Back

Identification and evaluation of 41 risk loci for juvenile idiopathic arthritis informs precision medicine: mechanistic implications of DNA topology and an HLA-A*02:01-ERAP2 interaction

Ainsworth, H. C.; Khvatkova, E. S.; Sudman, M.; Marion, M. C.; Jiang, K.; Bohnsack, J. F.; Comeau, M. E.; Gaffney, P. M.; Haas, J.-P.; Howard, T. D.; Hyrich, K.; James, J. A.; Nigrovic, P. A.; Nordal, E. B.; Rosenberg, A. M.; Rygg, M.; Smith, S. L.; Videm, V.; Wedderburn, L. R.; Wise, C. A.; Yeung, R. S.; Thompson, S. D.; Schulert, G. S.; Prahalad, S.; Bowes, J.; Angeles-Han, S. T.; Jarvis, J. N.; Langefeld, C. D.

2026-09-17 genetic and genomic medicine
10.64898/2026.09.16.26363161 medRxiv
Show abstract

Juvenile idiopathic arthritis (JIA), the most common class of pediatric rheumatic diseases, can lead to joint damage and extra-articular features including uveitis. Early treatment can improve functional outcomes; but many patients do not respond to first-line treatments. We present a JIA Immunochip-based association analysis of rheumatoid factor-negative polyarticular JIA and oligoarticular JIA. This study spans 3,939 cases and 14,412 controls, including 1,123 cases and 1,356 controls not previously analyzed. We identify 41 (14 novel) JIA risk regions and report a JIA polygenic risk score that also associates with age of onset. Functional mechanisms presented include: a novel interaction between HLA-A*02:01 and ERAP2, cluster-based approach for functional amino acid discovery within HLA-DRB1, and sequence-dependent DNA topology for variant prioritization. We present a comprehensive list of curated gene-drug/small molecule interactions for clinically focused studies. These findings provide a deepened understanding of the genetics of JIA and potential mechanisms to inform precision medicine.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.