Skip to navigation Skip to main content Skip to footer

Approved research

Genome-wide association study of Body mass index in short and tall individuals

Principal Investigator: Dr Bochao Lin
Approved Research ID: 44344
Approval date: January 23rd 2019

Lay summary

Body mass index (BMI) is defined as weight divided by the square of height (kg/m2), which has been a universal measure to classify obesity since 18th Century (Eknoyan 2008). Recent research shows a negative association between BMI and height, especially in women (Sperrin, Marshall et al. 2016). If equally fixed food portion sizes are served, shorter adults are at increased risk for obesity compared to taller adults, simply because their lean mass is lower and they require less energy for metabolism. When walking same distance, short people consume more calories per pound of body weight, because they have to take more steps than taller people (Weyand, Smith et al. 2010). Taller adults would be less susceptible because they require more energy intake for metabolism and their basal metabolic rate is higher. Thus, shorter adults are more susceptible to diabetes and cardiovascular disease, independent of BMI (Bosy-Westphal, Plachta-Danielzik et al. 2009). Therefore, height may be an important confounding factor when analyzing the genetic architecture of BMI. We hypothesized that the genetic architecture of BMI is different in short compared to tall individuals and propose a range of genetic analyses leveraging existing whole genome data to disentangle height-dependent differences in genetic determinants of BMI. First, a genome-wide association study of BMI and height-adjusted BMI will be conducted in the entire study population consisting of cohorts from the UK biobank and 23andme. Genetic overlap between BMI and height will be estimated in this entire cohort by GCTA (Genome-wide Complex Trait Analysis) (Lee, Yang et al. 2012). Secondly, to investigate SNP-based heritability and genetic architecture of BMI non-cofounded by height, GCTA and GWAS of BMI will be conducted in height-stratified sub populations. The association of BMI and environmental factors such as physical activity, 24 hours calorie intakes and diet will be tested in subpopulations separately. GWASs will be conducted in parallel for each data subset. Chicago plots which show the genetic hits for BMI compared with different height stratified groups will be generated. Finally, polygenetic risk scores (PRSs) for BMI and PRS scores for height will be calculated based on the GIANT consortium (Locke, Kahali et al. 2015) and another study (Wood, Esko et al. 2014). Association tests of BMI with PRSs will be conducted in the entire study. Summary statistics will be functionally annotated by FUMA and MAGMA, aiming to provide additional insight into functional and biological mechanisms.