Skip to main content
. 2023 Jan 19;6:73. doi: 10.1038/s42003-023-04462-5

Table 2.

Performance comparison under different protein language models.

TPR TNR Pre ACC F1 MCC AUROC AP
Dset_448
ESM-1b 0.725 0.652 0.246 0.662 0.368 0.264 0.759 0.349
ProGen2 0.714 0.602 0.220 0.617 0.336 0.218 0.715 0.287
ProtT5 0.448 0.916 0.455 0.852 0.451 0.366 0.810 0.442
Dset_72
ESM-1b 0.674 0.678 0.199 0.678 0.307 0.226 0.738 0.253
ProGen2 0.684 0.554 0.154 0.568 0.251 0.147 0.658 0.176
ProtT5 0.336 0.908 0.303 0.848 0.319 0.234 0.770 0.300
Dset_164
ESM-1b 0.732 0.542 0.261 0.576 0.375 0.211 0.690 0.324
ProGen2 0.758 0.436 0.229 0.495 0.352 0.153 0.650 0.288
ProtT5 0.308 0.909 0.427 0.800 0.358 0.248 0.740 0.384