Answers are grounded in this paper and cite their sources.
Computational mining of microbiomes has expanded antimicrobial peptide (AMP) discovery from individual organisms to millions of predicted sequences. The challenge is now to identify candidates that merit experimental investigation and to determine what their validation reveals about biological function. Ecological information may describe sequence provenance, support interpretation after prediction, guide candidate prioritization or contribute directly to model development. These uses are distinct, and the presence of ecological metadata does not by itself show that context improved discovery. Its added value must be tested against an appropriate alternative, such as sequence-only ranking under the same experimental budget, ablation of contextual variables or validation in held-out habitats. Functional evidence requires equally careful interpretation. Genomic coding potential, transcription, proteomic peptide detection, native mature-product identification and antimicrobial activity support different conclusions. Therapeutic efficacy in an infection model does not establish a peptide’s role in its source community, while ecological association cannot substitute for molecular detection or causal perturbation. By comparing representative computational and experimental studies, this Review examines where ecological context changes AMP discovery decisions, how that contribution can be evaluated and how validation should be matched to the biological question. Progress will depend less on expanding candidate catalogs alone than on transparent prioritization, preserved molecular identity and experiments designed for the claims being made.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: