Abstract MP20: Genome Wide Association Study of Early Childhood BMI in Ancestrally Diverse Populations Identifies Novel Loci

Introduction: Over the past 30 years, obesity prevalence has markedly increased in the United States, including tripling among children. Although numerous obesity genetic loci have been identified by genome-wide association studies (GWAS) in adults, less is known about the genetic architecture of early childhood obesity. Moreover, most childhood obesity GWAS have been restricted to severe obesity and primarily European ancestry populations. Hypothesis: We hypothesized that meta-analysis of GWAS of early childhood standardized body mass index z-scores (BMIz) from the ancestrally diverse ZOE 2.0 cohort, the Santiago Longitudinal Study (SLS), and two primarily European ancestry studies (Early Growth Genetics [EGG] Consortium and the Norwegian Mother and Child Cohort [MoBa]) would identify novel early childhood BMI genetic loci. Methods: To identify genetic loci associated with early childhood BMI, we performed GWAS of BMIz in the ZOE 2.0 cohort, a community-based sample of children enrolled in public preschools in North Carolina (total n = 6,054 children, mean age 4 years, 48% African American [n=2891], 20% Hispanic [n=1214], 18% White [n=1066], 2% American Indian/Alaska Native [n=146], 12% other or more than one race [n=737]), and the SLS, a longitudinal cohort of children recruited from community clinics in Santiago, Chile (n = 861 children, mean age 5.5 years), using SAIGE and SUGEN, respectively. Genetic data were imputed to the TOPMed Freeze 8 (ZOE 2.0) and 1000 Genomes Phase III admixed American (SLS) reference panels. GWAS models were adjusted for self-reported race/ethnicity and 8 ancestry principal components (PCs) (ZOE 2.0), and 5 PCs (SLS). We performed inverse variance weighted fixed-effect meta-analysis of these results with previously published summary statistics of BMIz of children in the EGG Consortium (ages 2-10 years, all European participants) and the MoBa cohort (age 3 years, all European participants), for a total N of 63,747 individuals. Variants were filtered for minor allele frequency (MAF) > 0.01 and effective N > 20. Results: We identified 3 genome-wide significant ( p < 5 x 10 -8 ) loci ( PTBP2, LOC374295, DTWD2 on chromosomes 1, 2, and 5) that have not been previously associated with childhood obesity traits, although PTBP2 has also been associated with adult BMI. We also identified 3 genome-wide significant loci previously associated with body size at younger ages: birth weight ( LCORL , HMGA2 , chromosomes 4 and 12) or BMI at less than 8 months ( SH3GL3, chromosome 15). LCORL and HMGA2 have also been associated with adult obesity traits. Conclusions: Our findings are consistent with previous literature reporting age-specific genetic effects across early childhood as well as some shared genetic architecture with adult BMI. We are currently pursuing replication of novel findings in diverse populations, and mechanistically validating our novel signals.