AI-Driven Peptide Discovery Identifies Novel Antimicrobial Treatment to Restore Gut Barriers in Ulcerative Colitis

Jilin University researchers use machine learning to identify a novel peptide that outperforms standard drugs in treating ulcerative colitis in animal models.

By: AXL Media

Published: Mar 24, 2026, 5:20 AM EDT

Source: Information for this report was sourced from First Hospital of Jilin University

Accelerating Therapeutic Discovery Through Artificial Intelligence

Ulcerative colitis (UC) remains a challenging chronic condition characterized by persistent intestinal inflammation and a frequent lack of response to traditional 5-aminosalicylic acid or antibiotic treatments. While antimicrobial peptides (AMPs) have long been recognized for their natural immunomodulatory properties, the traditional process of screening these sequences is notoriously labor-intensive. To overcome this, researchers at the First Hospital of Jilin University developed a machine learning pipeline combined with genetic algorithms. This computational approach allowed the team to evaluate more than 6,000 potential candidates simultaneously, bypassing years of manual laboratory work to identify 22 high-priority sequences with the potential to treat inflammatory bowel disease.

The Selection of Lead Peptide "LR"

From the AI-generated shortlist, five peptides were synthesized for rigorous experimental validation. Among these, a sequence named "LR"—referencing its specific terminal residues—emerged as the superior candidate. In vitro testing confirmed that LR possesses a rare balance of high bactericidal activity against pathogens like Escherichia coli and Staphylococcus aureus while maintaining exceptionally low cytotoxicity. Unlike many traditional antibiotics that can damage host cells, LR demonstrated high biocompatibility and minimal hemolytic activity, suggesting it could safely navigate the human circulatory and digestive systems without causing secondary harm.

Superior Efficacy in Preclinical Colitis Models

The therapeutic potential of LR was tested using a dextran sulfate sodium (DSS)–induced mouse model, which mimics the mucosal damage seen in human ulcerative colitis. The results were striking: mice treated with the LR peptide showed significant improvements in body weight retention, disease activity index scores, and colon length preservation. Histological analysis further confirmed that LR treatment reduced mucosal damage and the infiltration of inflammatory cells. Notably, the peptide demonstrated stronger therapeutic effects than both the standard anti-inflammatory drug 5-aminosalicylic acid and the conventional antibiotic ciprofloxacin, marking it as a highly promising lead for future pharmaceutical development.