± HEALTH Colorectal Cancer (CRC) is a common cause of cancer death and prognostic factors are used to determine manage…+ HEALTH This multicenter study developed and validated an interpretable machine learning model integrating granular n…+ HEALTH Multiple stakeholders need to locate results of registered clinical trials but frequently struggle to find th…+ HEALTH Accurate preoperative glioma grading and molecular subtyping are important for treatment. The vascular microe…+ HEALTH Gastrointestinal endoscopy generates extensive high-resolution video data, posing significant challenges for…+ HEALTH Non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) and invasive encapsulat…+ POLICY Artificial intelligence (AI) has the potential to strengthen real-world evidence (RWE) for regulatory decisio…± HEALTH Echocardiography remains the cornerstone of cardiovascular imaging. However, traditional workflows including…
TruaceTracing the truth around AIFriday, August 28, 2026
Subscribe
Health·The Trace·Dual reading·Published 2026-08-28

machine-learning classification of celiac disease status from naive adaptive immune receptor repertoires

Source article: Naïve adaptive immune receptor repertoires in celiac disease assessed by machine learning; impact of the HLA-DQ2.5 allotype on the TCR repertoire

Abstract: The adaptive immune receptor repertoire (AIRR) - the collection of an individual's B-cell and T-cell receptors (BCRs and TCRs, respectively) - encodes cumulative immune history and is shaped by both germline genetics and environmental exposures. Skewed repertoires have been linked to infections, vaccination responses and autoimmune diseases such as celiac disease (CeD) where biased usage of immunoglobulin and T-cell receptor genes reactive to disease relevant antigens has been reported. Motivated by evidence tha…

TRV-2026-0912Peer-reviewedPermanent record — cite & verify
Trace impact reading

Contested: both sides are scored from claims and sources, not community votes.

P 68The P score combines the specificity and measured human impact of the grounded problem claim with the strength of this Trace’s cited sources.G 68The G score combines the specificity and measured human impact of the grounded gain claim with the strength of this Trace’s cited sources.
Naïve adaptive immune receptor repertoires in celiac disease assessed by machine learning; impact of the HLA-DQ2.5 allotype on the TCR repertoire

Cross-reactivity hypothesis for the onset of dermatitis herpetiformis in patients with celiac disease by LBNBM OBIBT. CC BY-SA 4.0 · https://creativecommons.org/licenses/by-sa/4.0

The quick read

The study used machine learning on naive CD4+ TCR and naive BCR AIRR-seq repertoires to test classification of celiac disease versus controls, building on prior work linking germline HLA variation to naive repertoire composition and earlier BCR-based classification attempts.

It matters because the observed moderate TCR-based classification was not a direct disease signature but was driven by HLA-DQ2.5 frequency differences, raising uncertainty about whether naive repertoires alone contain independent celiac biomarkers and highlighting the need to control for germline HLA effects in future diagnostic models.

Main points
  • Study applied machine learning to naive CD4+ TCR and naive BCR AIRR-seq repertoires to distinguish celiac disease from controls.
  • TCR variable gene frequencies predicted HLA-DQ2.5 allotype with high accuracy.
  • Controlling for HLA-DQ2.5 abolished TCR-based diagnosis classification.
  • Naive BCR repertoires could not be used to classify celiac disease in this analysis.
Gain

Naive CD4+ TCR repertoire features enabled moderate machine-learning classification of celiac disease status and high-accuracy prediction of HLA-DQ2.5 status by publication date.

Problem

After accounting for HLA-DQ2.5 enrichment, machine-learning classification of celiac disease from naive TCR repertoires was abolished, and naive BCR repertoires failed to classify disease.

The rundown

Researchers analyzed naive CD4+ TCR and naive BCR AIRR-seq data with machine learning to test whether repertoire features distinguish celiac disease subjects from controls and to identify drivers.

Results showed moderate TCR-based diagnosis classification that tracked HLA-DQ2.5 enrichment, with TCR V-gene frequencies predicting HLA-DQ2.5 status, while controlling for HLA removed the diagnostic signal and BCR repertoires showed no classification ability.

What this doesn’t fix

TCR-based celiac classification was confounded by germline HLA variation, limiting direct disease signal from naive repertoires.

Sources

Reader signal

How should this claim be treated?

The debate