Machine Learning Picks a Heat-Stable Cationic Peptide, P2, That Inhibits Candida albicans and C. parapsilosis in Lab Tests
Researchers at Mexico's Ensenada Center for Scientific Research and Higher Education (CICESE) reported in the Journal of Applied Microbiology on October 22, 2026 that of four machine-learning-selected peptides, one called P2 had the best activity profile. It inhibited two Candida albicans reference strains and a C. parapsilosis clinical isolate at 62.5 µM, while C. glabrata needed 250 µM, and it still killed bacteria after 15 minutes at 100°C. Hemolysis reached 12% at 500 µM, and human 293T cells were unaffected up to 50 µM but badly damaged at 500 µM; the work is in lab cultures only.