 | OV 2999 | KIRC 3000 | LUAD 3001 | GBM 3002 | LIHC 3003 | BRCA 3004 | BLCA 3005 | LUAD 3006 | KIRP 3007 | LUAD 3008 | LUAD 3009 | PRAD 3010 | KIRC 3011 | ESCA 3012 | HNSC 3013 | PRAD 3014 | HNSC 3015 | BRCA 3016 | LIHC 3017 | ESCA 3018 | COAD 3019 | KIRP 3020 | ESCA 3021 | COAD 3022 | ESCA 3023 | GBM 3024 | HNSC 3025 | ESCA 3026 | ESCA 3027 | PRAD 3028 | PRAD 3029 | BRCA 3030 | GBM 3031 | LIHC 3032 | OV 3033 | UCEC 3034 | HNSC 3035 | OV 3036 | UCEC 3037 | OV 3038 | ESCA 3039 | LIHC 3040 | HNSC 3041 | LUAD 3042 | BRCA 3043 | BLCA 3044 | UCEC 3045 | COAD 3046 | BRCA 3047 | GBM 3048 | OV 3049 | LUSC 3050 | KIRP 3051 | UCEC 3052 | LUAD 3053 | LUAD 3054 | KIRC 3055 | PRAD 3056 | OV 3057 | BRCA 3058 | HNSC 3059 | BRCA 3060 | LUAD 3061 | LIHC 3062 | LUAD 3063 | PRAD 3064 | GBM 3065 | COAD 3066 | STAD 3067 | UCEC 3068 | PRAD 3069 | BRCA 3070 | HNSC 3071 | BRCA 3072 | HNSC 3073 | PRAD 3074 | OV 3075 | BRCA 3076 | KIRC 3077 | LIHC 3078 | BRCA 3079 | PRAD 3080 | BRCA 3081 | BRCA 3082 | KIRP 3083 | HNSC 3084 | UCEC 3085 | BRCA 3086 | OV 3087 | OV 3088 | KIRC 3089 | PRAD 3090 | BRCA 3091 | BRCA 3092 | KIRC 3093 | LUAD 3094 | OV 3095 | KIRP 3096 | KIRC 3097 | BRCA 3098 | PRAD 3099 | HNSC 3100 | LUAD 3101 | LIHC 3102 | UCEC 3103 | PRAD 3104 | BLCA 3105 | STAD 3106 | BRCA 3107 | HNSC 3108 | BRCA 3109 | OV 3110 | GBM 3111 | BRCA 3112 | HNSC 3113 | LIHC 3114 | KIRC 3115 | PRAD 3116 | ESCA 3117 | BRCA 3118 | BRCA 3119 | KIRC 3120 | BRCA 3121 | HNSC 3122 | ESCA 3123 | PRAD 3124 | GBM 3125 | HNSC 3126 | KIRC 3127 | UCEC 3128 | GBM 3129 | OV 3130 | KIRC 3131 | PRAD 3132 | HNSC 3133 | ESCA 3134 | UCEC 3135 | BRCA 3136 | BRCA 3137 | PRAD 3138 | HNSC 3139 | LIHC 3140 | BRCA 3141 | LUAD 3142 | BRCA 3143 | UCEC 3144 | BRCA 3145 | BLCA 3146 | ESCA 3147 | STAD 3148 | BRCA 3149 | BLCA 3150 | COAD 3151 | STAD 3152 | PRAD 3153 | BRCA 3154 | LUSC 3155 | GBM 3156 | KIRP 3157 | BRCA 3158 | PRAD 3159 | BRCA 3160 | BRCA 3161 | KIRC 3162 | HNSC 3163 | BLCA 3164 | COAD 3165 | BLCA 3166 | PRAD 3167 | STAD 3168 | LIHC 3169 | STAD 3170 | GBM 3171 | LUAD 3172 | HNSC 3173 | UCEC 3174 | BLCA 3175 | KIRC 3176 | HNSC 3177 | GBM 3178 | HNSC 3179 | KIRP 3180 | KIRP 3181 | LUAD 3182 | BRCA 3183 | LUSC 3184 | PRAD 3185 | KIRC 3186 | LIHC 3187 | READ 3188 | BLCA 3189 | UCEC 3190 | OV 3191 | PRAD 3192 | STAD 3193 | KIRC 3194 | PRAD 3195 | UCEC 3196 | LUSC 3197 | UCEC 3198 | BRCA 3199 | HNSC 3200 | COAD 3201 | BRCA 3202 | BRCA 3203 | HNSC 3204 | LUSC 3205 | STAD 3206 | PRAD 3207 | HNSC 3208 | OV 3209 | HNSC 3210 | READ 3211 | GBM 3212 | KIRP 3213 | UCEC 3214 | KIRC 3215 | OV 3216 | HNSC 3217 | PRAD 3218 | PRAD 3219 | LUAD 3220 | BLCA 3221 | LUSC 3222 | LUSC 3223 | BRCA 3224 | OV 3225 | COAD 3226 | OV 3227 | HNSC 3228 | LIHC 3229 | LUAD 3230 | STAD 3231 | HNSC 3232 | BRCA 3233 | KIRC 3234 | BRCA 3235 | LIHC 3236 | LUSC 3237 | KIRC 3238 | READ 3239 | BRCA 3240 | LUAD 3241 | UCEC 3242 | GBM 3243 | GBM 3244 | LUAD 3245 | PRAD 3246 | LUAD 3247 | KIRP 3248 | GBM 3249 | BLCA 3250 | COAD 3251 | BRCA 3252 | STAD 3253 | UCEC 3254 | BRCA 3255 | KIRC 3256 | GBM 3257 | HNSC 3258 | PRAD 3259 | KIRC 3260 | LIHC 3261 | BRCA 3262 | PRAD 3263 | UCEC 3264 | LUAD 3265 | KIRC 3266 | HNSC 3267 | KIRP 3268 | GBM 3269 | PRAD 3270 | LUAD 3271 | OV 3272 | LIHC 3273 | LUAD 3274 | KIRC 3275 | BRCA 3276 | PRAD 3277 | BRCA 3278 | KIRC 3279 | HNSC 3280 | LUAD 3281 | OV 3282 | BRCA 3283 | KIRC 3284 | BRCA 3285 | LUAD 3286 | PRAD 3287 | BRCA 3288 | KIRC 3289 | GBM 3290 | KIRP 3291 | OV 3292 | UCEC 3293 | BRCA 3294 | BRCA 3295 | BRCA 3296 | KIRC 3297 | BRCA 3298 | BRCA 3299 | COAD 3300 | BLCA 3301 | BRCA 3302 | BRCA 3303 | BRCA 3304 | STAD 3305 | KIRP 3306 | HNSC 3307 | LIHC 3308 | LUAD 3309 | LUAD 3310 | GBM 3311 | UCEC 3312 | OV 3313 | HNSC 3314 | LUAD 3315 | HNSC 3316 | BLCA 3317 | STAD 3318 | BRCA 3319 | PRAD 3320 | BLCA 3321 | LUSC 3322 | BRCA 3323 | ESCA 3324 | OV 3325 | KIRC 3326 | BRCA 3327 | COAD 3328 | LUAD 3329 | HNSC 3330 | ESCA 3331 | LUAD 3332 | ESCA 3333 | GBM 3334 | LUAD 3335 | LIHC 3336 | LUSC 3337 | KIRP 3338 | PRAD 3339 | KIRC 3340 | LIHC 3341 | BRCA 3342 | HNSC 3343 | STAD 3344 | BLCA 3345 | UCEC 3346 | GBM 3347 | LUAD 3348 | LUSC 3349 | COAD 3350 | BRCA 3351 | LIHC 3352 | GBM 3353 | BRCA 3354 | BLCA 3355 | BRCA 3356 | KIRC 3357 | KIRC 3358 | LUAD 3359 | KIRP 3360 | OV 3361 | LUSC 3362 | STAD 3363 | KIRC 3364 | UCEC 3365 | BRCA 3366 | OV 3367 | OV 3368 | HNSC 3369 | KIRC 3370 | GBM 3371 | HNSC 3372 | ESCA 3373 | LUSC 3374 | BRCA 3375 | HNSC 3376 | LUSC 3377 | GBM 3378 | KIRC 3379 | OV 3380 | LUAD 3381 | LUSC 3382 | PRAD 3383 | LUSC 3384 | KIRC 3385 | GBM 3386 | OV 3387 | COAD 3388 | PRAD 3389 | HNSC 3390 | STAD 3391 | OV 3392 | BRCA 3393 | BRCA 3394 | LIHC 3395 | UCEC 3396 | LUSC 3397 | HNSC 3398 | HNSC 3399 | OV 3400 | UCEC 3401 | HNSC 3402 | BLCA 3403 | STAD 3404 | GBM 3405 | BLCA 3406 | BRCA 3407 | BRCA 3408 | BRCA 3409 | LIHC 3410 | LUSC 3411 | BLCA 3412 | BRCA 3413 | HNSC 3414 | UCEC 3415 | HNSC 3416 | COAD 3417 | BRCA 3418 | HNSC 3419 | LUAD 3420 | LUSC 3421 | COAD 3422 | PRAD 3423 | HNSC 3424 | ESCA 3425 | GBM 3426 | BRCA 3427 | BRCA 3428 | ESCA 3429 | BRCA 3430 | LUAD 3431 | LIHC 3432 | BRCA 3433 | LUAD 3434 | UCEC 3435 | LIHC 3436 | OV 3437 | BLCA 3438 | KIRC 3439 | LUSC 3440 | COAD 3441 | STAD 3442 | UCEC 3443 | BRCA 3444 | PRAD 3445 | GBM 3446 | LUAD 3447 | LUAD 3448 | ESCA 3449 | BRCA 3450 | GBM 3451 | LUAD 3452 | OV 3453 | STAD 3454 | UCEC 3455 | BRCA 3456 | KIRC 3457 | HNSC 3458 | OV 3459 | GBM 3460 | PRAD 3461 | BRCA 3462 | STAD 3463 | KIRP 3464 | BRCA 3465 | KIRP 3466 | LUAD 3467 | PRAD 3468 | BLCA 3469 | OV 3470 | KIRC 3471 | UCEC 3472 | HNSC 3473 | LUAD 3474 | BRCA 3475 | BRCA 3476 | READ 3477 | ESCA 3478 | COAD 3479 | KIRC 3480 | OV 3481 | HNSC 3482 | BRCA 3483 | LUAD 3484 | BRCA 3485 | LUAD 3486 | BRCA 3487 | HNSC 3488 | PRAD 3489 | KIRC 3490 | KIRC 3491 | UCEC 3492 | HNSC 3493 | KIRP 3494 | COAD 3495 | LUAD 3496 | PRAD 3497 | BRCA 3498 | STAD 3499 | PRAD 3500 | ESCA 3501 | KIRC 3502 | LUAD 3503 | GBM 3504 | STAD 3505 | STAD 3506 | OV 3507 | UCEC 3508 | COAD 3509 | PRAD 3510 | PRAD 3511 | PRAD 3512 | LUAD 3513 | LIHC 3514 | PRAD 3515 | COAD 3516 | BRCA 3517 | HNSC 3518 | KIRC 3519 | LUAD 3520 | OV 3521 | BRCA 3522 | BRCA 3523 | PRAD 3524 | LIHC 3525 | ESCA 3526 | LIHC 3527 | UCEC 3528 | BRCA 3529 | BLCA 3530 | UCEC 3531 | KIRC 3532 | BRCA 3533 | LUAD 3534 | PRAD 3535 | KIRP 3536 | BRCA 3537 | COAD 3538 | KIRP 3539 | BRCA 3540 | LUAD 3541 | KIRC 3542 | READ 3543 | UCEC 3544 | LUAD 3545 | COAD 3546 | BRCA 3547 | KIRP 3548 | PRAD 3549 | LIHC 3550 | LUAD 3551 | GBM 3552 | BLCA 3553 | UCEC 3554 | KIRC 3555 | BRCA 3556 | LIHC 3557 | BLCA 3558 | BRCA 3559 | HNSC 3560 | HNSC 3561 | LIHC 3562 | HNSC 3563 | BLCA 3564 | UCEC 3565 | PRAD 3566 | COAD 3567 | STAD 3568 | KIRC 3569 | LIHC 3570 | KIRP 3571 | BRCA 3572 | BLCA 3573 | KIRC 3574 | BRCA 3575 | KIRC 3576 | LIHC 3577 | UCEC 3578 | LIHC 3579 | BRCA 3580 | PRAD 3581 | KIRC 3582 | STAD 3583 | COAD 3584 | PRAD 3585 | LUAD 3586 | UCEC 3587 | PRAD 3588 | LUSC 3589 | LUAD 3590 | KIRC 3591 | LIHC 3592 | PRAD 3593 | BLCA 3594 | BRCA 3595 | BRCA 3596 | KIRP 3597 | KIRC 3598 | LUAD 3599 | BRCA 3600 | BRCA 3601 | HNSC 3602 | OV 3603 | HNSC 3604 | BRCA 3605 | BRCA 3606 | BLCA 3607 | LIHC 3608 | BRCA 3609 | BRCA 3610 | BRCA 3611 | OV 3612 | KIRC 3613 | GBM 3614 | BRCA 3615 | HNSC 3616 | KIRC 3617 | OV 3618 | GBM 3619 | KIRP 3620 | BRCA 3621 | KIRP 3622 | BRCA 3623 | BLCA 3624 | UCEC 3625 | UCEC 3626 | BRCA 3627 | HNSC 3628 | UCEC 3629 | BRCA 3630 | BRCA 3631 | KIRC 3632 | KIRC 3633 | STAD 3634 | OV 3635 | BRCA 3636 | BRCA 3637 | LUSC 3638 | BRCA 3639 | PRAD 