I would either use min(4, 2-score) so you end up in the same 0-4 range.

Or leave it uncapped.

Both would conserve the C residue signal.

Hi @grst.bsky.social ,
Very exciting idea. We mainly used tcrBLOSUM with a standard alignment algorithm.
But you can adapt it for TCRdist in the same way as the original BLOSUM is used. 2 - #Value (as 2 is the max), and then setting the diagonal to 0. At least, that would be what I would do.

Excited to be at the ISLH conference today to talk about where we are with using TCR-epitope AI tools for hematology.

The preprint on IMMREP25 TCR-epitope prediction challenge is now out on preprint, which reveals that unseen epitopes remains an unsolved problem in the field!

doi.org/10.64898/202...

New White Paper: How Reliable Are Antigen-Specificity Annotations for T-Cell Receptors? ImmuneWatch white paper examining the reliability of computational TCR-antigen specificity annotations, covering database matching, seen epitope algorithms, and the validation behind DETECT.

Annotating novel T-cell receptors with their epitopes is something that I have put a lot of thought into the past decade. So I was excited to help actually frame an opinion on the topic in a white paper with ImmuneWatch:
www.immunewatch.com/news/white-p...

🚀 Registration is now OPEN for AIRR Community Meeting VIII: Decoding and Recoding Immunity! Join us June 8–11, 2026 at Yale University (USA) or attend virtually. 💸 Early-bird rates available until April 15! 🔗 Register: tinyurl.com/airrcmeeting8

#AIRRC8 #immunosky

Excited to have contributed to this review on how new technologies - including immunopeptidomics and TCR sequencing - might finally break the deadlock on a #Leishmania vaccine with Paul Kaye and Wim Adriaensen:

doi.org/10.1016/j.pt...

Multiomics and TCR Sequencing in Immunotherapy | Repertoire Room Ep. 5 YouTube video by ImmuneWatch

For the latest episode of The Repertoire Room Podcast, I sat down with Samra Turajlic after her talk at The Festival of Genomics and Biodata.

"We could only understand this because we were able to track the clones through TCR sequencing."

Check it out on YouTube:
www.youtube.com/watch?v=QnRd...

How can #AI help decode our #immune systems and help for next generation #diagnostics? Here is a technical overview from the @airr-community.bsky.social diagnostics working group on where we currently are on this problem: doi.org/10.1016/j.im...

Very proud to have been a jury member at the newly revived @iscb.bsky.social RSG Belgium hackathon! Exciting to see so many people giving their best efforts tackling challenging #bioinformatics problems!

Challenges and Future Directions of AIRR-seq-Based Diagnostics Adaptive Immune Receptor Repertoire sequencing (AIRR-seq) is a promising diagnostic method across various clinical conditions, yet its widespread impl…

The AIRR Diagnostics WG commentary is now available, wherein our members discuss both the challenges of current #AIRR-based #diagnostics, as well as the future directions that we see the field going into!
www.sciencedirect.com/science/arti...

That's it. #ATCR25 is over!
Thanks again to all attendees, speakers and support staff for making this an amazing meeting!

Last #ATCR25 speaker: Paul Thomas on deciphering the T cell receptor recognition code.

Fifth speaker on #ATCR25 day 2: Michael Birnbaum on high throughput TCR analysis and generation.

Fifth speaker, #ATCR25 day 2: Encarnita Mariotti-Ferrandiz on TCR repertoire signatures in auto immune and infectious disease.

Fourth speaker, #ATCR25 day 2: Hashem Koohy on decoding antigen-specific T cell recognition.

Third speaker, #ATCR25 day 2, Soumya Raychaudhuri on how the HLA affects the TCR repertoire, and how the TCR affects cell fate.

MHC peptide predictions are accurate but fail to identify the few immunodominant epitopes within the larger list of binders.

There many epitopes that could induce a strong T cell response, but do not get processed from their protein and do not get presented.

Second speaker, #ATCR25 day 2: Antonio Lanzavecchia on CD8 T cell immunodominance

Immunogenecity might not be the best sole criteria for the basis of vaccines, epitopes need to be protective too. This is hard but doable to figure out.

Organoids might be a way forward, but not the answer to everything (but could be the answer to a lot of things)

The spleen harbor a collection of different immune cells, and contain memory of infection and vaccination.

Next topic: two classes of regulatory T cells: CD4 Foxp3 T cells regulating B cells and a CD8 regulatory subset using granzyme B regulating T cell auto reactivity.

Exposure to pathogens (Mtb, CMV,...) remodels our immune system in a way that benefits them (and may hurt us)

Within TCR repertoires are motif clusters linked to disease progression, suggesting the existence of 'distractopes', epitopes that the disease wants you to react to, as a 'smoke screen' to redirect or exhaust the immune system.

Investigating the way that Mtb evades the immune system lead to the innovation of looking at the TCRs from tetramer sorting experiments.

Second topic: failures in vaccine trials. An infectious disease only becomes 'famous' when it learns how to evade our immune system.

First topic: how do human immune systems compare to inbred mice? Mice are the 'starter car', simple but lacking complexity, leading to failure in mice models.

First speaker, second #ATCR25 day: Mark Davis on analyzing the T cell response.