Note: Single-source report; awaiting corroboration.
Antibiotic resistance is a major public health challenge as many bacterial infections become harder to treat with current medications. Scientists are studying peptide molecules naturally produced by cells that kill bacteria by damaging their membranes, but selecting the most effective peptides is difficult. To address this, an NIH-funded team led by Drs. Jacob Gardner and César de la Fuente developed an artificial intelligence (AI) tool called ApexGo to design improved antibacterial peptides for laboratory testing.
ApexGo builds on a previous system, APEX, which identifies potential antibiotic peptides in large biological datasets. The new tool suggests modifications to existing peptides to enhance their bacteria-killing abilities and refines its predictions as more data is incorporated. The researchers tested ApexGo by optimizing 10 peptides derived from extinct organisms, generating 10 optimized versions for each original peptide.
Experimental results showed that 86 out of the 100 optimized peptides could kill at least one type of bacteria. Additionally, 68% of the optimized peptides showed improved antibacterial activity compared to their originals. Selected optimized peptides were also tested in mice infected with an antibiotic-resistant bacterial strain, demonstrating greater effectiveness against the infection.
This AI-driven approach may enable faster, more efficient development of new antibiotics to address bacterial resistance. The study was published in Nature Machine Intelligence on May 13, 2026.