Datasets Included

This page provides an overview of the datasets included in CoderData version 2.4. This package collects 18 diverse sets of paired molecular datasets with corresponding drug sensitivity data. All data here is reprocessed and standardized so it can be easily used as a benchmark dataset for drug response prediction machine learning models.

The dataset files are in csv format and are available at the link below:

Figshare record: https://api.figshare.com/v2/articles/33990850

Version: 2.4

Dataset Overview

Datasets and Modalities

Dataset	References	Model Type	Sample	Drug	Drug Descriptor	Experiments	Transcriptomics	Proteomics	Mutations	Copy Number
BeatAML	[1], [2]	ex vivo	1022	164	X	X	X	X	X
Bladder	[3]	patient derived organoid; patient derived xenograft; tumor; xenograft derived organoid	134	50	X	X	X		X	X
CCLE	[4]	cell line	502	24	X	X	X	X	X	X
Colorectal	[18]	patient derived organoid; tumor	61	10	X	X	X		X	X
CPTAC	[5]	tumor	1139				X	X	X	X
CTRPv2	[6], [7], [8]	cell line	847	460	X	X	X	X	X	X
FIMM	[9], [10]	cell line	52	52	X	X	X	X	X	X
GDSC v1	[23], [24], [25]	cell line	984	291	X	X	X	X	X	X
GDSC v2	[23], [24], [25]	cell line	806	169	X	X	X	X	X	X
gCSI	[21], [22]	cell line	570	43	X	X	X	X	X	X
HCMI	[11]	patient derived organoid; tumor; cell line	2129				X		X	X
Liver	[19]	patient derived organoid	62	76	X	X	X	X	X	X
MPNST	[12]	3D-MEDS; patient derived xenograft; tumor	133	39	X	X	X	X	X	X
NCI60	[13]	cell line	83	55712	X	X	X	X	X	X
Novartis	[20]	patient derived xenograft	385	25	X	X	X		X	X
Pancreatic	[14]	patient derived organoid	49	25	X	X	X		X	X
PRISM	[15], [16]	cell line	478	1416	X	X	X	X	X	X
Sarcoma	[17]	tumor; patient derived organoid	36	34	X	X	X		X

The table above lists the datasets included in CoderData version 2.4, along with references to their original publications, counts of samples and drugs, and the types of data available for each dataset.

CoderData includes the following data:

• Model Type - the biological model the samples derive from (cell line, patient derived organoid, patient derived xenograft, tumor, ex vivo, or 3D-MEDS microtissue)

• Sample - cell lines, patient-derived samples, or patient-derived organoids

• Drug - compounds tested for sensitivity

• Drug Descriptor - molecular descriptors for each drug (Morgan fingerprints from RDKit and 2D descriptors from Mordred; see Data Processing below)

• Experiments - dose-response experiments (various metrics such as AUC, IC50, etc.)

• Transcriptomics - gene expression (in transcripts per million, TPM)

• Proteomics - protein expression (log2 ratio to reference)

• Mutations - gene mutations (variant calls)

• Copy Number - gene copy number on the DepMap scale, log2(relative copy number + 1), where 1 is diploid, plus a copy call for each gene

An “X” indicates the presence of a particular data type for the corresponding dataset. Each sample in the datasets corresponds to either a cancer cell line, a patient-derived xenograft, or a patient-derived organoid, depending on the specific dataset.

Data Processing

Every dataset is reprocessed into the same formats, so data can be compared and combined across datasets. Genes are identified by Entrez ID, samples by improve_sample_id and drugs by improve_drug_id.

• Transcriptomics - gene expression in transcripts per million (TPM). CPTAC values are kept on the log2 scale provided by the cptac Python package.

• Proteomics - log2 ratio of each protein to the study’s reference. CPTAC values are kept on the log2 scale provided by the cptac package.

• Mutations - somatic variants for each gene, with variant types given MAF-style names (e.g. Missense_Mutation, Frame_Shift_Del, Silent). A few sources keep their own name for some types, e.g. Frameshift_Variant in BeatAML, Pancreatic and Sarcoma.

