CAREER: Deciphering the human regulome: omics-based analysis of intergenic genotype-to-trait associations, made accessible and powerful

Oregon State University, Corvallis OR

Investigators

Abstract

This project's research activities will advance the field of bioinformatics by creating a computational method that combines different types of information from genome-wide association studies. Genetic association studies measure sequence differences across an entire genome in order to identify what variants cause the occurrence and variability of traits like height and disease susceptibility. The goal of this project is to develop methods to precisely locate gene regulatory variants, that now are only known to be somewhere in a large region, and use them to understand how traits vary in a population. The methods developed in this research aim to combine a number of types of information, like gene expression levels for cells, measurement of traits in many individuals, and comparisons with traits in other species, in order to identify causal regulatory variants. The project's curriculum development activities will contribute to STEM education by creating and sharing a hands-on workshop on genome bioinformatics in the research area. By integrating research and educational activities students will (i) gain science literacy in the areas of genetics and bioinformatics; (ii) show how well hands-on methods work in genetics education; and (iii) use already performed genetic association studies to gain new knowledge in biology and in biomedicine. This research will create and evaluate an integrative machine-learning model for identifying regulatory variants within human intergenic GWAS regions. The model's inputs will include the reference genome, the local DNA 3-D shape, phylogenetic conservation, and transcriptomic and epigenomic measurements. The model's output will be predicted regulatory variants with significance scores. The model will be benchmarked against published methods using ground-truth regulatory variants. The machine-learning model's variant predictions will be incorporated into an open-source, web-based software tool for integrative post-analysis of GWAS data. Compatibility with a cloud-computing framework will position the tool for maximum impact. Through educational activities that are integrated with the project's research activities, Dr. Ramsey will create, evaluate, and disseminate a Genome Bioinformatics Workshop unit for high school educators and students. Participants will learn to use the tool to analyze and explore human GWAS-identified regions for a model trait (height); through this they would be expected to gain a better understanding of the potential of the field of personal genomics. The workshop's materials will be developed within an interdisciplinary workshop incubator consisting of pairs of STEM-underrepresented CS and biology undergraduate summer students. We will create, evaluate, and disseminate the workshop unit in partnership with three outreach programs for STEM-underrepresented students. Project results will be made available at the project website: lab.saramsey.org/regulome