Protein–protein interaction site prediction through combining local and global features with deep neural networks

Abstract

Motivation Protein–protein interactions (PPIs) play important roles in many biological processes. Conventional biological experiments for identifying PPI sites are costly and time-consuming. Thus, many computational approaches have been proposed to predict PPI sites. Existing computational methods usually use local contextual features to predict PPI sites. Actually, global features of protein sequences are critical for PPI site prediction. Results A new end-to-end deep learning framework, named DeepPPISP, through combining local contextual and global sequence features, is proposed for PPI site prediction. For local contextual features, we use a sliding window to capture features of neighbors of a target amino acid as in previous studies. For global sequence features, a text convolutional neural network is applied to extract features from the whole protein sequence. Then the local contextual and global sequence features are combined to predict PPI sites. By integrating local contextual and global sequence features, DeepPPISP achieves the state-of-the-art performance, which is better than the other competing methods. In order to investigate if global sequence features are helpful in our deep learning model, we remove or change some components in DeepPPISP. Detailed analyses show that global sequence features play important roles in DeepPPISP. Availability and implementation The DeepPPISP web server is available at http://bioinformatics.csu.edu.cn/PPISP/. The source code can be obtained from https://github.com/CSUBioGroup/DeepPPISP. Supplementary information Supplementary data are available at Bioinformatics online.

Developing Computational Model to Predict Protein-Protein Interaction Sites Based on the XGBoost Algorithm

“…It can be found that the values of six measures achieved in these three independent benchmark datasets are lower than that in the dataset used in this work, but it is reasonable because the IHT-XGB model is built on this original dataset. Table 4 also shows the prediction performance of five models, i.e., SSWRF, LORIS, PSIVER, SCRIBER, and DELPHI, in identification of protein-protein interaction sites [31][32][33][34][35][36][37]. The highest results in each performance measures for different models are highlighted as bold type.…”

EGRET: Edge Aggregated Graph Attention Networks and Transfer Learning Improve Protein-Protein Interaction Site Prediction

“…See Appendix K for additional discussion. scheme used by [15] (available at: https://github.com/CSUBioGroup/DeepPPISP). Table A1 shows the numbers of interaction and non-interaction sites in these datasets and the splits used in this study and [15].…”

Graphein - a Python Library for Geometric Deep Learning and Network Analysis on Protein Structures and Interaction Networks

“…As examples, we make available two graph-based protein structure datasets. The first, based on the collections outlined in (Zeng et al, 2019), consists of 420 proteins, with node labels indicating whether a residue is involved in a protein-protein interaction. The interaction status data and structure originate from structures of the complexes in the RCSB PDB.…”