Bioinformatics analysis and collection of protein post-translational modification sites in human viruses

In viruses, post-translational modifications (PTMs) are essential for their life cycle. Recognizing viral PTMs is very important for better understanding the mechanism of viral infections and finding potential drug targets. However, few studies have investigated the roles of viral PTMs in virus-human interactions using comprehensive viral PTM datasets. To fill this gap, firstly, we developed a viral post-translational modification database (VPTMdb) for collecting systematic information of viral PTM data. The VPTMdb contains 912 PTM sites that integrate 414 experimental-confirmed PTM sites with 98 proteins in 45 human viruses manually extracted from 162 publications and 498 PTMs extracted from UniProtKB/Swiss-Prot. Secondly, we investigated the viral PTM sequence motifs, the function of target human proteins, and characteristics of PTM protein domains. The results showed that (i) viral PTMs have the consensus motifs with human proteins in phosphorylation, SUMOylation and N-glycosylation. (ii) The function of human proteins that targeted by viral PTM proteins are related to protein targeting, translation, and localization. (iii) Viral PTMs are more likely to be enriched in protein domains. The findings should make an important contribution to the field of virus-human interaction. Moreover, we created a novel sequence-based classifier named VPTMpre to help users predict viral protein phosphorylation sites. Finally, an online web server was implemented for users to download viral protein PTM data and predict phosphorylation sites of interest. Author summaryPost-translational modifications (PTMs) plays an important role in the regulation of viral proteins; However, due to the limitation of data sets, there has been no detailed investigation of viral protein PTMs characteristics. In this manuscript, we collected experimentally verified viral protein post-translational modification sites and analysed viral PTMs data from a bioinformatics perspective. Besides, we constructed a novel feature-based machine learning model for predicting phosphorylation site. This is the first study to explore the roles of viral protein modification in virus infection using computational methods. The valuable viral protein PTM data resource will provide new insights into virus-host interaction.