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

Abstract

The study of protein-protein interaction is of great biological significance, and the prediction of protein-protein interaction sites can promote the understanding of cell biological activity and will be helpful for drug development. However, uneven distribution between interaction and non-interaction sites is common because only a small number of protein interactions have been confirmed by experimental techniques, which greatly affects the predictive capability of computational methods. In this work, two imbalanced data processing strategies based on XGBoost algorithm were proposed to re-balance the original dataset from inherent relationship between positive and negative samples for the prediction of protein-protein interaction sites. Herein, a feature extraction method was applied to represent the protein interaction sites based on evolutionary conservatism of proteins, and the influence of overlapping regions of positive and negative samples was considered in prediction performance. Our method showed good prediction performance, such as prediction accuracy of 0.807 and MCC of 0.614, on an original dataset with 10,455 surface residues but only 2297 interface residues. Experimental results demonstrated the effectiveness of our XGBoost-based method.

MicroRNA-126 Modulates Palmitate-Induced Migration in HUVECs by Downregulating Myosin Light Chain Kinase via the ERK/MAPK Pathway

“…Dephosphorylation of MLC results in cell relaxation. The catalytic activity of MLCP are inhibited by the phosphorylation of MYPT1 at multiple sites by several kind of kinases, therefore, led to decreased dephosphorylated MLC and thus, vascular smooth muscle contraction (Qiao et al, 2014;Chang et al, 2016;Gao et al, 2017;Deng et al, 2020). The current study found that miR-126 decreased phosphorylation levels of MYPT1 in palmitate-treated HUVECs, which may account for the observed decrease in MLC phosphorylation.…”

Integrated Analysis of DEAD-Box Helicase 56: A Potential Oncogene in Osteosarcoma

“…Given these limitations, the models generated here must be considered preliminary and incomplete. Other predictive tools, such as predictions of protein interactions based on molecular structure, specific groups, may be good cross-validations (Deng et al, 2020 ; Hu et al, 2020 ). Further studies should also be done to illustrate the underlying mechanisms.…”

LncLocation: Efficient Subcellular Location Prediction of Long Non-Coding RNA-Based Multi-Source Heterogeneous Feature Fusion

“…XGBoost was used on the benchmark dataset. XGBoost can automatically learn the optimal missing value according to the training loss and more effectively process different types of sparse patterns in the data, which is more stable for different batches of training [ 23 , 24 , 25 ].…”