Benchmark Data for Synergy

The benchmark data for Synergy was created using this pipeline.

Number of cell lines, drugs, and experiments in the benchmark datasets.

Dataset	Number of cell lines	Number of drugs	Number of experiments	Number of unique experiments
ALMANAC	57	103	296,166	294,152
FRIEDMAN	24	108	129,921	129,921
ONEIL	35	38	82,712	20,405
ASTRAZENECA	97	116	15,365	15,233
NCATS-1	10	1910	9123	7401
FORCINA	1	1818	1818	1818
FLOBAK	8	19	9984	1392
SMALL	6	34	264	208

Note: The ‘SMALL’ dataset is a collection of smaller experiments and should be used with caution.

Data Location

• ADD FTP

Data File Description

• y_data folder includes synergy measurements

  • synergy.tsv: file containing observations. Columns are as follows:

    • DepMapID: ID for cell line.

    • DrugID_row: ID for the first drug of the combination.

    • DrugID_col: ID for the second drug of the combination.

    • study: Study name identifying where the experiment originates from.

    • loewe: Loewe synergy value.

    • bliss: Bliss synergy value.

    • zip: ZIP synergy value.

    • hsa: HSA synergy value.

    • smean: Smean synergy value.

    • css: CSS synergy value.

• splits folder includes sample indices corresponding to rows in synergy.tsv in y_data folder. Sample index starts with 0.

• x_data folder includes multi-omics data of cancer cell lines and multi-modal data of drugs. The index of the files is the drug or cell ID and the columns are the features.

  • cell_cnv_continuous.tsv: copy number (CN), log2 transformed (log2(CN ratio + 1)) . Computed by a weighted average along the genomic coordinate.

  • cell_cnv_discretized.tsv: discretized gene copy numbers. -2, -1, 0, 1, 2 indicate deep deletion, heterozygous deletion, neutral, copy number gain, copy number amplification, respectively.

  • cell_mutation_delet.tsv: gene-level mutation counts in cell lines of all deleterious mutations.

  • cell_mutation_nonsynon.tsv: gene-level mutation counts in cell lines of all non-synonymous mutations.

  • cell_transcriptomics.tsv: TPM (transcripts per million reads mapped), log2 tranformed (log2(TPM+1)) gene expression values.

  • drug_SMILES.tsv: SMILES strings of drugs.

  • drug_SMILES.tsv: SMILES strings of drugs, canonicalized with RDKit.

  • drug_ecfp2_nbits256.tsv: ecfp2 binary fingerprints of drugs with 256 bits.

  • drug_ecfp2_nbits1024.tsv: ecfp2 binary fingerprints of drugs with 1024 bits.

  • drug_ecfp4_nbits256.tsv: ecfp4 binary fingerprints of drugs with 256 bits.

  • drug_ecfp4_nbits1024.tsv: ecfp4 binary fingerprints of drugs with 1024 bits.

  • drug_ecfp6_nbits256.tsv: ecfp6 binary fingerprints of drugs with 256 bits.

  • drug_ecfp6_nbits1024.tsv: ecfp6 binary fingerprints of drugs with 1024 bits.

  • drug_infomax.tsv: Infomax fingerprints of durgs.

  • drug_mordred.tsv: Mordred descriptors of drugs.

Data Sources

Source of data in the synergy benchmark dataset.

Data type	Source	Version	Link
Response	DrugComb	v1.5	https://drugcomb.org/download/
Drug SMILES	DrugComb	v1.5	https://drugcomb.org/drugs/
Cell line omics	DepMap	22Q2	https://depmap.org/portal/download/all/