Rxivist logo

Prediction of Protein-Protein Interactions Based on L1-Regularized Logistic Regression and Gradient Tree Boosting

By Bin Yu, Cheng Chen, Hongyan Zhou, Bingqiang Liu, Qin Ma

Posted 05 Mar 2020
bioRxiv DOI: 10.1101/2020.03.04.976365

Protein-protein interactions (PPIs) are of great importance to understand genetic mechanisms, disease pathogenesis, and guide drug design. With the increase of PPIs sequence data and development of machine learning, the prediction and identification of PPIs have become a research hotspot in proteomics. In this paper, we propose a new prediction pipeline for PPIs based on gradient tree boosting (GTB). First, the initial feature vector is extracted by fusing pseudo amino acid composition (PseAAC), pseudo-position-specific scoring matrix (PsePSSM), reduced sequence and index-vectors (RSIV) and autocorrelation descriptor (AD). Second, to remove redundancy and noise, we employ L1-regularized logistic regression to select an optimal feature subset. Finally, GTB-PPI model based on GTB is constructed. Five-fold cross-validation showed GTB-PPI achieved the accuracies of 95.15% and 90.47% on Saccharomyces cerevisiae and Helicobacter pylori, respectively. In addition, GTB-PPI could be applied to predict Caenorhabditis elegans, Escherichia coli, Homo sapiens, and Mus musculus independent test sets, the one-core PPIs network for CD9, and the crossover PPIs network. The results show that GTB-PPI can significantly improve prediction accuracy of PPIs. The code and datasets of GTB-PPI can be downloaded from https://github.com/QUST-AIBBDRC/GTB-PPI/.

Download data

  • Downloaded 251 times
  • Download rankings, all-time:
    • Site-wide: 109,252
    • In bioinformatics: 9,232
  • Year to date:
    • Site-wide: 99,996
  • Since beginning of last month:
    • Site-wide: 95,934

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide