Safer pHLA-Targeted Immunotherapies Through AI and Computational Immunology
About the poster
Off-target immunotoxicity remains a critical challenge in developing peptide HLA (pHLA)-targeted immunotherapies. Pinpointing problematic epitopes through experiments is a time-consuming and costly effort.
We developed an AI-driven pipeline (Ardigen’s ARDiTox) that integrates computational immunology and structural modeling to assess off-target risks in silico. The methodology was applied to several published cases of immunotherapy-related off-target toxicity and to a therapeutic candidate. In each case, our solution flagged the known problematic epitopes among its top predictions, demonstrating accuracy in identifying and prioritizing high-risk epitopes.
This in silico approach provides an efficient way to anticipate off-target immunotoxicity and guide the design of safer pHLA-targeted therapies. By highlighting the most relevant risks, ARDiTox helps direct experimental validation and supports the development of next-generation immunotherapies.
This poster was originally presented during the Discovery on Target Conference in Boston, US and Festival of Biologics, Switzerland
