Non-Hemolytic and Non-Toxic Antimicrobial Peptides Commonly Identified from Cephalopods PSG by Three Prediction Models

The presented dataset comprises subsets of non-hemolytic antimicrobial peptides (AMPs), devoid from toxic signatures. These AMP subsets were identified using a consensus prediction approach that involved three machine learning (ML) models. The toxicity models were applied to non-hemolytic AMPs libraries (DOI: 10.17632/pvptjh7kmv.1), which originated from previously predicted AMPs (DOI: 10.17632/wwk7zzcfhv.1). These AMP libraries were derived from the screening of 13 peptide libraries (DOI: 10.17632/6fjsdnvygb.1) based on omics data from Cephalopods' Posterior Salivary Glands (DOI: 10.17632/gxmkytwdhx.1). The provided dataset includes 13 peptide FASTA files, each meticulously annotated with information regarding the enzyme(s) involved in the in silico digestion process on the aforementioned omics data. The non-hemolytic and non-toxic AMPs are categorized based on the application of a single enzyme (OE) or two enzymes (TE), with the latter being applied either sequentially (S) or concurrently (C). These identified non-hemolytic and non-toxic AMPs represent a unique and yet unexplored chemical landscape, holding immense potential to drive the discovery of novel peptide-based pharmaceutical agents.