Predicting Incident Coronary Heart Disease Many Markers at a Time.

Abstract

After a decade of genome-wide association studies (GWASs), there are now thousands of common genetic variants associated with common clinic outcomes, subphenotypes, and human traits.1 As predicted from the common disease-common variant hypothesis, GWAS has established that each common genetic variant associated with a clinical outcome of interest confers only a small proportion of the overall risk attributable to genetics. From these findings, many studies have generated genetic risk scores, a summation across multiple loci, as an estimate of individual-level risk. The hope is that these scores can be used in clinical settings to target patients for prevention, intervention, and treatment strategies. Article, see p 531 There are multiple approaches to calculating genetic risk scores.2 The most straightforward calculation is the summation of the number of risk alleles a participant or patient has for a particular clinical outcome of interest. A common variation of this approach includes weighting the risk alleles by expected impact or effect size based on the previous studies. The genetic risk score is an appealing summation of risk, but there are many caveats and nuances related to its calculation, including the availability and quality of source data for the variants used in calculating the score, their associated effect sizes used in weighting, and their general sensitivity and specificity in predicting individual-level risk. With numerous established risk loci, coronary heart disease is an excellent example of how GWASs have transformed …