SciLifeLab The Svedberg seminar series 2014-08-21
MIT Department of Computer Science. MIT Department of Computer Science. Broad Institute of MIT and Harvard, USA.
Manolis Kellis is a Professor of Copmuter Science at MIT, a member of the Computer Science and Artificial Intelligence Laboratory and of the Broad Institute of MIT and Harvard, where he directs the MIT Computational Biology Group. His group works at the interface of computational and experimental genomics, and has pioneered methods for the discovery of regulatory elements and circuits using comparative genomics, epigenomics, and human genetics. His group has led or contributed to the integrative analysis efforts of several large-scale genomics projects, including the NIH Roadmap Epigenomics project, the ENCODE project, the comparative analysis of 29 mammals, and the Genotype Tissue-Expression (GTEx) project. He has recieved the US Presidential Early Career Award in Science and Engineering (PECASE), the NSF CAREER award, the Alfred P. Sloan Fellowship, and was recognized as a top young innovator by Technology Review, Genome Technology, and the Boston Museum of Science. He obtained his Ph.D. from MIT, where he recieved the Sprowls award for the best doctorate thesis in computer science. Prior to computational biology, he worked on artificial intelligence, machine vision, robotics, and computational geometry, at MIT and the Xerox Palo Alto Research Center. He lived in Greece and France before moving to the US.
Regulatory and Systems Genomics of Human disease
One of the greatest surprises of genetic studies of human disease and complex traits over the last 10 years is the realization the the vast majority of disease-associated loci have weak effects and lie outside protein-coding genes- Thus, understanding the molecular basis of human disease will require a systematic understanding of non-coding DNA elements, and the regulatory circuits they define. In this talk, I will describe our work towards this goal in the context of the NIH ENCODE and Roadmap Epigenomics projects. We use histone modification patterns and chromatin accessibility across 127 tissue/cell types to establish global maps of tissue-specific regulatory elements, define their activity patterns, and link them to their regulators and their target genes. We use this information to understand the molecular basis of genetic associations with human disease, to predict disease-relevant tissues, and to aggregate weak genetic signals across thousands of loci to reveal disease-relevant pathways and regulators. Lastly, we confirm our regulatory models based on epigenetic differences in brain-specific enhancer regions across 708 individuals in the context with Alzheimer’s disease, and using directed experimentation in human cell lines, and in mouse and zebrafish models for neurodegeneration, cardiovascular traits, and obesity phenotypes. Our results suggest that thousands of regulatory regions underlie complex traits and human disease, and provide a systematic model for discovering potential drug target and therapeutics by exploiting the cellular circuitry id disease.