• Copy Number - each source’s values are converted to the DepMap scale, log2(relative copy number + 1), where 1 is diploid. Each value is also binned into a copy call: deep del (below 0.52), het loss (0.52 to 0.73), diploid (0.73 to 1.21), gain (1.21 to 1.42) and amp (above 1.42).

• Drugs - drug names are matched to PubChem to get each drug’s structure (canonical SMILES and InChIKey), and drugs with the same structure share one improve_drug_id across datasets.

• Drug Descriptors - computed from each drug’s canonical SMILES. Each drug has a morgan fingerprint (1024 bits, radius 2, from RDKit) and about 1,600 2D descriptors from Mordred. The structural_descriptor column holds the descriptor name.

• Experiments - for datasets screened in culture (cell lines, organoids, microtissues and BeatAML’s ex vivo samples), a Hill curve (Einf, EC50 and Hill slope) is fit to each sample and drug’s dose-response data, and the curve metrics below are calculated from that fit. PDX datasets (MPNST and Novartis) use tumor volume metrics (ABC, LMM, mRECIST, TGI) instead. Sarcoma has the published AUC only.

Dataset Summary Statistics

The following table summarizes combination counts for each dataset. This includes the number of experimental sample-drug pairs, with different molecular data types. Each column represents the number of unique combinations of samples and drugs with the specified molecular data types available. For example, the “Sample-Drug-Transcriptomics-Mutations” column indicates the number of unique sample-drug pairs that have both transcriptomics and mutation data available. These counts let you estimate how much paired data is available for tasks like building predictive models with transcriptomics and drug response.

Dataset Summary Statistics

dataset	sample_drug_pairs	sample_drug_transcript_pairs	sample_drug_transcriptomics_mutation_pairs	sample_drug_transcriptomics_copynumber_pairs	sample_drug_mutation_copynumber_pairs
beataml	62707	52985	48724
bladder	3300	840	640	640	3100
ccle	11543	10933	10885	8375	8398
colorectal	140	60	60	60	140
cptac
ctrpv2	310133	300871	297687	216046	216830
fimm	2663	2457	2457	2098	2098
gcsi	13229	12368	12259	9186	9248
gdscv1	245489	242980	239787	239034	240351
gdscv2	113964	112911	111387	111085	111687
hcmi
liver	4453	4453	4453	4453	4453
mpnst	2736	2657	181	181	181
nci60	2940159	2343136	2343121	2085608	2085608
novartis	1757	1725	1725	1714	1714
pancreatic	190	190	185	185	185
prism	637674	632199	632199	497743	499088
sarcoma	275	234	187

Drug Curve Metrics Collected

The following table summarizes the number of drugs associated with each dose-response metric across the datasets.

To compare drug response across datasets, use fit_auc. It is calculated over the same dose range (10⁻¹⁰ to 10⁻⁴ M) for every dataset with dose-response curves, so values are on the same scale. auc, aac and dss use each study’s own dose range.

Drug Curve Metrics Summary

dataset

num_drugs

aac

abc

auc

delta_auc

dss

fit_auc

fit_ec50

fit_ec50se

fit_einf

fit_hs

fit_ic50

fit_r2

lmm

mrecist

published_auc

tgi

beataml

162

X

X

X

X

X

X

X

X

X

X

bladder

50

X

X

X

X

X

X

X

X

X

X

ccle

24

X

X

X

X

X

X

X

X

X

X

colorectal

10

X

X

X

X

X

X

X

X

X

X

ctrpv2

460

X

X

X

X

X

X

X

X

X

X

fimm

52

X

X

X

X

X

X

X

X

X

X

gcsi

43

X

X

X

X

X

X

X

X

X

X

gdscv1

291

X

X

X

X

X

X

X

X

X

X

gdscv2

169

X

X

X

X

X

X

X

X

X

X

liver

73

X

X

X

X

X

X

X

X

X

X

mpnst

39

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

nci60

55712

X

X

X

X

X

X

X

X

X

X

novartis

23

X

X

X

X

pancreatic

5

X

X

X

X

X

X

X

X

X

X

prism

1416

X

X

X

X

X

X

X

X

X

X

sarcoma

32

X

Types of dose-response metrics collected include:

• AAC - Area above the response curve; the complement value of AUC.

