Table 9 Performance comparison of feature representation methods using their respective optimized CNN architectures for Khib site prediction in the B. cinerea dataset. Boldface values indicate the best performance for each metric.

From: A deep learning framework for lysine 2-hydroxyisobutyrylation site prediction using evolutionary feature representation

Feature representation method

10-fold cross-validation set

Independent test set

ACC

F1

MCC

AUC

ACC

F1

MCC

AUC

ESM

0.706

0.635

0.368

0.750

0.676

0.691

0.373

0.748

One hot

0.693

0.681

0.370

0.748

0.670

0.667

0.324

0.737

CTD

0.735

0.740

0.443

0.778

0.708

0.767

0.465

0.794

PSSM

0.759

0.739

0.461

0.801

0.733

0.754

0.459

0.803

AAP

0.790

0.796

0.583

0.866

0.791

0.804

0.583

0.865

BLOSUM

0.800

0.806

0.604

0.882

0.819

0.833

0.635

0.903

Back to article page