HOME   |    PDF   |   


Prediction of transient and permanent protein interactions using AI methods



A. Kiran Kumar*, Syed Mohammad Shayez Karim, Mayank Kumar & Ravindranath Singh Rathore*



Department of Bioinformatics, Central University of South Bihar, Gaya, Bihar-824236, India; *Corresponding authors



Kiran Kumar - E-mail: kirankumar@cusb.ac.in

Shayez Karim - E-mail: shayazkarim@cusb.ac.in

Mayank Kumar - E-mail: mayank@cusb.ac.in

R.S. Rathore - E-mail: rsrathore@cusb.ac.in


Article Type

Research Article



Received June 1, 2023; Revised June 30, 2023; Accepted June 30, 2023, Published June 30, 2023



Protein-protein interactions (PPIs) can be classified as permanent or transient interactions based on their stability or lifetime. Understanding the precise details of such protein interactions will pave the way for the discovery of inhibitors and for understanding the nature and function of PPIs. In the present work, 43 relevant physicochemical, geometrical and structural features were calculated for a curated dataset from the literature, comprising of 402 protein-protein complexes of permanent and transient categories, and 5 different Supervised Machine Learning models were developed with Scikit-learn to predict transient and permanent PPI. Additionally, deep learning method with Artificial Neural Network was also performed using Tensor Flow and Keras. Predicted models achieved accuracy ranging from 76.54% to 82.71% and k-NN has achieved the highest accuracy. Detailed analysis of these methods revealed that Interface areas such as Percent interface accessible area, Interface accessible area and Total interface area and the parameters defining the shape of the PPI interface such as Planarity, Eccentricity and Circularity are the most discriminating factors between these two categories. The present method could serve as an effective tool to understand the mechanism of protein association and to predict the transient and permanent interactions, which could supplement the costly and time-consuming experimental techniques.



Transient and Permanent Protein-Protein Interactions; Machine Learning; Scikit-learn; Deep Learning; Tensor Flow.



Kiran Kumar et al. Bioinformation 19(6): 749-753 (2023)


Edited by

P Kangueane






Biomedical Informatics



This is an Open Access article which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. This is distributed under the terms of the Creative Commons Attribution License.