Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Feb 6;10(2):645-664.
doi: 10.1534/g3.119.400910.

A Comprehensive Genome-Wide and Phenome-Wide Examination of BMI and Obesity in a Northern Nevadan Cohort

Affiliations
Free PMC article

A Comprehensive Genome-Wide and Phenome-Wide Examination of BMI and Obesity in a Northern Nevadan Cohort

Karen A Schlauch et al. G3 (Bethesda). .
Free PMC article

Abstract

The aggregation of Electronic Health Records (EHR) and personalized genetics leads to powerful discoveries relevant to population health. Here we perform genome-wide association studies (GWAS) and accompanying phenome-wide association studies (PheWAS) to validate phenotype-genotype associations of BMI, and to a greater extent, severe Class 2 obesity, using comprehensive diagnostic and clinical data from the EHR database of our cohort. Three GWASs of 500,000 variants on the Illumina platform of 6,645 Healthy Nevada participants identified several published and novel variants that affect BMI and obesity. Each GWAS was followed with two independent PheWASs to examine associations between extensive phenotypes (incidence of diagnoses, condition, or disease), significant SNPs, BMI, and incidence of extreme obesity. The first GWAS examines associations with BMI in a cohort with no type 2 diabetics, focusing exclusively on BMI. The second GWAS examines associations with BMI in a cohort that includes type 2 diabetics. In the second GWAS, type 2 diabetes is a comorbidity, and thus becomes a covariate in the statistical model. The intersection of significant variants of these two studies is surprising. The third GWAS is a case vs. control study, with cases defined as extremely obese (Class 2 or 3 obesity), and controls defined as participants with BMI between 18.5 and 25. This last GWAS identifies strong associations with extreme obesity, including established variants in the FTO and NEGR1 genes, as well as loci not yet linked to obesity. The PheWASs validate published associations between BMI and extreme obesity and incidence of specific diagnoses and conditions, yet also highlight novel links. This study emphasizes the importance of our extensive longitudinal EHR database to validate known associations and identify putative novel links with BMI and obesity.

Keywords: BMI; GWAS; Obesity; PheWAS.

Figures

Figure 1
Figure 1
Manhattan plot of GWAS results of BMI including DM2-diagnosed individuals. This study includes DM2-diagnosed individuals and the statistical model includes DM2 as a bimodal covariate. The x-axis represents the genomic position of 500,508 SNPs. The y-axis represents -log10-transformed raw p-values of each genotypic association. The red horizontal line indicates the significance level 1x10-5.
Figure 2
Figure 2
Obesity Case-Control GWAS Manhattan Plot. This cohort includes DM2-diagnosed individuals. The x-axis represents the genomic position of 500,508 SNPs. The y-axis represents -log10-transformed raw p-values of each genotypic association. The red horizontal line indicates the significance level 1x10-5.
Figure 3
Figure 3
PheWAS results between BMI-significant SNPs and EHR Phenotypes. This figure shows the results of individual logistic regressions between incidence of 633 phenotype groups (phecodes) and the genotypes of 27 SNPs found to have statistically significant associations with BMI in a cohort with DM2 patients. Each point represents the p-value of one SNP and one of 633 phecodes with at least 20 cases assigned to it. The horizontal red line represents the significance level p=1.02x10-4, and the blue line represents the Bonferroni correction of p=3x10-6.
Figure 4
Figure 4
PheWAS results between BMI and EHR Phenotypes. This figure illustrates the results of individual linear regression between incidence of phenotype groups (phecodes) and the continuous BMI metric of all 6,645 individuals. Each of the 301 points represents the p-value of the association between one of 1,523 phecodes with at least 20 cases assigned to it, and BMI. Statistical significance was assessed by using the False Discovery Rate of 0.1, corresponding to a raw p-value of 1.96x10-2. Only associations with p < 1x10-15 are annotated for ease of viewing, represented by a horizontal line at 15 on the y-axis.
Figure 5
Figure 5
PheWAS results between obesity-significant SNPs and EHR Phenotypes. This figure presents results of individual logistic regressions between incidence of 372 phenotype groups (phecodes) and the genotypes of 34 SNPs found to be associated with extreme obesity. Each point represents the p-value of one SNP and one of 372 phecodes with at least 20 cases assigned to it. The horizontal red line represents the significance level p=3.85x10-4, and the blue line represents the Bonferroni correction of p=4x10-6.
Figure 6
Figure 6
PheWAS results between extreme obesity and EHR Phenotypes. This figure illustrates the results of individual linear regression between incidence of phenotype groups (phecodes) and the incidence of extreme obesity in 2,996 individuals. Each of the 191 points represents the p-value of the association between one of 1,362 phecodes with at least 20 cases assigned to it, and extreme obesity. Statistical significance was assessed by using the False Discovery Rate of 0.1, corresponding to a raw p-value of 1.4x10-2. Only associations with p < 1x10-15 are annotated for ease of viewing, represented by a horizontal line at 15 on the y-axis.

Similar articles

See all similar articles

References

    1. Alberti K. G. M. M., Eckel R. H., Grundy S. M., Zimmet P. Z., Cleeman J. I. et al. , 2009. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 1640–1645. 10.1161/CIRCULATIONAHA.109.192644 - DOI - PubMed
    1. Alexander J. K., Amad K. H., and Cole V. W., 1962. Observations on some clinical features of extreme obesity, with particular reference to cardiorespiratory effects. Am. J. Med. 32: 512–524. 10.1016/0002-9343(62)90052-9 - DOI
    1. An Y., Furber K. L., and Ji S., 2016. Pseudogenes regulate parental gene expression viaceRNA network. J. Cell. Mol. Med. 21: 185–192. 10.1111/jcmm.12952 - DOI - PMC - PubMed
    1. Anderson C. A., Pettersson F. H., Clarke G. M., Cardon L. R., Morris A. P. et al. , 2010. Data quality control in genetic case-control association studies. Nat. Protoc. 5: 1564–1573. 10.1038/nprot.2010.116 - DOI - PMC - PubMed
    1. Bañez L. L., Hamilton R. J., Partin A. W., Vollmer R. T., Sun L. et al. , 2007. Obesity-Related Plasma Hemodilution and PSA Concentration Among Men With Prostate Cancer. JAMA 298: 2275–2280. 10.1001/jama.298.19.2275 - DOI - PubMed

Publication types

LinkOut - more resources

Feedback