Datasets and models for Metagenomics-AI

Datasets and models for Metagenome-AI framework and related manuscript4 datasets in total: Global AMP prediction, Gram positive, Gram negative bacteria and toxicity.3 Base models: ESM2-3B, ESM3 and ProteinTrans.Proteins collected from extreme environments are given in 11379_corepeptides_scored_cAMP20250915-forRevision.tsv