Updated project metadata. A benchmark set of bottom-up proteomics data for training deep learning networks. It has data from 51 organisms and includes nearly 1 million peptides.