Skip to main content

Table 10 The performance of different modules on HLA-G*01:03 dataset

From: DeepHLAPred: a deep learning-based method for non-classical HLA binder prediction

HLA-G*01:03

SN

SP

ACC

MCC

AUC

Model

     

 CNN

0.9372

0.9039

0.9208

0.8424

0.9726

 Bi-LSTM

0.9257

0.9241

0.9248

0.8499

0.9716

 CNN + Bi-LSTM (In series)

0.9281

0.9402

0.9341

0.8686

0.9786

 CNN + CNN (In parallel)

0.9321

0.9333

0.9328

0.8656

0.9748

 Bi-LSTM + Bi-LSTM (In parallel)

0.9309

0.9437

0.9374

0.8749

0.9703

 DeepHLAPred

0.9454

0.9626

0.9541

0.9083

0.9812