Back

Comprehensive Genetic Investigation Reveals Heterogeneous Pathways To Obstructive Sleep Apnea

Justice, A. E.; Keenan, B. T.; Chittoor, G.; Pahl, M. C.; Hoskens, H.; Josyula, N. S.; Gupta, S.; Kim, S.; Thorleifsson, G.; Manduchi, E.; Gutierrez, A.; Andrade Oliveira, F.; Ackert-Bicknell, C. L.; Benediktsdottir, B.; Boero, J.; Borthwick, K. M.; Chen, N.-h.; Cistulli, P. A.; Favazzo, L. J.; Gottlieb, D. J.; Hakonarson, H.; Kirchner, H. L.; Magalang, U. J.; Malow, B. A.; Mazzotti, D. R.; McArdle, N.; Mentch, F. D.; Morgentheler, T. I.; Penzel, T.; Pippin, J. A.; Rasmussen-Torvik, L. J.; Shin, C.; Singh, B.; Sofer, T.; Veatch, O. J.; Villani, D. A.; Wells, A. D.; Williams, M. S.; Zee, P.; Zu

2026-01-11 genetic and genomic medicine
10.64898/2026.01.08.26343696 medRxiv
Show abstract

Obstructive sleep apnea (OSA) is a common, heritable disorder with diverse clinical presentations and etiologies. We conducted a genome-wide association meta-analysis of 492,107 individuals, including 46,028 OSA cases, identifying 14 genome-wide significant loci, eight of which are novel. Analyses adjusting for body mass index (BMI) revealed three loci independent of obesity, implicating alternative biological pathways contributing to disease. Integrative functional analyses, including chromatin interaction mapping, fine-mapping, and eQTL colocalization, prioritized candidate effector genes. Notably, implicated genes in chondrocytes were associated with craniofacial morphology in both mouse and human multivariate genotype-phenotype mapping, supporting a role for craniofacial structure in OSA risk. These findings highlight key genetic pathways (obesity-related, neurological, and craniofacial) underlying the development of OSA, offering new insights into its complex etiology.

Matching journals

The top 4 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.