Machine learning prediction of antiviral-HPV protein interactions for anti-HPV pharmacotherapy

Abstract

Persistent infection with high-risk types Human Papillomavirus could cause diseases including cervical cancers and oropharyngeal cancers. Nonetheless, so far there is no effective pharmacotherapy for treating the infection from high-risk HPV types, and hence it remains to be a severe threat to the health of female. Based on drug repositioning strategy, we trained and benchmarked multiple machine learning models so as to predict potential effective antiviral drugs for HPV infection in this work. Through optimizing models, measuring models’ predictive performance using 182 pairs of antiviral-target interaction dataset which were all approved by the United States Food and Drug Administration, and benchmarking different models’ predictive performance, we identified the optimized Support Vector Machine and K-Nearest Neighbor classifier with high precision score were the best two predictors (0.80 and 0.85 respectively) amongst classifiers of Support Vector Machine, Random forest, Adaboost, Naïve Bayes, K-Nearest Neighbors, and Logistic regression classifier. We applied these two predictors together and successfully predicted 57 pairs of antiviral-HPV protein interactions from 864 pairs of antiviral-HPV protein associations. Our work provided good drug candidates for anti-HPV drug discovery. So far as we know, we are the first one to conduct such HPV-oriented computational drug repositioning study.

Overview of methods for characterization and visualization of a protein–protein interaction network in a multi-omics integration context

“…Each pair of proteins is split into several residues, with two residues of two proteins in the same cluster interacting. In terms of performance, they achieved an accuracy value of 60%, which is generally lower than those of the SVM and ANN methods, due to the difference in the amount of information available on the proteins ( Ahmed, 2020 ; Jonathan et al, 2021 ; Lin et al, 2021 ).…”

“…; X m ,Y), where X{X1,X 2, X 3, ….,X m } is the m-dimensional input variable and Y is the output variable taking {0,1}. As input, this method can take either protein interaction datasets or genomic interaction datasets ( Jansen et al, 2003 ; Alashwal, Deris and Othman, 2009 ; Lin et al, 2021 ). In the end, the classifier gives a binary response, a zero indicating the interaction is not verified, and a one when there is a potential interaction.…”

A comprehensive review for machine learning based human papillomavirus detection in forensic identification with multiple medical samples

“…In a 2021 study, 57 antiviral HPV protein interactions were successfully predicted from the 864 antiviral HPV protein associations ( Lin H. H. et al, 2021 ). The investigators used data from DrugBank, Drugs@FDA, PubChem, Uniprot, and Therapeutic Targets, which collected data databases on antiviral drugs and their associated Targets.…”

Epigenetic and Transcriptomic Regulation Landscape in HPV+ Cancers: Biological and Clinical Implications

“…Regarding the combination of HDAC inhibitors with antivirals to eradicate HPV infections, hitherto, there are no HPV-specific antiviral agents approved for clinical use, however, machine learning and therapeutic repositioning are emerging strategies to identify antiviral molecules. Lin et al (2021) identified 57 antiviral drugs with potential interaction with E1, E2, E4, E5, E6, E7 L1, and E8^E2 ( Lin et al, 2021 ). However, the effectiveness of these candidates and their impact on epigenetic alterations have not yet evaluated.…”