DeepMVP is a deep learning framework for predicting PTM sites and variant-induced alterations across six modification types, including phosphorylation, acetylation, methylation, sumoylation, ubiquitination and N-glycosylation.