Dr. Andrew Hale

Supervisors: Dr. Eric R. Gamazon, PhD

Project Title: Integration of human genetics and multi-omics data implicates shared genetic architecture of the infectious disease phenome and neurodevelopmental traits

News Byte: Infectious diseases (ID) represent a significant proportion of morbidity and mortality across the world. Furthermore, the extent to which neurodevelopmental traits, for which central nervous system infections may be contributing factors, share genetic architecture with ID phenotypes is unknown. Using large-scale Electronic Health Records (EHR; ~3 million individuals), we identified comorbidity between neurodevelopmental traits and ID phenotypes (beyond known risk-factors), suggesting the presence of shared (genetic, environmental or interaction of the two) causal factors. We provide an atlas of potentially causal genes for 35 ID phenotypes, identifying 70 experiment-wide or individual-trait gene-level associations and replication in the independent UK Biobank and FinnGen consortia. We then performed systematic validation of identified genes using pathogen-exposure induced cellular phenotypes including intercellular spread, cytokine production, and pro-angiogenic growth factor host response, among others. These data provide potential functional and mechanistic insights into the role of host genetic variation. Finally, we developed a broadly-applicable methodology (Pleio-SCAN) to estimate both pleiotropic and trait-specific genetic effects using the genetic component of gene expression. Our study reveals unique insights into the genetic architecture of ID phenotypes and their relationship to neurodevelopmental disease.

Abstract: Hydrocephalus (HCP) is a component of ~200 genetic syndromes and a secondary consequence of many pathologies (infection, hemorrhage, etc.). Understanding the genetic architecture of HCP may lead to new therapeutic approaches. We use PrediXcan, which estimates the genetically-determined component of gene expression using common variant (>1%) data and a reference transcriptome (GTEx) of neurological tissues and whole blood. We identify MAEL as a genome-wide predictor of HCP across neurological tissues in individuals of European-ancestry and provide replication in the UK Biobank. Additional support is garnered using analysis of magnetic resonance imaging phenotypes in the UK Biobank, an exome scan in 29,713 patients, transcriptome analysis of choroid plexus in a murine model of HCP and protein expression analysis from cerebrospinal fluid isolated from patients with HCP. To extend our analysis to diverse populations, we use PrediXcan to identify gene-based associations with HCP in African-ancestry individuals; we identify TMEM208 reaching genome-wide significance and observe a significant enrichment of infection-related pathways among nominally-associated genes. We then use large-scale electronic health records (~3 million individuals) to quantify comorbidity between HCP and 35 infectious disease (ID) phenotypes, extending far beyond known risk-factors, suggesting shared (genetic, environmental or interaction of the two) causal factors. Finally, we identify 70 gene-level associations across 35 ID phenotypes, providing the basis of a broadly-applicable methodology (Pleio-SCAN) to estimate both pleiotropic and trait-specific genetic effects between two traits, which we apply to individual IDs and hydrocephalus/CSF shunt failure. Our data reveals insights into the polygenic architecture of hydrocephalus and highlights the shared genetic risk for IDs and HCP, even in the absence of infection as the antecedent cause of HCP.