Improved prediction of protein–protein interactions using novel negative samples, features, and an ensemble classifier

作者：

Highlights：

• A novel strategy of negative sample construction is proposed.

• Two novel types of feature extraction methods are proposed.

• Our machine learning based prediction model exhibits the efficiency and effectiveness to predict protein–protein interactions, significantly reducing false negatives.

摘要

•A novel strategy of negative sample construction is proposed.•Two novel types of feature extraction methods are proposed.•Our machine learning based prediction model exhibits the efficiency and effectiveness to predict protein–protein interactions, significantly reducing false negatives.

论文关键词：Protein-protein interaction,Negative samples,Feature selection,Ensemble classifier,Protein–protein non-interactions

论文评审过程：Received 20 December 2016, Revised 17 February 2017, Accepted 1 March 2017, Available online 4 March 2017, Version of Record 17 November 2017.

论文官网地址：https://doi.org/10.1016/j.artmed.2017.03.001