The goal of this NIH Pathway to Independence award is to provide Dr. Brittany Lasseigne with an extensive training program to prepare her to be an effective independent investigator who uses computational genomics to study complex human diseases. We propose a formal one-year training and mentoring program in genomics, computer science, statistics, and career development to build on her 8+ years of hands-on training, followed by a three-year structured and independent research program. Research will focus on the integration of multidimensional genomic data sets in the context of complex human diseases. A critical barrier in genomic research is the complexity of data integration: the ability to leverage overlapping and unique information captured by different genomic assays would improve our understanding of data integration and generate clinically relevant genomic signatures. To meet this need, we propose to integrate a combination of genomic data we generated with public data to (1) infer genomic instability signatures from different data types, (2) improve clinically relevant phenotype prediction by building multi-omics machine learning classifiers and reducing phenotype heterogeneity, and (3) create a cloud-enabled R package and associated Shiny application to accelerate future research. The proposed work will advance our understanding of data integration, allow inference of genomic instabilities across data sets, and generate high performance classifiers for assessing clinically relevant phenotypes in both cancer and psychiatric disease using frameworks that will be broadly applicable across other complex diseases. It will also facilitate prioritization of experiments in future studies by informing on the orthogonality of genomic assays, thereby allowing more efficient study designs to capture as much information as possible within a given sample size or scope of experimentation. Collectively, this additional training will allow Dr. Lasseigne to develop new multidimensional data integration approaches and translational questions applicable across complex diseases when independent. Dr. Richard Myers (HudsonAlpha) and Dr. Gregory Cooper (HudsonAlpha), leaders in applying genetics and genomics to complex human diseases, and an Advisory Committee of additional experts including Dr. Barbara Wold (Caltech), Dr. Eddy Yang (UAB), and Dr. Timothy Reddy (Duke), will provide mentoring throughout this award. The mentored phase will take place at the HudsonAlpha Institute for Biotechnology, an ideal environment for this training with extensive translational science collaborations, expert faculty and staff, and state-of-the art computational and laboratory resources devoted to genomics. This combination will maximize Dr. Lasseigne's training program, facilitating her transition to an independent, tenure-track investigator at a university with a strong commitment to data-driven approaches to complex human disease research, i.e. strong genomics research programs with clinical collaborators, ideally at, or affiliated with, an academic medical center.
The major outcome of this project will be a scientist with the necessary research, mentoring, teaching, and career development training to run an independent research program in computational genomics. The research proposed will apply novel strategies to further develop integrative machine learning analyses of multidimensional genomic data, discover clinically relevant predictive models, and create computational tools to accelerate future research.