3640 | LIHC 3641 | ESCA 3642 | KIRC 3643 | LUSC 3644 | BRCA 3645 | LUAD 3646 | LUAD 3647 | LUAD 3648 | KIRC 3649 | BRCA 3650 | OV 3651 | LUSC 3652 | GBM 3653 | BRCA 3654 | LUAD 3655 | LIHC 3656 | LUAD 3657 | PRAD 3658 | UCEC 3659 | PRAD 3660 | UCEC 3661 | STAD 3662 | HNSC 3663 | LIHC 3664 | HNSC 3665 | HNSC 3666 | BRCA 3667 | LUAD 3668 | BRCA 3669 | PRAD 3670 | LUAD 3671 | LUAD 3672 | BRCA 3673 | UCEC 3674 | BRCA 3675 | UCEC 3676 | HNSC 3677 | HNSC 3678 | HNSC 3679 | BRCA 3680 | KIRC 3681 | PRAD 3682 | KIRC 3683 | COAD 3684 | COAD 3685 | COAD 3686 | PRAD 3687 | BRCA 3688 | OV 3689 | LIHC 3690 | OV 3691 | BRCA 3692 | GBM 3693 | OV 3694 | LUAD 3695 | LUAD 3696 | BRCA 3697 | KIRC 3698 | LUAD 3699 | PRAD 3700 | BRCA 3701 | COAD 3702 | BRCA 3703 | OV 3704 | HNSC 3705 | BRCA 3706 | HNSC 3707 | BRCA 3708 | BRCA 3709 | KIRC 3710 | BRCA 3711 | OV 3712 | OV 3713 | HNSC 3714 | PRAD 3715 | OV 3716 | HNSC 3717 | BRCA 3718 | OV 3719 | BRCA 3720 | PRAD 3721 | KIRC 3722 | LUAD 3723 | LIHC 3724 | OV 3725 | LUAD 3726 | OV 3727 | KIRC 3728 | KIRC 3729 | PRAD 3730 | OV 3731 | HNSC 3732 | BLCA 3733 | KIRC 3734 | BRCA 3735 | BRCA 3736 | KIRC 3737 | BLCA 3738 | HNSC 3739 | KIRC 3740 | KIRC 3741 | KIRP 3742 | LIHC 3743 | UCEC 3744 | STAD 3745 | BRCA 3746 | BRCA 3747 | KIRC 3748 | KIRC 3749 | UCEC 3750 | OV 3751 | BRCA 3752 | KIRC 3753 | PRAD 3754 | HNSC 3755 | HNSC 3756 | UCEC 3757 | BLCA 3758 | HNSC 3759 | PRAD 3760 | KIRC 3761 | LIHC 3762 | KIRC 3763 | KIRC 3764 | OV 3765 | ESCA 3766 | PRAD 3767 | STAD 3768 | LUSC 3769 | ESCA 3770 | GBM 3771 | GBM 3772 | LUAD 3773 | ESCA 3774 | BRCA 3775 | OV 3776 | COAD 3777 | GBM 3778 | ESCA 3779 | STAD 3780 | KIRC 3781 | BLCA 3782 | LUAD 3783 | BRCA 3784 | KIRP 3785 | GBM 3786 | UCEC 3787 | HNSC 3788 | OV 3789 | GBM 3790 | BRCA 3791 | BRCA 3792 | LUAD 3793 | HNSC 3794 | BRCA 3795 | KIRP 3796 | GBM 3797 | BRCA 3798 | COAD 3799 | BRCA 3800 | BRCA 3801 | STAD 3802 | KIRC 3803 | PRAD 3804 | LUAD 3805 | KIRC 3806 | UCEC 3807 | LIHC 3808 | COAD 3809 | BRCA 3810 | PRAD 3811 | KIRP 3812 | KIRP 3813 | ESCA 3814 | LIHC 3815 | STAD 3816 | LIHC 3817 | BLCA 3818 | OV 3819 | PRAD 3820 | PRAD 3821 | STAD 3822 | KIRP 3823 | LUAD 3824 | HNSC 3825 | PRAD 3826 | BLCA 3827 | BRCA 3828 | HNSC 3829 | BRCA 3830 | LUAD 3831 | KIRC 3832 | ESCA 3833 | HNSC 3834 | OV 3835 | BLCA 3836 | READ 3837 | UCEC 3838 | PRAD 3839 | LIHC 3840 | KIRC 3841 | LUSC 3842 | PRAD 3843 | LUAD 3844 | READ 3845 | KIRC 3846 | LUSC 3847 | BRCA 3848 | LUAD 3849 | LUAD 3850 | KIRP 3851 | HNSC 3852 | HNSC 3853 | BLCA 3854 | PRAD 3855 | OV 3856 | BRCA 3857 | UCEC 3858 | COAD 3859 | LUAD 3860 | BRCA 3861 | BLCA 3862 | OV 3863 | BRCA 3864 | ESCA 3865 | ESCA 3866 | LUAD 3867 | BRCA 3868 | BRCA 3869 | KIRP 3870 | HNSC 3871 | KIRC 3872 | KIRC 3873 | ESCA 3874 | BLCA 3875 | BRCA 3876 | UCEC 3877 | BRCA 3878 | LIHC 3879 | HNSC 3880 | ESCA 3881 | ESCA 3882 | HNSC 3883 | LIHC 3884 | KIRP 3885 | GBM 3886 | KIRC 3887 | GBM 3888 | HNSC 3889 | OV 3890 | BRCA 3891 | OV 3892 | STAD 3893 | GBM 3894 | BRCA 3895 | BRCA 3896 | LIHC 3897 | BLCA 3898 | BRCA 3899 | LUAD 3900 | PRAD 3901 | OV 3902 | READ 3903 | LUSC 3904 | BRCA 3905 | BRCA 3906 | LUAD 3907 | LUAD 3908 | BRCA 3909 | COAD 3910 | BRCA 3911 | BRCA 3912 | HNSC 3913 | KIRC 3914 | BRCA 3915 | BRCA 3916 | LIHC 3917 | LUSC 3918 | BRCA 3919 | LUAD 3920 | PRAD 3921 | BRCA 3922 | KIRC 3923 | GBM 3924 | KIRC 3925 | KIRC 3926 | KIRP 3927 | STAD 3928 | BRCA 3929 | LUAD 3930 | HNSC 3931 | LIHC 3932 | GBM 3933 | HNSC 3934 | HNSC 3935 | OV 3936 | LUSC 3937 | BRCA 3938 | UCEC 3939 | BLCA 3940 | ESCA 3941 | BRCA 3942 | BRCA 3943 | PRAD 3944 | KIRC 3945 | GBM 3946 | PRAD 3947 | BRCA 3948 | COAD 3949 | KIRC 3950 | KIRC 3951 | KIRP 3952 | BRCA 3953 | PRAD 3954 | STAD 3955 | BRCA 3956 | KIRC 3957 | BRCA 3958 | STAD 3959 | GBM 3960 | LUAD 3961 | LUAD 3962 | LUSC 3963 | LUAD 3964 | READ 3965 | LUSC 3966 | LUAD 3967 | PRAD 3968 | BRCA 3969 | OV 3970 | HNSC 3971 | KIRC 3972 | OV 3973 | HNSC 3974 | OV 3975 | LUSC 3976 | HNSC 3977 | BRCA 3978 | OV 3979 | KIRC 3980 | GBM 3981 | HNSC 3982 | BRCA 3983 | BLCA 3984 | HNSC 3985 | LUSC 3986 | STAD 3987 | BLCA 3988 | KIRP 3989 | KIRC 3990 | HNSC 3991 | KIRC 3992 | BRCA 3993 | STAD 3994 | GBM 3995 | BRCA 3996 | KIRP 3997 | KIRC 3998 | BRCA 3999 | PRAD 4000 | HNSC 4001 | UCEC 4002 | LUAD 4003 | BRCA 4004 | OV 4005 | KIRC 4006 | BRCA 4007 | READ 4008 | HNSC 4009 | HNSC 4010 | STAD 4011 | LIHC 4012 | PRAD 4013 | PRAD 4014 | GBM 4015 | HNSC 4016 | PRAD 4017 | KIRC 4018 | LUSC 4019 | BRCA 4020 | KIRC 4021 | GBM 4022 | COAD 4023 | PRAD 4024 | UCEC 4025 | BRCA 4026 | HNSC 4027 | HNSC 4028 | BRCA 4029 | PRAD 4030 | UCEC 4031 | PRAD 4032 | PRAD 4033 | BRCA 4034 | BLCA 4035 | LUSC 4036 | LIHC 4037 | UCEC 4038 | PRAD 4039 | KIRC 4040 | LIHC 4041 | ESCA 4042 | OV 4043 | BRCA 4044 | LUSC 4045 | BRCA 4046 | KIRC 4047 | READ 4048 | KIRC 4049 | LUAD 4050 | OV 4051 | BLCA 4052 | KIRC 4053 | LUAD 4054 | BRCA 4055 | BRCA 4056 | PRAD 4057 | OV 4058 | BLCA 4059 | BRCA 4060 | LUAD 4061 | LIHC 4062 | KIRC 4063 | BRCA 4064 | STAD 4065 | BRCA 4066 | ESCA 4067 | BLCA 4068 | UCEC 4069 | ESCA 4070 | LUAD 4071 | OV 4072 | HNSC 4073 | PRAD 4074 | LUAD 4075 | KIRP 4076 | COAD 4077 | COAD 4078 | HNSC 4079 | BRCA 4080 | OV 4081 | BLCA 4082 | OV 4083 | BRCA 4084 | PRAD 4085 | BLCA 4086 | GBM 4087 | BRCA 4088 | PRAD 4089 | HNSC 4090 | OV 4091 | KIRC 4092 | BRCA 4093 | BRCA 4094 | PRAD 4095 | KIRC 4096 | OV 4097 | PRAD 4098 | BLCA 4099 | LUSC 4100 | LUAD 4101 | BRCA 4102 | LUAD 4103 | LUAD 4104 | KIRC 4105 | LUAD 4106 | HNSC 4107 | LUAD 4108 | BRCA 4109 | OV 4110 | GBM 4111 | KIRC 4112 | READ 4113 | COAD 4114 | LUAD 4115 | BLCA 4116 | BLCA 4117 | STAD 4118 | GBM 4119 | KIRC 4120 | KIRC 4121 | PRAD 4122 | LUAD 4123 | KIRC 4124 | BLCA 4125 | BRCA 4126 | HNSC 4127 | LUAD 4128 | READ 4129 | COAD 4130 | READ 4131 | KIRC 4132 | STAD 4133 | HNSC 4134 | OV 4135 | KIRC 4136 | HNSC 4137 | BRCA 4138 | OV 4139 | BRCA 4140 | OV 4141 | BRCA 4142 | HNSC 4143 | HNSC 4144 | KIRC 4145 | PRAD 4146 | HNSC 4147 | UCEC 4148 | LUAD 4149 | HNSC 4150 | BRCA 4151 | PRAD 4152 | LIHC 4153 | OV 4154 | BRCA 4155 | BRCA 4156 | KIRC 4157 | OV 4158 | PRAD 4159 | HNSC 4160 | PRAD 4161 | STAD 4162 | BRCA 4163 | KIRC 4164 | ESCA 4165 | COAD 4166 | UCEC 4167 | HNSC 4168 | GBM 4169 | OV 4170 | COAD 4171 | STAD 4172 | BRCA 4173 | KIRC 4174 | KIRC 4175 | LUSC 4176 | STAD 4177 | HNSC 4178 | BRCA 4179 | KIRC 4180 | KIRC 4181 | PRAD 4182 | KIRC 4183 | BRCA 4184 | UCEC 4185 | BLCA 4186 | BLCA 4187 | BRCA 4188 | COAD 4189 | BRCA 4190 | PRAD 4191 | BRCA 4192 | UCEC 4193 | KIRP 4194 | OV 4195 | READ 4196 | KIRC 4197 | KIRC 4198 | HNSC 4199 | LUAD 4200 | GBM 4201 | KIRC 4202 | HNSC 4203 | STAD 4204 | BRCA 4205 | BRCA 4206 | BRCA 4207 | LIHC 4208 | PRAD 4209 | LIHC 4210 | UCEC 4211 | STAD 4212 | BRCA 4213 | LIHC 4214 | LUAD 4215 | BRCA 4216 | OV 4217 | STAD 4218 | BLCA 4219 | BRCA 4220 | UCEC 4221 | KIRC 4222 | BLCA 4223 | LUAD 4224 | COAD 4225 | OV 4226 | PRAD 4227 | BRCA 4228 | OV 4229 | LUSC 4230 | LUAD 4231 | KIRC 4232 | LIHC 4233 | BRCA 4234 | BRCA 4235 | HNSC 4236 | HNSC 4237 | COAD 4238 | HNSC 4239 | KIRP 4240 | KIRC 4241 | BRCA 4242 | PRAD 4243 | HNSC 4244 | ESCA 4245 | LUSC 4246 | BRCA 4247 | BRCA 4248 | KIRP 4249 | KIRP 4250 | BLCA 4251 | BRCA 4252 | UCEC 4253 | PRAD 4254 | LUSC 4255 | COAD 4256 | KIRC 4257 | OV 4258 | KIRC 4259 | STAD 4260 | KIRC 4261 | KIRC 4262 | COAD 4263 | KIRP 4264 | BRCA 4265 | KIRP 4266 | HNSC 4267 | GBM 4268 | BRCA 4269 | KIRC 4270 | LUSC 4271 | BRCA 4272 | BRCA 4273 | LUAD 4274 | KIRC 4275 | UCEC 4276 | LIHC 4277 | BLCA 4278 | KIRC 4279 | BRCA 4280 | ESCA 4281 | COAD 4282 | HNSC 4283 | BRCA 4284 | BRCA 4285 | UCEC 4286 | OV 4287 | BRCA 4288 | BRCA 4289 | KIRC 4290 | READ 4291 | KIRC 4292 | LUAD 4293 | HNSC 4294 | HNSC 4295 | ESCA 4296 | BRCA 4297 | PRAD 4298 | PRAD 4299 | KIRC 4300 | BRCA 4301 | LUAD 4302 | OV 4303 | PRAD 4304 | HNSC 4305 | ESCA 4306 | STAD 4307 | ESCA 4308 | KIRC 4309 | OV 4310 | LIHC 4311 | COAD 4312 | PRAD 4313 | GBM 4314 | KIRC 4315 | LUAD 4316 | HNSC 4317 | OV 4318 | READ 4319 | HNSC 4320 | BRCA 4321 | BRCA 4322 | LUAD 4323 | BRCA 4324 | KIRC 4325 | HNSC 4326 | BLCA 4327 | HNSC 4328 | STAD 4329 | LUAD 4330 | KIRC 4331 | BRCA 4332 | KIRC 4333 | LUSC 4334 | UCEC 4335 | BRCA 4336 | KIRP 4337 | HNSC 4338 | BRCA 4339 | KIRC 4340 | READ 4341 | LUSC 4342 | KIRC 4343 | HNSC 4344 | LUAD 4345 | BRCA 4346 | LIHC 4347 | OV 4348 | HNSC 4349 | GBM 4350 | LIHC 4351 | LUAD 4352 | BRCA 4353 | BRCA 4354 | PRAD 4355 | PRAD 4356 | LUSC 4357 | BRCA 4358 | HNSC 4359 | HNSC 4360 | KIRC 4361 | HNSC 4362 | KIRC 4363 | LUAD 4364 | BRCA 4365 | BRCA 4366 | BLCA 4367 | UCEC 4368 | STAD 4369 | BRCA 4370 | HNSC 4371 | LUAD 4372 | ESCA 4373 | STAD 4374 | BRCA 4375 | OV 4376 | STAD 4377 | OV 4378 | UCEC 4379 | STAD 4380 | BRCA 4381 | LUAD 4382 | LUAD 4383 | BLCA 4384 | READ 4385 | LIHC 4386 | HNSC 4387 | COAD 4388 | BRCA 4389 | HNSC 4390 | BRCA 4391 | UCEC 4392 | LUAD 4393 | BRCA 4394 | BRCA 4395 | GBM 4396 | LUAD 4397 | PRAD 4398 | OV 4399 | BRCA 4400 | STAD 4401 | HNSC 4402 | PRAD 4403 | PRAD 4404 | KIRC 4405 | ESCA 4406 | LUAD 4407 | KIRC 4408 | BRCA 4409 | STAD 4410 | BRCA 4411 | BRCA 4412 | KIRC 4413 | KIRP 4414 | LUAD 4415 | GBM 4416 | LUAD 4417 | PRAD 4418 | BRCA 4419 | STAD 4420 | BLCA 4421 | UCEC 4422 | UCEC 4423 | LIHC 4424 | COAD 4425 | KIRC 4426 | COAD 4427 | KIRC 4428 | GBM 4429 | ESCA 4430 | PRAD 4431 | COAD 4432 | ESCA 4433 | BRCA 4434 | BLCA 4435 | BRCA 4436 | KIRP 4437 | BRCA 4438 | UCEC 4439 | HNSC 4440 | BRCA 4441 | PRAD 4442 | BLCA 4443 | HNSC 4444 | LUAD 4445 | LUAD 4446 | BRCA 4447 | BRCA 4448 | BRCA 4449 | LUAD 4450 | GBM 4451 | KIRC 4452 | LUAD 4453 | BRCA 4454 | BLCA 4455 | HNSC 4456 | BRCA 4457 | KIRP 4458 | OV 4459 | PRAD 4460 | READ 4461 | KIRC 4462 | LUSC 4463 | BRCA 4464 | PRAD 4465 | BRCA 4466 | PRAD 4467 | BLCA 4468 | BRCA 4469 | -------------------------------------------------------------------------------- /prepare_dataset.py: -------------------------------------------------------------------------------- 1 | """ Prepare the dataset.pkl from more widely used data format. 2 | 3 | """ 4 | 5 | import os 6 | import argparse 7 | import numpy as np 8 | import pickle 9 | import pandas as pd 10 | 11 | __author__ = "Yifeng Tao" 12 | 13 | 14 | parser = argparse.ArgumentParser() 15 | parser.add_argument("--input_dir", help="directory of input files", type=str, default="mydata") 16 | parser.add_argument("--output_dir", help="directory of output files", type=str, default="mydata") 17 | args = parser.parse_args() 18 | 19 | 20 | # input file 1: DEG matrix in csv format 21 | df = pd.read_csv(os.path.join(args.input_dir, "deg.csv"), index_col=0) 22 | 23 | tmr = list(df.index) 24 | 25 | idx2deg = {i:d for i, d in enumerate(df.columns)} 26 | 27 | deg = df.values 28 | 29 | 30 | # input file 2: cancer type list in txt format 31 | can = [] 32 | with open(os.path.join(args.input_dir,"cancer_type.txt"), "r") as f: 33 | for line in f: 34 | can.append(line.strip()) 35 | 36 | idx2can = {i:c for i, c in enumerate(list(set(can)))} 37 | can2idx = {idx2can[i]:i for i in idx2can.keys()} 38 | 39 | can = [can2idx[c] for c in can] 40 | 41 | can = np.array(can) 42 | 43 | 44 | # input file 3: SGA list in txt format 45 | sga = [] 46 | with open(os.path.join(args.input_dir, "sga.txt"), "r") as f: 47 | next(f) 48 | for line in f: 49 | line = line.strip().split(", ") 50 | sga.append(line[1:]) 51 | 52 | idx2sga = {i:s for i,s in enumerate(list(set([l for line in sga for l in line])))} 53 | sga2idx = {idx2sga[i]:i for i in idx2sga.keys()} 54 | 55 | sga = [[sga2idx[l] for l in line] for line in sga] 56 | 57 | 58 | # output file: dictionary in pickle format 59 | dataset = { 60 | "can": can, 61 | "deg": deg, 62 | "idx2can": idx2can, 63 | "idx2deg": idx2deg, 64 | "idx2sga": idx2sga, 65 | "sga": sga, 66 | "tmr": tmr} 67 | 68 | with open(os.path.join(args.output_dir, "dataset.pkl"), "wb") as f: 69 | pickle.dump(dataset, f, protocol=2) 70 | 71 | -------------------------------------------------------------------------------- /test_run.py: -------------------------------------------------------------------------------- 1 | """ Demo of training and evaluating GIT model and its variants. 2 | 3 | """ 4 | import os 5 | import argparse 6 | 7 | from utils import bool_ext, load_dataset, split_dataset, evaluate 8 | from models import GIT 9 | 10 | __author__ = "Yifeng Tao" 11 | 12 | 13 | parser = argparse.ArgumentParser() 14 | 15 | parser.add_argument("--train_model", help="whether to train model or load model", type=bool_ext, default=True) 16 | 17 | parser.add_argument("--input_dir", help="directory of input files", type=str, default="data") 18 | parser.add_argument("--output_dir", help="directory of output files", type=str, default="data") 19 | 20 | parser.add_argument("--embedding_size", help="embedding dimension of genes and tumors", type=int, default=512) 21 | parser.add_argument("--hidden_size", help="hidden layer dimension of MLP decoder", type=int, default=1024) 22 | parser.add_argument("--attention_size", help="size of attention parameter beta_j", type=int, default=400) 23 | parser.add_argument("--attention_head", help="number of attention heads", type=int, default=128) 24 | 25 | parser.add_argument("--learning_rate", help="learning rate for Adam", type=float, default=1e-4) 26 | parser.add_argument("--max_iter", help="maximum number of training iterations", type=int, default=3072*20) 27 | parser.add_argument("--max_fscore", help="Max F1 score to early stop model from training", type=float, default=0.7) 28 | parser.add_argument("--batch_size", help="training batch size", type=int, default=16) 29 | parser.add_argument("--test_batch_size", help="test batch size", type=int, default=512) 30 | parser.add_argument("--test_inc_size", help="increment interval size between log outputs", type=int, default=256) 31 | parser.add_argument("--dropout_rate", help="dropout rate", type=float, default=0.5) 32 | parser.add_argument("--weight_decay", help="coefficient of l2 regularizer", type=float, default=1e-5) 33 | 34 | # GIT variants: 35 | # args.initializtion=False -> GIT-init 36 | parser.add_argument("--initializtion", help="whether to use pre-trained gene embeddings or not", type=bool_ext, default=True) 37 | # args.attention=False -> GIT-attn 38 | parser.add_argument("--attention", help="whether to use attention mechanism or not", type=bool_ext, default=True) 39 | # args.cancer_type=False -> GIT-can 40 | parser.add_argument("--cancer_type", help="whether to use cancer type or not", type=bool_ext, default=True) 41 | # args.deg_shuffle=False -> DEG-shuffled 42 | parser.add_argument("--deg_shuffle", help="whether to shuffle DEGs or not", type=bool_ext, default=False) 43 | 44 | args = parser.parse_args() 45 | 46 | # Above default hyper parameters tuned in full GIT model; 47 | # Ablated GIT variants have slight different tuned hyper parameters: 48 | 49 | #if args.cancer_type == False: 50 | # args.max_iter = 3072*40 51 | #elif args.attention == False: 52 | # args.max_iter = 3072*40 53 | # args.learning_rate = 0.0003 54 | 55 | print("Loading dataset...") 