In plain English
The team used AI to narrow a yeast genome to peptide candidates, then made the peptides and tested whether they actually inhibited bacteria.
How the study worked
A plain-language walk through the work behind the result.
Generated peptide fragments from two Malassezia furfur genomes.
Filtered candidates by physical properties and deep-learning antimicrobial-peptide scores.
Synthesized selected peptides and tested activity, structure, membrane interaction, and keratinocyte toxicity.
What they found
- The genome-guided AI screen enriched for peptides with measurable activity against Staphylococcus aureus.
- Experimental results also exposed bias and generalization limits in the prediction models.
Why it matters
The study tests the part of AI peptide discovery that matters most: whether ranked sequences survive synthesis and functional validation.
The catch
- The record is a preprint and has not completed peer review.
- This summary is based on the abstract; methods and supplementary analyses were not independently rechecked.