Jun
23
Thu
BMIR SPECIAL RESEARCH COLLOQUIUM: “Critical Care endotypes: Collecting and analyzing large datasets to optimize diagnosis in the ICU” @ MSOB, Conference Room X-275
Jun 23 @ 4:00 pm – 5:15 pm

David Maslove

David Maslove, MD, MS
Assistant Professor
Department of Medicine and Critical Care
Queens University

ABSTRACT:
The diseases treated in the Intensive Care Units (ICU) are syndromic in nature, largely defined by a number of vague or arbitrary criteria. As a result, there exists significant case mixing within ICU syndromes, whereby different syndrome subtypes – or endotypes – are all treated the same way. This limits the effectiveness of evidence-based practices evaluated in large clinical trials, and makes critical care particularly suitable to a precision approach. Further enabling this approach is the incredible abundance of data generated in the ICU, upon which distinctions between patients can be made. Making the most of these data requires that they be systematically collected, vetted, merged, and analyzed, ideally in a near real-time environment that respects the fast-paced nature of ICU practice. We will explore some key challenges to this approach, with a focus on specific examples involving physiologic, clinical, and genomic data generated from critically ill and injured patients.

Feb
27
Thu
Center for Population Health Sciences Seminar Series: Elliot M. Tucker-Drob, University of Texas, Austin @ Li Ka Shing Learning and Knowledge Center, Room 320
Feb 27 @ 2:00 pm – 3:00 pm
Center for Population Health Sciences Seminar Series: Elliot M. Tucker-Drob, University of Texas, Austin @ Li Ka Shing Learning and Knowledge Center, Room 320

Using Genome-Wide Data to Investigate the Joint Genetic Architecture of Complex Traits

Methods for using genome-wide data to estimate genetic overlap between pairwise combinations of traits have produced “atlases” of genetic architecture. These atlases have revealed pervasive genetic sharing across different social, behavioral, and mental health outcomes, and individual risk variants are often found to be associated with multiple such outcomes. My group has recently a developed formal multivariate framework for analyzing the joint genetic architectures of constellations of traits using summary statistics from existing Genome-Wide Association Studies. This method, Genomic Structural Equation Modeling (Genomic SEM), can be used to identify variants with effects on both general and specific dimensions of cross-trait liability, boost power for discovery, and calculate more predictive polygenic scores. Genomic SEM can further be used to construct and test hypothesized cascade models of trait development and disease progression. In this talk, I describe the Genomic SEM framework and present results of a series of substantive applications of the method in the areas of psychology and mental health.

Register here