• ABC - Area between curves, the difference between the AUC of the control and the treated cells. (PDX)

• AUC - Area under the fitted hill slope curve across all doses present. Lower AUC signifies lower levels of growth.

• delta_auc - Difference in AUC between a single drug and its combination (single-drug AUC minus combination AUC); used for MPNST treated-microtissue combination experiments.

• DSS - A multiparametric dose response value, that takes into account control and treated cells; the Drug Sensitivity Score is described in Yadav et al., 2014.

• fit_auc - Area under the fitted hill slope curve across the common interval of −log10[M], where the molar concentration ranges from 10⁻⁴ to 10⁻¹⁰.

• fit_ec50 - The fitted curve prediction of the −log10M concentration at which 50% of the maximal effect is observed.

• fit_ec50se - Standard error of the Fit_EC50 estimate.

• fit_einf - The fraction of cells that are unaffected even at an infinite dose concentration. Calculated as the lower asymptote of the hill slope function.

• fit_hs - The estimated hill slope binding cooperativity, calculated as the slope of the sigmoidal hill curve.

• fit_ic50 - The fitted curve prediction of the −log10M concentration required to reduce tumor growth by 50%.

• fit_r2 - Coefficient of determination between observed growth and the fitted hill slope curve, indicating goodness of fit.

• lmm - The resulting “time and treatment interaction” in a linear mixed model with fixed effects as time and treatment and with patient as a random effect. This indicates how much the treatment changes the slope of log(volume) over time compared to the control. (PDX)

• mRECIST - Disease status is classified into four categories: PD (progressive disease), SD (stable disease), PR (partial response), and CR (complete response). This is calculated as Day 10 percent volume change (%ΔV) and its cumulative average response under the following methods: CR if best %ΔV < −95% and avg < −40%, PR if %ΔV < −50% and avg < −20%, SD if %ΔV < +35% and avg < +30%, otherwise PD. (PDX)

• published_auc - Published Area Under the Curve

• TGI - The tumor growth inhibition (TGI) between the control and treatment time-volume curves. (PDX)

References

[1]

Bottomly D, Long N, Schultz AR, et al. Integrative analysis of drug response and clinical outcome in acute myeloid leukemia. Cancer Cell. 2022;40(8):850-864.e9. doi:10.1016/j.ccell.2022.07.002

[2]

Pino JC, Posso C, Joshi SK, et al. Mapping the proteogenomic landscape enables prediction of drug response in acute myeloid leukemia. Cell Rep Med. 2024;5(1):101359. doi:10.1016/j.xcrm.2023.101359

[3]

Lee SH, Hu W, Matulay JT, et al. Tumor Evolution and Drug Response in Patient-Derived Organoid Models of Bladder Cancer. Cell. 2018;173(2):515-528.e17. doi:10.1016/j.cell.2018.03.017

[4]

Barretina J, Caponigro G, Stransky N, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483(7391):603-607. doi:10.1038/nature11003

[5]

Lindgren CM, Adams DW, Kimball B, et al. Simplified and Unified Access to Cancer Proteogenomic Data. J Proteome Res. 2021;20(4):1902-1910. doi:10.1021/acs.jproteome.0c00919

[6]

Rees MG, Seashore-Ludlow B, Cheah JH, et al. Correlating chemical sensitivity and basal gene expression reveals mechanism of action. Nat Chem Biol. 2016;12(2):109-116. doi:10.1038/nchembio.1986

[7]

Seashore-Ludlow B, Rees MG, Cheah JH, et al. Harnessing Connectivity in a Large-Scale Small-Molecule Sensitivity Dataset. Cancer Discov. 2015;5(11):1210-1223. doi:10.1158/2159-8290.CD-15-0235

[8]

