Table 2 Comparison of the performance of our predictors using different pre-trained models test data sets.

From: Accurate prediction of protein–ATP binding sites based on a protein pretrained large language model and a fractional-order convolutional neural network

Pre-trained model

Acc

Sen

Spe

Pre

MCC

AUC

A: ATP-17 test set

ProtBert

0.989

0.222

0.998

0.514

0.333

0.930

ESM-1b

0.973

0.544

0.991

0.718

0.611

0.958

ESM-2

0.990

0.391

0.997

0.642

0.497

0.965

B: ATP-41 test set

ProtBert

0.981

0.151

0.999

0.744

0.192

0.925

ESM-1b

0.983

0.194

0.999

0.857

0.403

0.941

ESM-2

0.984

0.488

0.994

0.621

0.543

0.952

Back to article page