Lightnews — Scholar-powered news

Reposted by Q-Immuno lab at UCL

Joe Sabatino @jsabatino37.bsky.social · 5d

A really nice overview of key Treg discoveries relevant for today's exciting Nobel announcement
www.nobelprize.org/uploads/2025...

www.nobelprize.org

3 6

Q-Immuno lab at UCL @qimmuno.bsky.social · 14d

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

LDN-Q-Immuno · The Q-Immuno Lab

qimmuno.com

Q-Immuno lab at UCL @qimmuno.bsky.social · 14d

Join us for the next London Quantitative Immunology meet-up at the ICR in Chelsea on 19th November 2-5pm!

Free sign-up and short talk submission here: forms.gle/YZT1y2Sp6YVj...

Quantitative Immunology Nov 25'

Quantitative Immunology Meet up - November

forms.gle

1 1

Reposted by Q-Immuno lab at UCL

UCL Global Engagement @ucl-global.bsky.social · 19d

Driven to understand why transplants fail, Dr Andreas Tiffeau-Mayer (@qimmuno.bsky.social) used a UCL – University of Sydney Strategic Ignition Grant to combine computational and clinical expertise with collaborators in Australia ⬇️

Find out more about the Grants and apply by 13 Oct: bit.ly/4pFUd2S

Using machine learning to combat rejection in medical transplants

Dr Andreas Tiffeau-Mayer (UCL Division of Infection and Immunity) used the UCL – University of Sydney Strategic Ignition Grant to further knowledge about why some medical transplants fail.

www.ucl.ac.uk

1 2

Q-Immuno lab at UCL @qimmuno.bsky.social · Aug 19

How can repertoire analysis inform the selection of TCRs for cancer immunotherapy? We've summarised emerging experimental and computational approaches in a review led by @UDemael doi.org/10.3390/cell...

doi.org

2

Reposted by Q-Immuno lab at UCL

James Reading @jamesreading.bsky.social · Aug 5

🚨 PhD Position available in our lab 🚨 exploring the power of blood immune multi-omics to detect lung cancer years prior to clinical diagnosis in a unique cohort of >10,000 CT screened individuals.
✅ Wet & dry lab
✅ September 2025 enrolment
✅ UK tuition fees only

www.ucl.ac.uk/medical-scie...

Pre-Cancer Immunology

The Pre-Cancer Immunology Lab (James Reading Lab) is mapping pre-invasive T cell dynamics during carcinogenesis to detect and intercept cancer development.

www.ucl.ac.uk

15 16

Reposted by Q-Immuno lab at UCL

Trevor Graham @trevorgraham.bsky.social · May 22

Calling data scientists: we have opening in our Data Science core @icr.ac.uk to lead work around cancer spatial biology and single cell analysis. These are staff scientist type positions, ideal for someone who wants to help drive computational research. Apply here: jobs.icr.ac.uk/vacancies/12...

Data Scientist in Sutton | The Institute of Cancer Research

View details and apply for this Data Scientist vacancy in Sutton. Salary : Salary range £39,805 to £49,023 (depending on the experience) Reporting to: Professor Trevor Graham ...

jobs.icr.ac.uk

9 7

Q-Immuno lab at UCL @qimmuno.bsky.social · May 2

Very important work! We've been waiting for someone to do this. It has been clear for some time that a proportion of aggregated TCR-pMHC data might not be reliable. It will be interesting to reassess performance in light of this more robustly validated data.

bioRxiv Immunology @biorxiv-immuno.bsky.social · May 2

A functionally validated TCR-pMHC database for TCR specificity model development https://www.biorxiv.org/content/10.1101/2025.04.28.651095v1

1 5

Q-Immuno lab at UCL @qimmuno.bsky.social · May 1

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. 🎉

Q-Immuno lab at UCL @qimmuno.bsky.social · May 1

How to learn rules of TCR-pMHC binding that generalize? We propose that co-specificity rules are likely more universal and can thus be learned from less data.
arxiv.org/abs/2412.13722

Data-driven Discovery of Biophysical T Cell Receptor Co-specificity Rules

The biophysical interactions between the T cell receptor (TCR) and its ligands determine the specificity of the cellular immune response. However, the immense diversity of receptors and ligands has ma...

arxiv.org

1 6

Reposted by Q-Immuno lab at UCL

Mahdad Noursadeghi @mnoursad.bsky.social · Apr 21

Interested in T cell immunity in TB? Check out our new pre-print on antigen-agnostic identification of generalisable human in vivo Mtb-reactive T cell responses: www.biorxiv.org/content/10.1...
with @qimmuno.bsky.social & @innate2adaptive.bsky.social

Evolution of T cell responses in the tuberculin skin test reveals generalisable Mtb-reactive T cell metaclones.

