VPTMdb: a viral post-translational modification database

Yujia Xiang; Quan Zou; Lilin Zhao

The VPTMdb web interface was written in the R programming language using the Rshiny web development framework (Chang, et al., 2018) . The MySQL database management system was used to store structured PTM data. The base machine learning predictor (such as SVM) was supported by the caret R package (Kuhn, 2020) ; the ROC curve was analysed using ROCR (Sing, et al., 2005) ; and MRMR and t-SNE were analysed using mRMRe (De Jay, et al., 2013) and Rtsne (Krijthe, 2015) . Software ggplot2 was used to plot beautiful pictures (Wickham, 2016) . The website is free and can be browsed in most modern browsers. Figure 1A shows that the VPTMdb webserver consists of two parts: VPTM database and VPTMpre. The VPTM database currently includes 414 unique experimentally determined PTM sites with 8 modification types from 45 viruses. In summary, 162 manually checked references were collected in the database. Each entry in VPTMdb includes the (i) virus name, (ii) virus protein name in the UniProt database, (iii) viral modification site, (iv) residue sequences, and (v) PubMed id. PTM data from UniProtKB/Swiss-Prot contain two types: 199 phosphorylation sites and 299 glycosylation sites (N-lined and O-lined). author/funder. All rights reserved. No reuse allowed without permission.