Publication Date



Open access

Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PHD)


Human Genetics and Genomics (Medicine)

Date of Defense


First Committee Member

William K. Scott

Second Committee Member

Margaret A. Pericak-Vance

Third Committee Member

Abigail S. Hackam

Fourth Committee Member

Jonathan L. Haines


AMD is the leading cause of irreversible vision loss in developed countries. Its multifactorial etiology includes both genetic and environmental risk factors. As much as 65% of the genetic contribution to AMD risk may be explained by known genes. The remaining genetic factors may be genes or pathways containing multiple variants with small effect, or variants that interact with the environment to modulate disease. However, gene-environment interactions have been relatively understudied in AMD, and there were no methods for gene or pathway analysis that incorporated environmental effects. Therefore, the overall goal of this project was to address this gap in knowledge and identify novel AMD loci with significant gene-environment interaction effects that were missed in previous studies of strong main effects only. I examined an existing imputed GWAS dataset of 1,207 white AMD cases and 686 unaffected white controls. First, I conducted a genome-wide screen for gene-environment interactions with two established AMD environmental risk factors, smoking and exogenous estrogen by using the two degree of freedom (2df) joint test of single nucleotide polymorphism (SNP) main effect and gene-environment interaction. There were no novel genome-wide significant variants associated with AMD with joint effects with smoking, hormone replacement therapy or birth control pills (BCP). However, I identified several interesting, suggestively significant genes for follow up analyses and genotyped additional variants within two candidate genes, but effects did not replicate in two independent cohorts. Next, I developed the set-based 2df joint test of genetic and environmental factors. This is the first pathway analysis method to account for gene-environmental interactions. It extended the single variant tests and analyzed the joint effect of genetic variants and environment exposures, averaged across genes or pathways. Simulations proved that my new method had the correct type 1 error rate and had higher power to detect associations when effect modification was present than analyzing genes or pathways with main genetic effects alone. Next, I tested the association of AMD with candidate pathways, while accounting for genetic and environmental effects. I found that the VEGF signaling pathway was not associated with AMD when main effects were analyzed alone. However, the pathway was associated with neovascular AMD in women when accounting for history of BCP use (P= 0.017), and six independent genes within VEGF’s Proliferation subpathway were responsible for the effect. Stratification by BCP use revealed that effects were isolated to women who had taken BCPs and were not detected while analyzing genetic effects alone. The novel genes identified in this pathway warrant closer examination for AMD association in the context of estrogen usage. These results illustrate that some potential AMD associations may only be revealed when considering more complex relationships among risk factors. This study also demonstrates the importance of incorporating environmental exposures in gene discovery efforts at the SNP, gene, and pathway level.


Age-related macular degeneration; computational genetics; genetic epidemiology; gene-environment interactions; GWAS; pathway analysis