Abstract P070: Comparison of Heritability Estimation Using the Principal Components Calculated by Maximizing Variance and Heritability for OSA Traits

Abstract

RATIONALE: Obstructive sleep apnea (OSA) is a common disorder characterized by repetitive pauses in breathing during sleep associated with significant cardiovascular morbidity. Accumulating study results suggest that there are strong genetic contributions for this disease as defined by the apnea hypopnea index (AHI), the number of breathing pauses per hour of sleep. Using this metric, it has been shown that OSA significantly aggregates within families (1-2); heritability for the AHI is estimated to vary from about 20 to 40% (2-3). However, the AHI is only one index from several that describe OSA severity, and may not be the most relevant phenotype for genetic studies. It has not, however, been clear how to systematically assess heritability of multiple correlated and uncorrelated OSA traits. METHODS: To identify OSA related phenotypes that are most heritable to inform discovery of genetic risk variants for the disease, heritability analysis of multiple OSA related phenotypes were conducted using family data from Cleveland Family Study which comprises 700 African Americans from 147 families and 669 European Americans from 139 families. Principal-components analysis was performed using two approaches for combining data for six OSA traits that describe several dimensions of physiological and clinical impairment: the AHI and three other sleep study metrics (average hypopnea duration, average oxygen saturation, and percent time at oxygen saturation of < 90%) and two OSA symptoms (habitual snoring and excessive sleepiness). One method uses the traditional principal-components (PC) approach assuming all subjects are independent. The other developed by Rabinowitz and Ott (4) calculates principal components by maximizing the heritability. RESULTS: In European Americans, the maximum heritability for the traditional multi-trait PC was 0.513 (SE=0.1010) while it was 0.512 (SE=0.100) for the PC calculated maximizing heritability. The maximum individual trait heritability was for average oxygen saturation, which is 0.388 (SE=0.107). In African Americans, the maximum heritability of traditional multi-trait 0.50 (SE=0.094) while it was 0.616 (SE=0.091) for the PC calculated maximizing heritability. The maximum individual trait heritability was for average hypopnea duration, which is 0.570 (SE=0.096). CONCLUSIONS: In general, both principal components approaches (maximizing variance and heritability) of six OSA related traits result higher heritability estimates than individual trait heritability estimates. The principal components approach based on maximizing heritability has the potential to improve trait heritability and therefore can be useful in future association analysis for searching genes underlying obstructive sleep apnea.