Data for: Predicting eukaryotic protein secretion without signals

Datasets extracted from UniProt for benchmarking methods for predicting non-classical secretion. Data are in two formats: FASTA sequences (*.fast) and lists of UniProt accession codes (*.ac.txt). The *_secreted_* files are proteins with subcellular location confirmed as secreted, but without signal peptides. The *_non_secreted_* files are proteins with subcellular location confirmed as cytoplasmic or nuclear. The mammalia_* files are from mammals; the eukaryota_* files are from eukaryotes excluding mammals.