Polygenic scores for obstructive sleep apnoea reveal pathways contributing to cardiovascular disease

Figures

• 

  Fig. 1. OSA PGS development and assessment. Development and assessment of OSA PGSs. The steps are composed of (a) PGS training using GWAS summary statistics, reference panels, and a separate population for computing PGS summation weights; (b) evaluation step used to select PGS out of multiple candidates; (c) validation of associations with OSA in new independent datasets; and (d) follow up analyses addressing OSA PGS associations with OSA within various strata, associations with related sleep phenotypes, comorbidities, and sequelae of OSA. Analyses in step (d) were in datasets from previous steps. OSA, obstructive sleep apnoea; PGS, polygenic score; GWAS, genome-wide association study.

• 

  Fig. 2. OSA PGS associations with OSA in TOPMed individuals. Panel a: distributions of BMIadjOSA- and BMIunadjOSA-PGSs in strata defined by self-reported race/ethnicity. Panel b: percentages and numbers of individuals with normal, mild, moderate, and severe OSA (defined by cut-points of REI/AHI of 5, 15, and 30), in quintiles of the OSA PGSs by self-reported race/ethnicity (Asian group excluded due to low sample size). Limited to individuals with measured REI/AHI. Panel c: estimated OSA PGS associations with OSA in TOPMed combined and stratified samples. Analyses were adjusted for age, sex (unless sex-stratified), self-reported race/ethnicity (unless stratified by that), and BMI (both linear and squared terms). AHI, apnoea-hypopnea index; OSA, obstructive sleep apnoea; PGS, polygenic score; REI, respiratory event index.

• 

  Fig. 3. Estimated associations of OSA PGSs with OSA- and sleep-related phenotypes in TOPMed. Results from association analyses of OSA PGSs with OSA-related sleep measures. Panel a: measures that tend to be higher with more severe OSA, panel b: measures that tend to be lower with more severe OSA. AHI, apnoea-hypopnea index; NREM, non-rapid eye movement sleep; REI, respiratory event index; REM, rapid eye movement sleep; Avg SatO2, average oxyhaemoglobin saturation during sleep; Min SatO2, minimum oxyhaemoglobin saturation during sleep; Avg DesatO2, average oxyhaemoglobin desaturation during respiratory events.

• 

  Fig. 4. OSA PGS associations with OSA in validation studies. Estimated associations of BMIadjOSA- and BMIunadjOSA-PGSs with OSA in three validation datasets: All of Us, Geinsinger health system's MyCode, and the Human Phenotype Project, in combined and stratified analyses. BMI, body mass index; OSA, obstructive sleep apnoea; PGS, polygenic score.

• 

  Fig. 5. OSA PGS associations with quantitative OSA phenotypes in the Human Phenotype Project. Estimated associations of BMIadjOSA-PGS and BMIunadjOSA-PGS with sleep monitoring, log-transformed OSA-related measures in the HPP, stratified by REM and NREM sleep and combined. All measures were averaged over the three nights of sleep monitoring. AHI, peripheral arterial tonometry-derived apnoea-hypopnea index; HPP, Human Phenotype Project; ODI, oxygen desaturation index; RDI, respiratory disturbance index; REM, rapid eye movement; NREM, non-REM.

• 

  Fig. 6. OSA PGS associations with other clinical outcomes in All of Us. Estimated associations of BMIadjOSA and BMIunadjOSA PGSs with clinical outcomes in the All of Us study. Associations were adjusted for age, sex, BMI (linear and squared terms), and 10 PCs of genetic ancestry. BMI, body mass index; COPD, chronic obstructive pulmonary disease; OSA, obstructive sleep apnoea; PGS, polygenic score; PC, principal component.

• 

  Fig. 7. Sex-stratified associations OSA PGSs with cardiometabolic traits and body fat distribution measures from DXA scan. Panel a: Estimated adjusted odds ratio of OSA PGSs with hypertension and T2D in All of Us. Panels b and c: estimated associations of BMIadjOSA-PGS with DXA measures in HPP. All associations were adjusted for age, sex, BMI (linear and squared terms), and the 10 first genetic principal components. Results from association analyses that used BMIunadjOSA-PGS and was not adjusted for BMI are provided in Supplementary Figure S18. BMI, body mass index; DXA, dual-energy X-ray absorptiometry; HPP, human phenotype project; OSA, obstructive sleep apnoea; PGS, polygenic score; OR, odds ratio; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue; FM, fat mass; TFM, total scan fat mass; T2D, type 2 diabetes.