Oculomics

Summary¶

Oculomics is an emerging clinical and research paradigm that utilizes the eye as a diagnostic window into systemic health. Because the retinal vasculature and neural layers are directly and noninvasively accessible using modern imaging modalities, deep learning-derived and expert-defined retinal phenotypes can serve as biomarkers for systemic disease. Oculomics research seeks to map associations between ophthalmic imaging features and cardiovascular, neurological, metabolic, and genomic traits to enable early disease prediction and personalized medicine.

Overview and Physiological Basis¶

The eye is uniquely positioned among human organs in that it permits direct, noninvasive visualization of both the vasculature (via color fundus photography, CFP) and the central nervous system (via optical coherence tomography, OCT). Physiological changes in these structures often mirror systemic pathology: - Cardiovascular Indicators: Changes in retinal vessel width, branching complexity, tortuosity, and choroidal vasculature reflect systemic cardiovascular conditions such as hypertension, coronary artery disease, and stroke. - Neurological Indicators: The retina is an embryological extension of the brain. Retinal nerve fiber layer (RNFL) thinning, ganglion cell layer loss, and photoreceptor (ellipsoid zone) alterations have been linked to neurodegenerative diseases, including Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis (ALS).

Technical Frameworks¶

To capture the breadth of oculo-systemic associations without relying solely on manual, expert-labeled features, recent research integrates machine learning and multi-omic pipelines: - Unsupervised Representation Learning: Tools like Ret-AAE compress high-resolution retinal scans into low-dimensional vector embeddings, preserving abstract structural features that correlate with health outcomes. - Automated Segmentation: Pipelines like AutoMorph enable automated quality control and segmentation of retinal structures (e.g., vasculature masking). - Multi-Omic Integration: Correlating retinal embeddings with genomic datasets (GWAS), circulating metabolites (e.g., HDL, LDL, and VLDL cholesterol), physiological measures, and brain/cardiac MRI features allows researchers to reconstruct the molecular and anatomical pathways linking the eye to systemic health.

Clinical Applications¶

Disease Risk Prediction: Oculomics enables the development of models to predict future onset of diseases such as ischemic heart disease, heart failure, vascular dementia, and Parkinson's disease.
Genomic Associations: GWAS of retinal phenotypes have identified associations with genes regulating pigmentation (TYR, OCA2, DCT, TSPAN10), retinal development (VSX2), visual cycle (RDH5), and vascular smooth muscle contraction (PDE3A).
Metabolic Markers: Retinal imaging features show strong association with lipid metabolism, identifying circulating lipoprotein composition as a key biological axis between ophthalmic structure and cardiometabolic health.

Birney Lab — computational genomics and cohort-scale phenotypic modeling.
Frangi Lab — medical image computing and computational medicine.
Sergouniotis Lab — ophthalmic genetics and genomic medicine.

Citations¶

Julian, T. H. et al. (2026). Multi-omic analysis of deep learning-derived phenotypes links ophthalmic imaging to cardiovascular and neurological traits. Nature Cardiovascular Research, 5, 541–554. DOI: 10.1038/s44161-026-00815-5. Source paper: s44161-026-00815-5.pdf