Table 1.

Number of additional datasets used by participants for training and algorithm class

Team	SC1	SC2	No. of drug-AUC datasetsa	No. of drug-mRNA datasetsb	No. of drug-target datasetsc	Total training datasets	Algorithm class
Netphar	12.6	70.9	6	1	4	11	similarity
SBNB	11.7	59.2	6	3	2	11	similarity
Xielab	13.8	50.3	6	2	1	9	similarity
Atom	17.4	49.3	–	2	4	6	NN
DMIS_PDA	13.8	35.2	–	2	1	3	NN
Theragen	15.1	17.3	–	2	1	3	similarity
Signal	6.3	6.1	–	1	2	4	regression
TeamAxolotl	6.2	1.1	–	–	2	3	NN
AMbeRland	3.3	1.1	–	–	–	0	unsup.
SenthamizhaV	7.4	0.9	–	–	–	0	unsup.

^aDrug sensitivity (AUC) databases include: NCI60,²⁷ GDSC,²² CTRP,²⁸ gCSI,²⁹ CCLE,²¹ and other manually curated data.

^bDrug mRNA perturbation databases include: L1000-drugs, L1000-shRNA,²³ and CREEDS.³⁰

^cDrug target datasets include: DrugBank,³¹ ChEMBL,³² KEGG,³³ and MATADOR.³⁴