Fig. 1From: Decrypting orphan GPCR drug discovery via multitask learningFlowchart of the stacked ensemble multitask learning models for the GPCR bioactivities. The human GPCR-ligand pair activities database was extracted from GPCRdb. The models were generated from the individual or integrated receptors training datasets. The validation datasets and test orphan datasets validated the models independentlyBack to article page