Skip to main content

Table 4 Experimental results in the unknown drug data sets from Kiba

From: GeneralizedDTA: combining pre-training and multi-task learning to predict drug-target binding affinity for unknown drug discovery

Model MSE \(R^{2}\)
DeepDTA 0.5437 0.3605
GraphDTA 0.4950 0.2953
SAGDTA 0.6237 0.2311
MGraphDTA 0.4667 0.3766
Our method (\(\lambda _{\text{ atom } }=0\)) 0.7311 0.2039
Our method (\(\lambda _{\text{ atom } }=0.5\)) 0.4331 0.2831
Our method (\(\lambda _{\text{ atom } }=1.0\)) 0.4582 0.3906
Our method (\(\lambda _{\text{ atom } }=2.0\)) 0.6067 0.1781