Basu A, Bodycombe NE, Cheah JH, et al. An interactive resource to identify cancer genetic and lineage dependencies targeted by small molecules. Cell. 2013;154(5):1151-1161. doi:10.1016/j.cell.2013.08.003

[9]

Mpindi JP, Yadav B, Östling P, et al. Consistency in drug response profiling. Nature. 2016;540(7631):E5-E6. doi:10.1038/nature20171

[10]

Pemovska T, Kontro M, Yadav B, et al. Individualized systems medicine strategy to tailor treatments for patients with chemorefractory acute myeloid leukemia. Cancer Discov. 2013;3(12):1416-1429. doi:10.1159/2159-8290.CD-13-0350

[11]

Human Cancer Models Initiative (HCMI). dbGaP accession phs001486. https://cancer.gov/ccg/research/functional-genomics/hcmi

[12]

Dehner C, Moon CI, Zhang X, et al. Chromosome 8 gain is associated with high-grade transformation in MPNST. JCI Insight. 2021;6(6):e146351. doi:10.1172/jci.insight.146351

[13]

Shoemaker RH. The NCI60 human tumour cell line anticancer drug screen. Nat Rev Cancer. 2006;6(10):813-823. doi:10.1038/nrc1951

[14]

Tiriac H, Belleau P, Engle DD, et al. Organoid Profiling Identifies Common Responders to Chemotherapy in Pancreatic Cancer. Cancer Discov. 2018;8(9):1112-1129. doi:10.1158/2159-8290.CD-18-0349

[15]

Corsello SM, Nagari RT, Spangler RD, et al. Discovering the anti-cancer potential of non-oncology drugs by systematic viability profiling. Nat Cancer. 2020;1(2):235-248. doi:10.1038/s43018-019-0018-6

[16]

Yu C, Mannan AM, Yvone GM, et al. High-throughput identification of genotype-specific cancer vulnerabilities in mixtures of barcoded tumor cell lines. Nat Biotechnol. 2016;34(4):419-423. doi:10.1038/nbt.3460

[17]

Al Shihabi A, Tebon PJ, Nguyen HTL, et al. The landscape of drug sensitivity and resistance in sarcoma. Cell Stem Cell. 2024;31(10):1524-1542.e4. doi:10.1016/j.stem.2024.08.010

[18]

van de Wetering M, Francies HE, Francis JM, et al. Prospective derivation of a living organoid biobank of colorectal cancer patients. Cell. 2015;161(4):933-945. doi:10.1016/j.cell.2015.03.053

[19]

Ji S, Feng L, Fu Z, et al. Pharmaco-proteogenomic characterization of liver cancer organoids for precision oncology. Sci Transl Med. 2023;15(706):eadg3358. doi:10.1126/scitranslmed.adg3358

[20]

Gao H, Korn JM, Ferretti S, et al. High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response. Nat Med. 2015;21(11):1318–1325. doi:10.1038/nm.3954

[21]

Haverty PM, Lin E, Tan J, et al. Reproducible pharmacogenomic profiling of cancer cell line panels. Nature. 2016;533(7603):333–337. doi:10.1038/nature17987

[22]

Klijn C, Durinck S, Stawiski EW, et al. A comprehensive transcriptional portrait of human cancer cell lines. Nat Biotechnol. 2015;33(3):306–312. doi:10.1038/nbt.3080

[23] (1,2)

Garnett MJ, Edelman EJ, Heidorn SJ, et al. Systematic identification of genomic markers of drug sensitivity in cancer cells. Nature. 2012;483(7391):570–575. doi:10.1038/nature11005

[24] (1,2)

Iorio F, Knijnenburg TA, Vis DJ, et al. A Landscape of Pharmacogenomic Interactions in Cancer. Cell. 2016;166(3):740–754. doi:10.1016/j.cell.2016.06.017

[25] (1,2)

Yang W, Soares J, Greninger P, et al. Genomics of Drug Sensitivity in Cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. Nucleic Acids Res. 2013;41(Database issue):D955–D961. doi:10.1093/nar/gks1111