| dc.description.abstract | Polygenic scores (PGS) aim to capture a person’s genetic predisposition to a disease or trait. There is substantial interest in applying PGS to clinical settings to identify patients at elevated risk for disease. Given its high heritability, height is a particularly promising trait for the study of clinical applications of PGS. We first performed a cross-ancestry validation of a recently developed height PGS in adult (N=61,701) and pediatric (N=9,088) cohorts. We show that variance explained by the PGS decreases when applied in non-European compared to European ancestries. Variance explained also decreases when applied in children compared to adults, particularly during onset of the pubertal growth spurt. We then show that the PGS associates with having short stature and clinical utilization related to short stature, including ICD codes for short stature and referrals to endocrinology. We then curated a cohort of pediatric patients evaluated by endocrinology for short stature. This cohort included 225 patients with short stature secondary to pathology and 114 patients without an explanation for their short stature (i.e., idiopathic short stature) even after extensive testing. We show that a lower height PGS is associated with idiopathic short stature, while the current standard for height prediction, mid-parental height, is not. We show that a PGS improves discrimination between idiopathic and pathologic short stature when added to a baseline model. In summary, a PGS associates with height and short stature-related clinical outcomes and discriminates idiopathic from pathologic short stature. However, portability between ancestries remains a barrier to implementation. | |
