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

Liyong Wang

Third Committee Member

Margaret A. Pericak-Vance

Fourth Committee Member

David J. Lee


Age-related macular degeneration (AMD) is shaped by genetic and non-genetic factors. Examining such factors—along with interactions between these factors—that contribute to AMD is crucial because permanent vision loss can arise, and better understanding the disease pathogenesis will foster potential treatments. Consideration of gene-gene and gene-environment interactions may identify additional risk factors and explain some of the variability in disease presentation and progression. Several studies have revealed associations between AMD risk and mitochondrial (mt) variants [mt single nucleotide polymorphisms (SNPs)]; additionally, mtDNA damage or mt dysfunction in retinal pigment epithelium (RPE) cells can affect AMD-related processes. Overall diet quality has been inversely associated with AMD risk. This project focused on mtSNP-nuclear SNP (nSNP) interactions for advanced AMD risk and gene-dietary exposure interactions for time to progression to advanced AMD; both types of interaction are understudied in AMD. Thus, my principal research question was whether novel genetic loci for both outcomes could be identified when including either interactions with mtSNPs or interactions with dietary exposures in genetic analyses. Joint effects of mtSNPs and nSNPs were analyzed in a genome-wide association study (GWAS) of the large IAMDGC exome chip dataset (17,832 controls and 16,144 advanced AMD cases). Novel locus TRPM1 was identified with genome-wide significant joint effects (p<5.0x10^-8) of intronic TRPM1 nSNPs and AMD-associated non-synonymous MT-ND2 mtSNP A4917G. Stratified analysis by mt allele identified association only in 4917A (major allele) carriers who had decreased AMD risk. Intronic and intergenic ABHD2/RLBP1 nSNPs demonstrated both genome-wide significant joint effects and nominally statistically significant interaction effects (p<0.01) with MT-ND5 synonymous mtSNP G12771A. Although positive association was detected in both strata, association was stronger in 12771A subjects. These findings illustrate that the mt genome may aid in deciphering mechanisms of AMD, especially those contributing to vision loss. GWAS analysis of a second dataset with longitudinal examination data identified three novel loci associated with time to progression to advanced AMD. In a dataset of 397 intermediate AMD cases (606 eyes, 178 progressors), three novel loci were associated with a shorter time to progression in Cox proportional hazards (PH) models: RRAGC (chromosome 1), CPNE4/ACPP (chromosome 3), and LOC100506393/PDE3A (chromosome 12) [p<5x10^-8]. RRAGC, which encodes Ras-related GTP binding protein C (a GTPase), is a member of the mTOR pathway. This pathway is involved in neovascularization, which defines the most severe AMD phenotype subtype, and has not been a top AMD risk-associated pathway or implicated by known risk loci. Analysis of dietary exposures and progression to advanced AMD identified four dietary factors associated with time to progression. Nutrient intake was estimated using a food frequency questionnaire in 168 intermediate AMD cases at baseline (260 eyes, 50 progressors). From Cox PH analyses of 108 dietary exposures, four were statistically significant (p<0.05): total number of vegetable servings (longer time to progression), number of other vegetable servings (longer time to progression), delta-tocopherol (shorter time to progression), and fatty acid 22:1 (shorter time to progression). Association with vegetable servings has been noted for risk. Since vegetables are sources of antioxidants, a counter to oxidative stress, antioxidants may play both a role in progression and risk. Given significant SNPs in the three novel progression loci and a genetic risk score (GRS), I conducted analyses to determine if any dietary exposures’ effects were modified by these candidate SNPs and GRS. None of the interaction effects were nominally statistically significant (p<0.000116), indicating that other novel genetic loci may be involved in effect modification. Finally, a genome-wide analysis of joint genetic and dietary exposure effects identified a novel locus associated with progression time to advanced AMD. When considering interactions with the four significantly associated dietary factors in Cox PH models of 149 individuals (230 eyes, 48 progressors), an intronic ZNF286B SNP (chromosome 17) achieved genome-wide significance. Upon stratifying by exposure level, higher intake of vegetables was associated with a shorter time to progression while lower intake was associated with a longer time to progression. Examination of complex gene-gene and gene-diet interactions revealed six novel loci associated with either AMD progression or risk. This stresses the vitality of examining a complex disease’s multiple outcomes and including statistical interactions to better understand the biology and application of preventative strategies, especially regarding environmental exposures, such as diet.


age-related macular degeneration (AMD); progression rate; gene-gene interactions; gene-environment interactions; mitochondrial DNA; diet and nutrition