Multilocus genetic associations with obesity outcomes in Hispanic and non-Hispanic whites using a principal components regression approach: the San Luis Valley Diabetes Study
Date
2013
Authors
Booth-Kalajian, Andrea D., author
Nelson, Tracy, advisor
Melby, Chris, advisor
Hokanson, John, committee member
Peel, Jennifer, committee member
Bellows, Laura, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Introduction: The overweight and obesity crisis in America has reached alarming rates with little progress of reversing the trend, despite much effort. Heritability has been estimated at up to 70%, though it is still unclear how genetics respond to environmental pressures. Evaluating groups of genes that are known to influence metabolic pathways has given some insight into the variation we see in body composition and prevalence of metabolic diseases. Methods: Data from the San Luis Valley Diabetes Study's third examination were utilized (1997-1998, n=837). One hundred seven single nucleotide polymorphisms (SNP) were selected from 22 genes that have previously been associated with obesity and type 2 diabetes (T2D) in a cohort of Hispanic and non-Hispanic white (NHW) individuals. Genetic data were reduced to a smaller set of derived factors using principal components analysis (PC). Associations were determined between factors and obesity outcomes. Results: Hispanics were more likely to have T2D than NHW (19% vs. 11%). Sample minor allele frequencies for 100 analyzable SNPs varied between the two groups with the minor alleles of rs8059937 (A2BP1) and rs6822807 (UCP1) being significantly more prevalent in Hispanics and rs11724758 (FABP2) and rs2239179 (VDR) significantly more prevalent in NHWs. SNP variance was redistributed into orthogonal components and 32 were retained for analysis, accounting for 77% of the total variance in genetic data. The combined genetic information increased predictive power of increases in body mass index (BMI) from the study baseline by 5.6% in Hispanics. Genetic data increased predictability of BMI and waist circumference (WC) in NHWs by 7.5% and 5.1%, respectively. Both groups had a significant increase in knowledge gained (18%) for the prevalence of T2D when genetic information was added to the base model. SNPs from UCP1 loaded strongly onto PC4, which was associated with BMI change in Hispanics and BMI, WC, and T2D in NHWs. PC7 represented SNPs from RBP4 and FABP2, which was associated with diabetes status in both groups. All obesity outcomes were associated with PC15 in Hispanics, symbolizing SNPs on the PPARD and RBP4 genes. NHWs showed additional associations with components having strong loadings from SNPs on multiple genes, including ADIPOQ, GC, VDR, PPARG, PPARGC1A, PPARD, UCP2, UCP3, and AlOX15. Conclusions: When combined together, multilocus genetic data show a larger influence on obesity outcomes than single polymorphisms alone. Base variations in UCP1, RBP4, and FABP2 gene sequences are associated with change in BMI and diabetes status in Hispanics and BMI, WC, and diabetes status in NHWs.
Description
Rights Access
Subject
principal components
obesity
type 2 diabetes
genetics
Hispanics