Andrew DeWan, PhD, MPH

My overarching research mission is to understand how variation in the human genome contributes to complex human diseases. My work is centered around a diverse set of complex traits using a strategy of narrowly defined phenotypes and stratification by ancestry to reduce heterogeneity and increase statistical power as well as extending analytical methods to look at genetic interactions and pleiotropy.

My research leverages data from high-throughput genotyping and sequencing studies (both primary and secondary data analysis) to identify genetic variants associated with the phenotype of interest. Using a highly homogeneous set of cases and controls, my work identified an HTRA1promoter variant significantly associated with the wet form of age-related macular degeneration. Additional approaches to reducing phenotypic heterogeneity have led to mapping a suggestive association with a genetic marker in PDE11A for children with atopic asthma, a marker on chromosome 17q21.2 that significantly interacts with asthma to increase body mass index (BMI), and variants in TCFL2 and ERBB4 associated with BMI in Hispanics and African Americans, respectively. My work has recently expanded to include childhood cancer, with a leading role in a GWAS for childhood acute lymphoblastic leukemia (ALL) that identified two novel loci and included the largest sample of Hispanic cases to date. I am now leveraging this data to study genetic variants in microRNAs that are associated with ALL (funded through my R03) as well as a better understanding of the genetic risk factors for ALL in Hispanics, Asians and African Americans.

Building on the success of exome sequencing a single family with multiple asthmatics, I obtained R01 funding to recruit and conduct exome sequencing of approximately 250 nuclear families with multiple children with asthma with the goal of identify rare variants segregating with asthma. A family-based strategy for mapping rare variants has numerous power advantages over a population-based design including a potential for higher genetic load within families with multiple affected subjects, avoiding population stratification and reducing environmental effects through shared living. This represents a paradigm shift in association mapping, as this technique has been largely limited to unrelated case-control studies. Beyond the immediate scope of this project, the data and biological samples that were collected will be able to be utilized in future studies of asthma, including studies to look at the involvement of microRNAs in asthma and rare variants contributing to virus-induced asthma (F32 to L. Wang).

I am regularly sought out for my expertise in gene mapping studies and thus my work has made significant contributions in the application of advanced mapping methods to a host of other projects. Recent papers have utilized principal components analysis to identify subjects of Jewish ancestry and conduct an analysis of sixteen genetic variants for pancreatic cancer among this narrowly defined ancestral population; genome-wide imputation to identify variants in piRNAs associated with glioma; risk score analysis to identify multiple variants associated with bipolar disease; and meta-analyses of genetic variants associated with infectious diseases to summarize previous evidence for genetic associations with West Nile and Dengue diseases.

My statistical work is focused on extending methods to identify genetic interactions as well as loci exhibiting pleiotropy. Given the current number of single-nucleotide polymorphisms (SNPs) genotyped simultaneously in a GWAS, examining all SNP-SNP interactions becomes daunting and increasingly difficult to identify statistically significant and replicable results. I am working on developing analytical approaches for conducting multi-stage analyses to identify replicable interactions in genome-wide data. One strategy to reduce phenotypic heterogeneity is to analyze multiple correlated phenotypes jointly and look for pleiotropic loci and a recent paper from my group has summarized the statistical analysis techniques that can be utilized for different data types to identify cross phenotype associations and then dissect these findings to identify pleiotropy. These dissection approaches are being applied to a large dataset to look for pleiotropic loci for asthma and obesity and comparing the power of univariate and multivariate methods (F31 to Y. Salinas).

Moving forward, I am expanding this work to develop novel methods for incorporating rare and imputed variants for mapping pleiotropic loci that also leverages other types of omics data including gene expression and whole genome sequence data. Pleiotropy, although an important phenomenon in genetic etiology, has not been adequately studied and there are limited methods to detect pleiotropy for rare and imputed variants. My work will address this gap using a multi-prong approach of pleiotropic association testing, estimating tissue-specific disease heritability and detecting tissue-specific pleiotropy. This highly innovative work is supported by two NIH R01 grants. Not only does the study of pleiotropic loci improve our understanding of the genetic etiology for complex diseases and traits, but it also has high public health significance; pleiotropic effects will improve our ability to estimate genetic risk and provide insight into drug targets for the development of treatments for multiple diseases due to shared genetic architecture.

One major challenge in genetic epidemiology is to identify additional risk loci that may have smaller effect sizes, affect multiple phenotypes, and/or non-additive effects. Through the genetic analyses of various phenotypes, models and diverse ancestral populations, and in concert with other omics data and clinical investigations, we will gain a better understanding of the genetic architecture of complex traits and begin to build the foundation for incorporating personalized medicine into clinical practice. My work will continue to make significant contributions to this goal.