Detection of clonal hematopoiesis of indeterminate potential via genome or exome sequencing profoundly underestimates disease associations

Clonal hematopoiesis of indeterminate potential (CHIP) occurs when at least 4% of blood cells harbor somatic mutations in leukemogenic genes and is associated with increased risk for cardiovascular disease, malignancy, and mortality. While deep sequencing (>1,000x coverage) is the gold standard for CHIP detection, most large-scale studies rely on shallow genome/exome sequencing ([~]35x coverage) from biobanks. However, the sensitivity and specificity of genome-based CHIP detection remain unknown, raising concerns about the accuracy of reported disease associations. We performed both deep targeted sequencing and genome sequencing on identical DNA samples from 9,925 participants to characterize genome-based CHIP detection performance. Genome sequencing showed poor sensitivity (28%) and positive predictive value (44%) compared to deep sequencing, with performance highly dependent on clone size. Simulation studies revealed that these ascertainment errors dramatically reduce statistical power and underestimate true effect sizes by >80%. These findings indicate that genome-based studies profoundly underestimate CHIP-disease associations, necessitating targeted deep sequencing for accurate clinical risk assessment.