Table 1.

Data collected from multiple platforms

Dataset	Data type	Platform	Samples
GDSC	Gene expression	Affymetrix Human Genome U219 array	49 cell lines ×  14,770 genes
	CNV	Affymetrix SNP6 array	49 cell lines ×  3037 genes
	Mutation	Whole exome sequencing	49 cell lines ×  8849 genes
	Drug response	IC50	49 cell lines ×  220 drugs
CCLE	Gene expression	Illumina Hiseq 2000	28 cell lines ×  14,770 genes
	Drug response	IC50	28 cell lines × 13 drugs
TCGA	Gene expression	Illumina Hiseq 2000	1100 tumors × 14,770 genes
	Drug response	RECIST response categories a	110 tumors × 5 drugs
Drugs	Chemical structure	rcdk b	220 drugs × 1024 fingerprints
	Target	Curated	220 drugs × 272 targets
MSigDB	Canonical pathways	Curated	1329 pathway gene sets

^a Response Evaluation Criteria in Solid Tumours (RECIST), a standard way to categorize treatment response of a cancer patient, including complete response, a partial response, progressive disease, and stable disease

^b An R package which can take the SMILES string of a drug as input and output the fingerprints, 1D- and 2D-structres of the drug