BIDS Monthly Seminar

Modeling Immune Recognition with Language and Structural Models: New Directions in Computational Proteomics

Recent advances in computational proteomics are opening new avenues to understand how the immune system recognizes and responds to disease. María Rodríguez Martínez, PhD, will present her group's work on modeling immune receptor binding—specifically, how T cell receptors (TCRs) and B cell receptors (antibodies) engage with their targets.

She will discuss sequence-based approaches, including fine-tuning protein language models for calibrated receptor binding predictions and ensemble methods to identify autoreactive T cell receptors with applications in autoimmune disease.

She will then cover structure-based modeling, describing how her team combines structural predictions with graph-based methods to rank antibody binders, models TCR flexibility using generative AI trained on molecular dynamics simulations, and integrates confidence scores. with energetic and geometric criteria to assess interaction reliability.