The effect of common disease associated genetic variants on transcriptomic and early life phenotypes

Scientists have recently made substantial progress in identifying genes that increase the risk of ?common diseases? like diabetes, heart disease, asthma, and various forms of cancer. However, in many cases the biological pathways through which these genes exert their negative effects are unknown. The aim of this proposal is to better characterize the biological systems that genetically predispose some people to future disease. 101 genetic markers associated with common diseases will be genotyped in 10,000 children from the Avon Longitudinal Study of Parents and their Children (ALSPAC). ALSPAC is a very large study containing detailed information on hundreds of biological variables relevant to future health, including measures of genetic functioning in 1000 of these children. The planned research revolves around investigating the genetic markers (i.e. which are associated with disease), and their relationship to the biological variables and measures of genetic functioning contained within ALSPAC. In order to integrate the many different data types and make maximum use of the information, a combination of existing and new statistical methodology will be employed. In the event that an interesting biological effect is identified, the experiment will be repeated in two other large studies that contain similar information- the RAINE study in Australia, and the Pelotas study in Brazil. This will give an indication of whether the finding is genuine or whether it was simply due to chance. It is expected that this sort of approach will increase our understanding of the causal pathways involved in disease etiology, and enhance our ability to detect disease-genes above and beyond what could be achieved through standard methods alone.

Explosive growth in knowledge related to the genomic architecture of disease and genotyping technologies coupled with effective study designs has resulted in a rapidly expanding list of genes known to affect complex human phenotypes and diseases. However, in many cases the underlying biological mechanisms responsible for the development of disease and consistently observed associations remain unknown. The aim of this application is to genotype SNPs that are robustly associated with disease in several large population based cohorts in which detailed transcriptomic and endophenotype measures are available, with the intention of elucidating the biological pathways through which these genetic variants predispose to future disease. 101 SNPs reliably associated with common diseases will be genotyped in 10,000 offspring from the Avon Longitudinal Study of Parents and their Children (ALSPAC) cohort. These children have been measured longitudinally on hundreds of detailed endophenotypes relevant to future disease development (i.e. cardiovascular disease, obesity, diabetes, cancer, psychiatric morbidity and neuro-cognitive functioning), and in the case of 1000 of them, will also have genome-wide mRNA expression levels assayed in lymphoblastoid cell lines (~48,000 transcripts).

SNPs will be tested for association with transcriptomic and endophenotype variables using standard linear regression in the case of continuous measures, and logistic regression in the case of binary outcomes. Putative associations will be tested for replication in the Australian RAINE (N = 3,000) and/or Brazillian Pelotas cohorts (N = 11,000 in total) depending on the availability of similar endophenotypes. Complex relationships between SNPs, mRNA levels and endophenotypes will be investigated more thoroughly by existing and novel multivariate approaches including the construction of Bayesian co-expression gene networks. In this regard, a central aim of the proposal will be the development and characterization of statistical methodology for analyzing high dimensional multivariate genetic, transcriptomic and phenotypic data. Statistical tools and software for the (multivariate) analysis of genetic data will be made available to the academic community. It is expected that integrating information from genetic, transcriptomic and phenotypic sources will not only identify molecular pathways involved in disease pathogenesis, but also significantly enhance ability to detect disease-gene associations above and beyond what could be achieved through genetic association studies alone.