Natural products form the basis for most drugs in clinical use. Advances in genome sequencing and bioinformatic tools have revealed thousands of biosynthetic gene clusters encoding these products. However, linking natural products identified by genome mining to their corresponding products in untargeted metabolomics data remains a key challenge. Here we present a platform, DeepRiPP, which integrates genomic and metabolomic data to automate the discovery of new ribosomally synthesized posttranslationally modified peptides (RiPPs), a subclass of natural products with diverse chemistry and activities. We apply DeepRiPP to discover 3 novel RiPPs.