Coronary risk scores stop giving one number: eight biological axes separate people with the same genetic liability
Synopsis
Polygenic risk for coronary artery disease no longer collapses into a single number. Conventional scores identify who is at high risk but not which biological processes drive that risk. Decomposing 526 genome-wide significant variants by their associations with 78 CAD-related traits yielded eight latent components: cholesterol-enriched lipoproteins, triglyceride-rich lipoproteins, lipoprotein(a), metabolically unhealthy normal weight, obesity, endothelial dysfunction, coagulation and cardiac remodeling. Process-specific scores built in UK Biobank, SCAPIS and Mass General Brigham Biobank tracked clinical coronary events, subclinical atherosclerosis and distinct biomarker and comorbidity profiles.
Interpretation
Coronary artery disease genetic risk can now be split by biological process instead of reported as one aggregate number. Earlier polygenic scores combined hundreds of variants into a single metric that flags high-risk individuals but does not say which processes make a given person susceptible. Latent component decomposition over 526 independent CAD risk variants and 78 traits converged on eight stable components across 50 independent runs, with 7 to 50 variants per component.
Each component becomes an individual-level process-specific polygenic score measuring a person's load on one genetic axis. Each score sums the component's variants weighted by their component-specific weights; correlations with the total CAD score are modest and correlations among the eight scores are low. Scores were constructed and tested in 458,905 UK Biobank participants, 24,904 in SCAPIS and 53,848 in Mass General Brigham Biobank.
The eight scores relate to clinical coronary events, subclinical plaque and distinct biomarker and comorbidity profiles. The total score gives only an overall association, whereas these scores show different directions and strengths of association. Comparing top with bottom decile, cardiac remodeling, lipoprotein(a) and metabolically unhealthy normal weight gave odds ratios of 1.84, 1.59 and 1.56, and coagulation the lowest at 1.15; directions were consistent in SCAPIS, which did not contribute to the decomposition.
Perspective
The framework resolves CAD genetic liability into eight biological axes and yields per-axis individual load scores in three large cohorts, for predominantly European-ancestry participants in deeply phenotyped and imaged settings. The authors propose testing in intervention studies targeting specific lipoprotein metabolism mechanisms whether these scores can guide treatment response.
Component detection and interpretation partly reflect the strength of available phenotype and cell-type data, and the decomposition considered only genome-wide significant variants, so a larger variant set might resolve additional mechanisms. Components were derived in predominantly European-ancestry participants; seven of eight scores reproduced in Asian-ancestry participants while the obesity score showed no association, and other groups may require ancestry-specific trait curation and decomposition. Interactions between risk factors and the scores were observational and confirmation will require repeated measurement or intervention trials.