T cells contribute to immune protection and pathogenesis in tuberculosis, but measurements of polyclonal responses have failed to resolve correlates of outcome. We report the first temporal…

www.biorxiv.org

1 8 13

Reposted by Q-Immuno lab at UCL

T cell Vaccines Lab (Swadling UCL) 🧬👩🏽‍🔬 @tcellvaccineucl.bsky.social · Apr 10

Hi, I’m Peter the Postdoc and today I’ll be taking over the feed from the 3rd London Quantitative Immunology Day! I’ll be sharing updates and insights as the day progresses, so stay tuned!

1 2

Reposted by Q-Immuno lab at UCL

T cell Vaccines Lab (Swadling UCL) 🧬👩🏽‍🔬 @tcellvaccineucl.bsky.social · Apr 10

(1/2)

Another successful #qimmuno day has drawn to a close, but discussions continue in the pub around the corner! Congrats to the organising committee @qimmuno.bsky.social for an excellent day and to all the speakers for engaging talks.

1 3

Q-Immuno lab at UCL @qimmuno.bsky.social · Apr 8

Two days to go until the 3rd LDN Q-Immuno Day at UCL! We are looking forward to welcome you on Thursday for an incredible line-up of speakers. Full schedule now available here: qimmuno.com/ldnday/

LDNQImmuno · April 10th

qimmuno.com

1 2

Q-Immuno lab at UCL @qimmuno.bsky.social · Mar 27

Two weeks to go for our London Quantitative Immunology day on April 10th! Last chance to submit abstracts for contributed talks is today 👉 qimmuno.com/ldnday/

LDNQImmuno · April 10th

qimmuno.com

3 2

Reposted by Q-Immuno lab at UCL

Damon May @the-dee-em.bsky.social · Mar 19

The recent "Mal-ID" paper in Science diagnoses diseases from immune repertoires. But a powerful model like Mal-ID can learn differences between groups of immune repertoires that are unrelated to the diseases of interest. We explore these pitfalls in our eLetter response.

Disease diagnostics using machine learning of B cell and T cell receptor sequences

Clinical diagnosis typically incorporates physical examination, patient history, various laboratory tests, and imaging studies but makes limited use of the human immune system’s own record of antigen ...

www.science.org

7 20

Q-Immuno lab at UCL @qimmuno.bsky.social · Mar 4

Come join us for the 3rd LDN Quantitative Immunology Day on 10th of April @ucl.ac.uk ! All info and free registration 👉 qimmuno.com/ldnday/

LDNQImmuno · April 10th

A community day for researchers in the quantitative life sciences from across London.

qimmuno.com

3 3

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 17

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.

3

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 17

3. We benchmarked the embeddings for single chain alpha or beta only data (thx reviewer 2!), and found that the chain dropping during pre-training makes single chain embeddings competitive with alignment as well!

1 2

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 17

2. We demonstrated that supervised contrastive learning reshapes the embedding space to make use of 'easy negatives'. This is one reason why discrimination between a fixed set of pMHCs is easier than completely blind prediction (as also seen in Immrep23).

1 3

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 17

Since the preprint, we made a number of additional discoveries:

1. We showed that contrastive training might enable SCEPTR to weight sequence similarity with respect to recombination biases (Pgen).

1

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 17

Give SCEPTR embeddings a try for your TCR analysis applications: sceptr.readthedocs.io/en/stable/

SCEPTR — sceptr 1.1.0-post1 documentation

sceptr.readthedocs.io

1

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 17

Our work on contrastive learning of T cell receptor representations is now out in Cell Systems!

www.cell.com/cell-systems...

Contrastive learning of T cell receptor representations

How should we train protein language models to predict protein function? This study shows that pre-training using an autocontrastive loss greatly improves transfer learning compared with masked-langua...

www.cell.com

1 2 11

Q-Immuno lab at UCL @qimmuno.bsky.social · Jan 7

Thanks for sharing Chris!

Reposted by Q-Immuno lab at UCL

Chris Thorpe @drchristhorpe.bsky.social · Jan 7

Two very exciting roles in this brilliant group. The deadline for one is next week, the other the end of January, so get applying and do some work that really matters with excellent people!

Q-Immuno lab at UCL @qimmuno.bsky.social · Dec 18

We are hiring! Please help us spread the word about this brilliant PhD opportunity in quantitative immunology!! 💥

qimmuno.com/openings/

1 1