Abstract
The article presents a dataset containing nine classes of calculated sequence-derived descriptors for 78 peptide sequences, 21 of which demonstrate the ability to bind with gastric cancer cells. The datasaet was used in the paper "A screening algorithm for gastric cancer binding peptides" [1] for the creation of a classification model that can predict the ability of a given peptide sequence to bind with gastric cancer cells. The 78 peptide sequences were extracted from a systematic literature search, and the various peptide descriptors were calculated using the R package "Peptides". The nine calculated sequence-derived descriptor classes are the Blosum indices, Cruciani properties, FASGAI vectors, Kidera factors, ProtFP, ST-scales, T-scales, VHSE scales, and Z-scales. The resulting dataset, which is composed of over 4000 data points, offers a rich resource for further protochemometric analyses of the curated peptide sequences relevant to cancer diagnostics and therapeutics.