Table 2 Performance of Models with Two Independent Modalities and Transformer Replaced by LSTM.

From: A multimodal model for protein function prediction

Method

\(F_{max}\)

AUPRC

MF

BP

CC

MF

BP

CC

Sequence only

0.423

0.211

0.398

0.433

0.295

0.302

Structure only

0.461

0.299

0.455

0.479

0.288

0.352

GCN with LSTM

0.685

0.578

0.634

0.579

0.316

0.491

GCN with transformer

0.769

0.632

0.695

0.721

0.401

0.495

Back to article page