56 | dataset = load_dataset(input_dir=args.input_dir, deg_shuffle=args.deg_shuffle) 57 | train_set, test_set = split_dataset(dataset, ratio=0.66) 58 | 59 | args.can_size = dataset["can"].max() # cancer type dimension 60 | args.sga_size = dataset["sga"].max() # SGA dimension 61 | args.deg_size = dataset["deg"].shape[1] # DEG output dimension 62 | args.num_max_sga = dataset["sga"].shape[1] # maximum number of SGAs in a tumor 63 | 64 | print("Hyperparameters:") 65 | print(args) 66 | 67 | model = GIT(args) # initialize GIT model 68 | model.build() # build GIT model 69 | 70 | if args.train_model: # train from scratch 71 | print("Training...") 72 | model.train(train_set, test_set, 73 | batch_size=args.batch_size, 74 | test_batch_size=args.test_batch_size, 75 | max_iter=args.max_iter, 76 | max_fscore=args.max_fscore, 77 | test_inc_size=args.test_inc_size) 78 | model.load_model(os.path.join(args.output_dir, "trained_model.pth")) 79 | else: # or directly load trained model 80 | model.load_model(os.path.join(args.input_dir, "trained_model.pth")) 81 | 82 | # evaluation 83 | print("Evaluating...") 84 | labels, preds, _, _, _, _, _ = model.test(test_set, test_batch_size=args.test_batch_size) 85 | precision, recall, f1score, accuracy = evaluate(labels, preds, epsilon=model.epsilon) 86 | print("prec=%.3f, recall=%.3f, F1=%.3f, acc=%.3f"%(precision, recall, f1score, accuracy)) 87 | # prec=0.702, recall=0.565, F1=0.626, acc=0.788 88 | 89 | -------------------------------------------------------------------------------- /utils.py: -------------------------------------------------------------------------------- 1 | """ Shared utilities for models.py and test_run.py. 2 | 3 | """ 4 | import os 5 | import random 6 | import numpy as np 7 | import pickle 8 | 9 | import torch 10 | from torch.autograd import Variable 11 | 12 | __author__ = "Yifeng Tao" 13 | 14 | 15 | def bool_ext(rbool): 16 | """ Solve the problem that raw bool type is always True. 17 | 18 | Parameters 19 | ---------- 20 | rbool: str 21 | should be True of False. 22 | 23 | """ 24 | 25 | if rbool not in ["True", "False"]: 26 | raise ValueError("Not a valid boolean string") 27 | 28 | return rbool == "True" 29 | 30 | 31 | def load_dataset(input_dir="data", deg_shuffle=False): 32 | """ Load dataset, modify, and shuffle`. 33 | 34 | Parameters 35 | ---------- 36 | input_dir: str 37 | input directory of dataset 38 | deg_shuffle: bool 39 | whether to shuffle DEG or not 40 | 41 | Returns 42 | ------- 43 | dataset: dict 44 | dict of lists, including SGAs, cancer types, DEGs, patient barcodes 45 | """ 46 | 47 | # load dataset 48 | data = pickle.load( open(os.path.join(input_dir, "dataset.pkl"), "rb") ) 49 | can_r = data["can"] # cancer type index of tumors: list of int 50 | sga_r = data["sga"] # SGA index of tumors: list of list 51 | deg = data["deg"] # DEG binary matrix of tumors: 2D array of 0/1 52 | tmr = data["tmr"] # barcodes of tumors: list of str 53 | 54 | # shift the index of cancer type by +1, 0 is for padding 55 | can = np.asarray([[x+1] for x in can_r], dtype=int) 56 | 57 | # shift the index of SGAs by +1, 0 is for padding 58 | num_max_sga = max([len(s) for s in sga_r]) 59 | sga = np.zeros( (len(sga_r), num_max_sga), dtype=int ) 60 | for idx, line in enumerate(sga_r): 61 | line = [s+1 for s in line] 62 | sga[idx,0:len(line)] = line 63 | 64 | # shuffle DEGs 65 | if deg_shuffle: 66 | rng = list(range(deg.shape[1])) 67 | for idx in range(deg.shape[0]): 68 | random.shuffle(rng) 69 | deg[idx] = deg[idx][rng] 70 | 71 | # shuffle whole dataset 72 | rng = list(range(len(can))) 73 | random.Random(2019).shuffle(rng) 74 | can = can[rng] 75 | sga = sga[rng] 76 | deg = deg[rng] 77 | tmr = [tmr[idx] for idx in rng] 78 | 79 | dataset = {"can":can, "sga":sga, "deg":deg, "tmr":tmr} 80 | 81 | return dataset 82 | 83 | 84 | def split_dataset(dataset, ratio=0.66): 85 | """ Split the dataset according to the ratio