CFAGO:

cross-fusion of network and attributes based on attention mechanism for protein function prediction


Introduction

CFAGO is a new computational method for frotein function prediction via fusion network structure and node attribute by using attention mechanism. The architecture of CFAGO is shown in Fig. 1.

Architecture of CFAGO
Fig 1. The flowchart of CFAGO. (A) Structure of self-supervised learning step, where MLP stands for multilayer perceptron. (B) Structure of fine-tuning step. (C) Multi-head attention layer of Encoder and Decoder. (D) Scaled Dot-Product Attention of Multi-head attention layer.

source code

The source code of CFAGO can be download from below link:

Download CFAGO

Download datasets

Download Dataset



CONTACT US

Bin Liu
bliu@bliulab.net
School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen, Guangdong 518055, China.

Copyright © by Liu Lab, Harbin Institute of Technology

网站备案号:粤ICP备19041859号-1