Episode 4: Prof David Gfeller:Beyond Black Boxes: How Interpretable AI Is Decoding T Cell Recognition

Artificial intelligence has dramatically improved our ability to predict T cell recognition—but can it explain why T cells recognise specific targets?

In this episode of Immunity by Design, Prof Hashem Koohy speaks with Prof David Gfeller about a new generation of interpretable AI models that reveal the biological rules governing T cell recognition, rather than simply making predictions.

The discussion explores three landmark studies introducing TCR Specificity Profiles (TSPs), TEMPO, and new evidence that challenges long-held assumptions about TCR sequence prediction, alpha-beta chain pairing, and cross-reactivity.

Together they discuss how interpretable machine learning is reshaping computational immunology, improving our understanding of immune recognition, and accelerating the development of safer T cell therapies.