Dissertations & Theses (Open Access)
Date of Award
8-2023
Degree Name
Doctor of Philosophy (PhD)
Advisor(s)
Goo Jun
Second Advisor
Han Chen
Third Advisor
Yun-Xin Fu, Craig L Hanis
Abstract
Worsening glycemia, prediabetes and diabetes, is one of the essential diseases in public health, considering their high prevalence, the enormous impact on multiple organs, and the economic burden on the community. Various factors can affect the pathogenesis of worsening glycemia, and this study focused on macronutrient intake, genes, and their interactions with metabolome data. 616 self-reported Mexican American participants in Starr County were recruited with informed consent. 308 identified and 2,471 unidentified metabolites were used for the analysis, and all the metabolites were inverse normalized with less than half of the missing rate. Each of the five glycemic and lipid traits was selected, and insulin and HOMA-IR were only log-transformed to correct skewed distribution. Macronutrient intake was calculated from 110-item food frequency questionnaires by the formula of nutrient density. All the analyses were adjusted for age, gender, and BMI as covariates. The analyses to find associations across glycemic and lipid traits, nutrients, and metabolites used linear regression Question Response models. We also compared the mean difference of metabolites across the glycemic status group with ANOVA model adjusted covariates. Genetic associations on the metabolites were calculated by GMMAT, and gene-environment interactions were investigated by MAGEE. 3- hydroxybutyric acid, CAR (5:1), DG (18:1_18:1), DG (32:0), DG (32:1), DG (34:1), DG (34:2), PC (32:1), and 9 unidentified metabolites were associated with macronutrient intake, glycemic traits, and lipid traits. 28 identified and 232 unidentified metabolites were associated with specific SNPs in the cutoff of 5.0E-08. Among the metabolites, DG (32:1) was associated with the SNPs located on the LRFN2 gene (top signal p-value 8.95E-09), and the other 16 metabolite-gene pairs were newly found. In the SNP-nutrient interactions, 13 SNP-nutrient interaction pairs on identified metabolites and 40 SNP- nutrient interaction pairs on unidentified metabolites were significant, but no significant SNPs overlapped compared to the GMMAT results. Moreover, more than half of the significant signals by MAGEE were located on noncoding DNA regions, so further study should be needed to reveal their functions.
Recommended Citation
Chung, Shinhye, "Diet By Genomic Interactions In Determining Metabolic Traits And Profiles On Glucose Metabolism Among Mexican Americans In Starr County, Texas" (2023). Dissertations & Theses (Open Access). 240.
https://digitalcommons.library.tmc.edu/uthsph_dissertsopen/240