Skip to main content
Browse by:
GROUP

Reconstructing the evolutionary history of tumors

Event Image
Wednesday, October 29, 2014
3:30 pm - 4:30 pm
Quaid Morris, University of Toronto
Machine Learning Seminar

12noon lunch and lecture ¿Perspectives in Machine Learning¿3:30-4:30pm Seminar reception followingTumors consist of genetically diverse subpopulations of cells that differ in their response to therapy and their metastatic potential. The short read sequencing commonly used (and on the verge of becoming clinical practice) to characterize tumor heterogeneity only provides the allelic frequencies of the tumor somatic mutations, not full genotypes of individual cells. I will describe my lab¿s efforts to recover these full genotypes by reconstructing the tumor¿s evolutionary history. We do this by fitting subpopulation phylogenies to the allele frequency data. In some circumstances, a full reconstruction is possible but often multiple phylogenies are consistent with the data. We have developed a number of methods (PhyloSub, PhyloWGS, treeCRP) that use Bayesian inference in non-parametric models to distinguish ambiguous and unambiguous portions of the phylogeny thereby explicitly representing reconstruction uncertainty. Our methods incorporate simple somatic mutations (single nucleotide variants and small indels) as well as copy number variations; have excellent results on real and simulated data; and can take as input allele frequencies from single or multiple tumor samples where these frequencies are estimated using either targeted or whole genome sequencing.Bio:Quaid MORRIS is an associate professor in the Donnelly Centre at the University of Toronto in Canada.

Type: LECTURE/TALK
Contact: Sayan Mukherjee