Adaptive Residue-wise Profile Fusion for Low Homologous Protein SecondaryStructure Prediction Using External Knowledge

TL;DR

This paper introduces an adaptive residue-wise profile fusion method utilizing external self-supervised BERT features to improve protein secondary structure prediction for low homologous proteins, outperforming existing methods.

Contribution

It proposes a novel residue-wise attention mechanism and feature consistency loss to effectively fuse profile and BERT features, enhancing prediction accuracy in low homologous scenarios.

Findings

Outperforms state-of-the-art methods by 4.7% on BC40 dataset for low homologous proteins.

Demonstrates the superiority of profile over PSSM in low homologous PSSP.

Validates the effectiveness of BERT-based pseudo profiles and adaptive fusion in improving prediction accuracy.

Abstract

Protein secondary structure prediction (PSSP) is essential for protein function analysis. However, for low homologous proteins, the PSSP suffers from insufficient input features. In this paper, we explicitly import external self-supervised knowledge for low homologous PSSP under the guidance of residue-wise profile fusion. In practice, we firstly demonstrate the superiority of profile over Position-Specific Scoring Matrix (PSSM) for low homologous PSSP. Based on this observation, we introduce the novel self-supervised BERT features as the pseudo profile, which implicitly involves the residue distribution in all native discovered sequences as the complementary features. Further-more, a novel residue-wise attention is specially designed to adaptively fuse different features (i.e.,original low-quality profile, BERT based pseudo profile), which not only takes full advantage of each feature but also avoids noise disturbance. Be-sides, the feature consistency loss is proposed to accelerate the model learning from multiple semantic levels. Extensive experiments confirm that our method outperforms state-of-the-arts (i.e.,4.7%forextremely low homologous cases on BC40 dataset).