RegVar is a deep neural network-based computational server for prioritizing tissue-specific noncoding regulatory variants. RegVar integrates the sequential, epigenetic and evolutionary profiles of single nucleotide polymorphisms (SNPs) and their potential target genes in 17 human tissues, and gives tissue-specific predictions of regulatory probabilities of the provided SNPs on provided genes. To cite RegVar: Lu, H. et al. RegVar: Tissue-specific Prioritization of Noncoding Regulatory Variants. Genom Proteom Bioinform (2021) doi:10.1016/j.gpb.2021.08.011.