Dataset Viewer
Auto-converted to Parquet Duplicate
sample_id
stringlengths
15
15
expression_subtype
stringclasses
4 values
TCGA-39-5031-01
basal
TCGA-39-5024-01
basal
TCGA-18-3419-01
basal
TCGA-66-2781-01
basal
TCGA-33-4538-01
classical
TCGA-66-2778-01
basal
TCGA-56-6546-01
secretory
TCGA-22-4599-01
primitive
TCGA-66-2785-01
secretory
TCGA-46-3765-01
classical
TCGA-60-2711-01
classical
TCGA-33-4533-01
primitive
TCGA-43-6647-01
secretory
TCGA-22-5485-01
classical
TCGA-66-2765-01
basal
TCGA-60-2720-01
basal
TCGA-21-5786-01
basal
TCGA-18-3407-01
classical
TCGA-22-4613-01
classical
TCGA-66-2768-01
basal
TCGA-66-2800-01
classical
TCGA-33-4586-01
classical
TCGA-60-2710-01
classical
TCGA-46-3768-01
primitive
TCGA-21-1077-01
classical
TCGA-60-2698-01
secretory
TCGA-22-5473-01
classical
TCGA-37-3789-01
basal
TCGA-22-5489-01
secretory
TCGA-66-2759-01
classical
TCGA-66-2771-01
secretory
TCGA-66-2755-01
secretory
TCGA-66-2773-01
basal
TCGA-63-6202-01
secretory
TCGA-18-3411-01
classical
TCGA-63-5131-01
primitive
TCGA-18-4721-01
classical
TCGA-66-2757-01
primitive
TCGA-60-2725-01
secretory
TCGA-51-4080-01
classical
TCGA-66-2782-01
basal
TCGA-43-6770-01
classical
TCGA-22-1011-01
basal
TCGA-37-3783-01
basal
TCGA-34-5232-01
basal
TCGA-22-5471-01
classical
TCGA-43-2578-01
primitive
TCGA-66-2777-01
secretory
TCGA-18-3421-01
basal
TCGA-66-2789-01
secretory
TCGA-39-5037-01
classical
TCGA-22-4604-01
classical
TCGA-18-4083-01
classical
TCGA-34-5928-01
secretory
TCGA-60-2708-01
basal
TCGA-21-1076-01
secretory
TCGA-34-5236-01
classical
TCGA-39-5019-01
classical
TCGA-39-5036-01
classical
TCGA-33-4547-01
basal
TCGA-43-6143-01
primitive
TCGA-33-4582-01
classical
TCGA-22-0944-01
classical
TCGA-70-6723-01
classical
TCGA-34-5241-01
basal
TCGA-60-2719-01
primitive
TCGA-46-6026-01
primitive
TCGA-39-5028-01
basal
TCGA-60-2724-01
basal
TCGA-85-6560-01
primitive
TCGA-39-5016-01
classical
TCGA-18-3409-01
secretory
TCGA-66-2783-01
classical
TCGA-43-6771-01
basal
TCGA-37-4141-01
primitive
TCGA-22-1012-01
classical
TCGA-18-3412-01
classical
TCGA-21-1071-01
classical
TCGA-39-5021-01
primitive
TCGA-39-5035-01
secretory
TCGA-34-5239-01
secretory
TCGA-21-5787-01
secretory
TCGA-66-2727-01
classical
TCGA-66-2754-01
primitive
TCGA-18-3406-01
basal
TCGA-56-5897-01
basal
TCGA-60-2713-01
primitive
TCGA-56-5898-01
classical
TCGA-66-2793-01
primitive
TCGA-85-6175-01
secretory
TCGA-34-2608-01
secretory
TCGA-34-5929-01
primitive
TCGA-66-2770-01
classical
TCGA-39-5022-01
secretory
TCGA-66-2788-01
basal
TCGA-18-3416-01
classical
TCGA-22-5472-01
basal
TCGA-46-3769-01
secretory
TCGA-66-2794-01
classical
TCGA-70-6722-01
secretory
End of preview. Expand in Data Studio

This dataset combines expression data from TCGA with protein interaction network from STRING. An example implementation of a GCN trained on this dataset is available here.

How

I built node indices from filtered human ppi in notebooks/process_human_ppi.ipynb to make a custom PyG Dataset. human ppi -> filter for high confidence (score > 0.7) -> filter to keep only genes in expression dataset -> gene ids -> integer ids -> edge indices GCN expects a graph dataset; each graph represents a patient sample. Nodes represent genes, and edges represent protein-protein interactions between genes. The structure of the graph is the same for all samples (given by edge indices), only the node features (expression levels for each gene) vary.

Data

  • expression: TCGA LUSC HiSeqV2 gene expression data (178 samples, 20530 genes), UCSC Xena
  • subtype labels: TCGA Nature 2012 paper supplementary data
  • STRING PPI network: Homo sapiens interaction data; high-confidence interactions (score > 0.7)
  • class distribution:
classical: 65 (36.5%)
basal: 43 (24.2%)
secretory: 43 (24.2%)
primitive: 27 (15.2%)
Downloads last month
46