Genetic discovery using deep learning-derived optic nerve integrity phenotypes
- 1QIMR Berghofer
Abstract
The peripapillary retinal nerve fibre layer (pRNFL) thickness and Bruchs membrane opening-minimum rim width (BMO-MRW) are three-dimensional retinal biomarkers for glaucoma. We aimed to demonstrate that AI-derived thickness from two-dimensional fundus images can act as proxies, enabling the discovery of novel neurodegenerative loci. AI-derived pRNFL and BMO-MRW strongly correlated with OCT-derived thickness (r: 0.69 and 0.79, respectively). After validation, these phenotypes were predicted in two cohorts lacking disc-centred OCT: the UK Biobank and the Canadian Longitudinal Study on Aging. The predicted phenotypes showed strong genetic correlations with directly measured phenotypes from a previous study for both phenotypes (0.70 and 0.96, respectively). This data increased statistical power, identifying 29 loci for pRNFL thickness and 122 loci for BMO-MRW, including 14 loci that were independent of VCDR. We observed shared and sector-specific thickness loci overlapping glaucoma loci and revealed loci independent of known risk factors. Together, these results emphasise that multidimensional inferences can be drawn from 2D imaging, enabling downstream genetic analyses.
Related
- Field-of-view confounding shapes genetic discovery from self-supervised cardiac-imaging phenotypes
- Machine learning-derived Alzheimer's disease dimensions: neuroanatomical, cognitive, clinical, functional, and genetic risk heterogeneity signatures
- Neuroblastoma-derived Extracellular Vesicles Disrupt the Integrity of Central Nervous System Barriers via Tight Junction Modulation
- Shared genetic architecture between constipation and aging-related phenotypes : Insights from large-scale genome-wide cross-trait analysis
- Vitamin A and D Status, Selected Genetic Variants, and Allergy-Related Phenotypes: An Exploratory Retrospective Study
BibTeX
@misc{aman2026genetic,
title = {{Genetic discovery using deep learning-derived optic nerve integrity phenotypes}},
author = {A. M. Aman and E. Zaher and S. Diaz-Torres and S. J. Driessen and V. A. de Vries and F. C. van der Heide and A. Kolovos and J. M. Schmidt and H. N. Marshall and L. Saleh and A. Schulze and G. A. Blokland and C. A. B. Webers and C. J. H. van der Kallen and A. Wesselius and I. Arts and F. van Asten and M. Gorski and M. E. Zimmermann and K. J. Stark and I. M. Heid and T. L. Young and L. R. Pasquale and A. V. Segre and J. L. Wiggs and A. P. Khawaja and D. J. Zack and R. C. B. Wong and A. W. Hewitt and A. K. Schuster and T. T. J. M. Berendschot and A. A. H. J. Thiadens and K. A. van Garderen and C. C. W. Klaver and P. G. Hysi and C. J. Hammond and C. Brandl and J. Craig},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.08.26362398},
url = {https://www.medrxiv.org/content/10.64898/2026.09.08.26362398v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Genetic discovery using deep learning-derived optic nerve integrity phenotypes
AU - A. M. Aman
AU - E. Zaher
AU - S. Diaz-Torres
AU - S. J. Driessen
AU - V. A. de Vries
AU - F. C. van der Heide
AU - A. Kolovos
AU - J. M. Schmidt
AU - H. N. Marshall
AU - L. Saleh
AU - A. Schulze
AU - G. A. Blokland
AU - C. A. B. Webers
AU - C. J. H. van der Kallen
AU - A. Wesselius
AU - I. Arts
AU - F. van Asten
AU - M. Gorski
AU - M. E. Zimmermann
AU - K. J. Stark
AU - I. M. Heid
AU - T. L. Young
AU - L. R. Pasquale
AU - A. V. Segre
AU - J. L. Wiggs
AU - A. P. Khawaja
AU - D. J. Zack
AU - R. C. B. Wong
AU - A. W. Hewitt
AU - A. K. Schuster
AU - T. T. J. M. Berendschot
AU - A. A. H. J. Thiadens
AU - K. A. van Garderen
AU - C. C. W. Klaver
AU - P. G. Hysi
AU - C. J. Hammond
AU - C. Brandl
AU - J. Craig
PY - 2026
DA - 2026/09/08
PB - medRxiv
DO - 10.64898/2026.09.08.26362398
UR - https://www.medrxiv.org/content/10.64898/2026.09.08.26362398v1
AB - The peripapillary retinal nerve fibre layer (pRNFL) thickness and Bruchs membrane opening-minimum rim width (BMO-MRW) are three-dimensional retinal biomarkers for glaucoma. We aimed to demonstrate that AI-derived thickness from two-dimensional fundus images can act as proxies, enabling the discovery of novel neurodegenerative loci. AI-derived pRNFL and BMO-MRW strongly correlated with OCT-derived thickness (r: 0.69 and 0.79, respectively). After validation, these phenotypes were predicted in two cohorts lacking disc-centred OCT: the UK Biobank and the Canadian Longitudinal Study on Aging. The predicted phenotypes showed strong genetic correlations with directly measured phenotypes from a previous study for both phenotypes (0.70 and 0.96, respectively). This data increased statistical power, identifying 29 loci for pRNFL thickness and 122 loci for BMO-MRW, including 14 loci that were independent of VCDR. We observed shared and sector-specific thickness loci overlapping glaucoma loci and revealed loci independent of known risk factors. Together, these results emphasise that multidimensional inferences can be drawn from 2D imaging, enabling downstream genetic analyses.
ER -