Skip to main content
. 2022 Nov 14;53(12):15246–15260. doi: 10.1007/s10489-022-04280-y

Table 5.

Comparison with state-of-the-art methods in terms of the ROC-AUC scores of each task on Tox21. The best performance is denoted by bold

Tasks Smi2Vec*-LSTM Smi2Vec-BiGRU TranGRU
NR-AR 0.6914 0.7114 0.8241
NR-AR-LBD 0.7477 0.8243 0.847
NR-AhR 0.678 0.8793 0.8327
NR-Aromatase 0.4964 0.6985 0.7838
NR-ER 0.6231 0.736 0.6907
NR-ER-LBD 0.5308 0.8675 0.8433
NR-PPAR-gamma 0.5659 0.7494 0.8375
SR-ARE 0.6414 0.761 0.7014
SR-ATAD5 0.5 0.7632 0.7267
SR-HSE 0.612 0.7845 0.7362
SR-MMP 0.7425 0.8599 0.8156
SR-p53 0.518 0.7321 0.81

The performance of Smi2Vec*-LSTM and Smi2Vec-BiGRU below are from the original papers