Looking forward to connecting with many of you!

Later on I will give an invited talk in the session "AI driven insights into immune responses: vaccines, infections, and disease" laying out a vision for combining AI with TCRseq to study human immunity.

Carolin Turner and Peter Thomas will talk in the Decoding and harnessing diversity in adaptive immune receptor repertoires session on T cell immunity to TB and SARS-CoV-2, respectively

To keep up-to-date with future events, also consider signing up for the LDN Q-Immuno mailing list. Instructions here: qimmuno.com/ldnqimmuno/

In our updated preprint, we show that mlTCRdist weights are predictive for pMHCs highly dissimilar (≥ 6 edits) to any seen during training! We are excited about these findings as such wide extrapolation has so far remained out of reach for sequence-based machine learning approaches. 🎉

4. Finally, we used our information theory framework to explain why predictions for some epitopes are easier regardless of model and we introduce a simple technique to provide error bars for comparing model performance which removes the variance resulting from this effect.