Skip to main content
. 2023 May 30;39(6):btad355. doi: 10.1093/bioinformatics/btad355

Table 3.

Performance evaluation on more realistic settings of Davis datasets.

Scenario Method MSE CI rm2
Cold drug GraphDTA 0.920(0.029) 0.678(0.036) 0.160(0.019)
GEFA 0.847(0.012) 0.709(0.028) 0.182(0.015)
FusionDTA 0.581(0.094) 0.737(0.012) 0.187(0.034)
MgraphDTA 0.563(0.065) 0.729(0.022) 0.192(0.021)
NHGNN(Ours) 0.554(0.091) 0.752(0.017) 0.207(0.030)
Cold target GraphDTA 0.510(0.086) 0.729(0.012) 0.154(0.014)
GEFA 0.433(0.022) 0.759(0.009) 0.289(0.016)
FusionDTA 0.364(0.021) 0.826(0.011) 0.435(0.023)
MgraphDTA 0.359(0.023) 0.813(0.008) 0.425(0.028)
NHGNN(Ours) 0.344(0.029) 0.855(0.016) 0.479(0.021)
All cold GraphDTA 0.968(0.096) 0.579(0.017) 0.026(0.016)
GEFA 0.944(0.092) 0.610(0.029) 0.032(0.022)
FusionDTA 0.876(0.091) 0.645(0.043) 0.072(0.048)
MgraphDTA 0.874(0.090) 0.636(0.021) 0.071(0.041)
NHGNN(Ours) 0.857(0.096) 0.665(0.038) 0.087(0.051)

Bold corresponds to the best performance for each metric, and underline indicates the second best.