Yana Bromberg, Ph.D.
School of Environmental and biological Sciences
Noon, Monday November 4, 2019
Auditorium, Life Sciences Building
145 Bevier Rd. Busch Campus, Piscataway, NJ 08854
"Identifying Functional Signatures of Disease"
At least a fifth of the exonic non-synonymous variants in our genomes alter the molecular functions of the genes that they affect. Distinguishing between the disease-affected and healthy individual genome combinations of functionally non-neutral variants in relevant pathways contributes to our understanding of the genetic mechanisms of complex disease. Our lab’s novel computational methods leverage functional effects of genome variants in disorder-specific genes to predict individual disease susceptibility. For Crohn’s disease (CD), in particular, our machine learning method, AVA,Dx (Analysis of Variation for Association with Disease), uses whole exome sequencing data of only 111 individuals to accurately differentiate patients from healthy controls. By additionally accounting for batch effects, we are able to predict individual CD status for nearly 3,000 individuals from other studies. As an added bonus, AVA,Dx highlights previously unknown pathogenesis pathways. Our work motivates new experimentally testable hypothesis regarding the biological mechanisms of disease and may eventually provide a means for earlier prognosis, more accurate diagnosis and the development of better treatments.