of training/test sets. 86 | 87 | Parameters 88 | ---------- 89 | dataset: dict 90 | dict of lists, including SGAs, cancer types, DEGs, patient barcodes 91 | ratio: float 92 | size(train_set)/size(train_set+test_set) 93 | 94 | Returns 95 | ------- 96 | train_set, test_set: dict 97 | 98 | """ 99 | 100 | num_sample = len(dataset["can"]) 101 | num_train_sample = int(num_sample*ratio) 102 | 103 | train_set = {"sga":dataset["sga"][:num_train_sample], 104 | "can":dataset["can"][:num_train_sample], 105 | "deg":dataset["deg"][:num_train_sample], 106 | "tmr":dataset["tmr"][:num_train_sample]} 107 | test_set = {"sga":dataset["sga"][num_train_sample:], 108 | "can":dataset["can"][num_train_sample:], 109 | "deg":dataset["deg"][num_train_sample:], 110 | "tmr":dataset["tmr"][num_train_sample:]} 111 | 112 | return train_set, test_set 113 | 114 | 115 | def wrap_dataset(sga, can, deg, tmr): 116 | """ Wrap default numpy or list data into PyTorch variables. 117 | 118 | """ 119 | 120 | dataset = {"sga": Variable(torch.LongTensor(sga)), 121 | "can": Variable(torch.LongTensor(can)), 122 | "deg": Variable(torch.FloatTensor(deg)), 123 | "tmr": tmr} 124 | 125 | return dataset 126 | 127 | 128 | def get_minibatch(dataset, index, batch_size, batch_type="train"): 129 | """ Get a mini-batch dataset for training or test. 130 | 131 | Parameters 132 | ---------- 133 | dataset: dict 134 | dict of lists, including SGAs, cancer types, DEGs, patient barcodes 135 | index: int 136 | starting index of current mini-batch 137 | batch_size: int 138 | batch_type: str 139 | batch strategy is slightly different for training and test 140 | "train": will return to beginning of the queue when `index` out of range 141 | "test": will not return to beginning of the queue when `index` out of range 142 | 143 | Returns 144 | ------- 145 | batch_dataset: dict 146 | a mini-batch of the input `dataset`. 147 | 148 | """ 149 | 150 | sga = dataset["sga"] 151 | can = dataset["can"] 152 | deg = dataset["deg"] 153 | tmr = dataset["tmr"] 154 | 155 | if batch_type == "train": 156 | batch_sga = [ sga[idx%len(sga)] 157 | for idx in range(index,index+batch_size) ] 158 | batch_can = [ can[idx%len(can)] 159 | for idx in range(index,index+batch_size) ] 160 | batch_deg = [ deg[idx%len(deg)] 161 | for idx in range(index,index+batch_size) ] 162 | batch_tmr = [ tmr[idx%len(tmr)] 163 | for idx in range(index,index+batch_size) ] 164 | elif batch_type == "test": 165 | batch_sga = sga[index:index+batch_size] 166 | batch_can = can[index:index+batch_size] 167 | batch_deg = deg[index:index+batch_size] 168 | batch_tmr = tmr[index:index+batch_size] 169 | 170 | batch_dataset = wrap_dataset( 171 | batch_sga, 172 | batch_can, 173 | batch_deg, 174 | batch_tmr) 175 | 176 | return batch_dataset 177 | 178 | 179 | def evaluate(labels, preds, epsilon=1e-4): 180 | """ Calculate performance metrics given ground truths and prediction results. 181 | 182 | Parameters 183 | ---------- 184 | labels: matrix of 0/1 185 | ground truth labels 186 | preds: matrix of float in [0,1] 187 | predicted labels 188 | epsilon: float 189 | a small Laplacian smoothing term to avoid zero denominator 190 | 191 | Returns 192 | ------- 193 | precision: float 194 | recall: float 195 | f1score: float 196 | accuracy: float 197 | 198 | """ 199 | 200 | flat_labels = np.reshape(labels,-1) 201 | flat_preds = np.reshape(np.around(preds),-1) 202 | 203 | accuracy = np.mean(flat_labels == flat_preds) 204 | true_pos = np.dot(flat_labels, flat_preds) 205 | precision = 1.0*true_pos/flat_preds.sum() 206 | recall = 1.0*true_pos/flat_labels.sum() 207 | 208 | f1score = 2*precision*recall/(precision+recall+epsilon) 209 | 210 | return precision, recall, f1score, accuracy 211 | 212 | --------------------------------------------------------------------------------