ABSTRACT:

Cancer is a disease driven by rounds of Darwinian selection on somatic genetic mutations, and recent advances in sequencing technologies is offering new opportunities as well as revealing new challenges in this field.    In this talk, I will describe two statistical problems in the genetic analysis of tumors.  In the first part, I will describe the problem of allele-specific copy number estimation.  Copy number change is a basic type of DNA alteration in tumors, and understanding how they affect the tumor’s genome at the allelic level is fundamental to understanding the tumor’s genetic signature.  I will describe a bivariate binomial mixture process for this problem, and a method for detecting change-points in this process.  In the second part, I will describe the problem of inferring a tumor’s clonal evolutionary history through repeated bulk DNA sampling.  This is similar to classic phylogenetic inference problems, with the key difference being that the observed data are slices of a mixed population.    I will describe a framework that we developed to estimate the underlying evolutionary tree by joint modeling single nucleotide mutation and allele-specific copy number profiles. 

BIO:

Nancy R. Zhang obtained her doctoral degree in Statistics at Stanford University in 2005.  After a year’s post-doctoral study at UC Berkeley, she returned to Stanford as assistant professor in the department of Statistics.  She was promoted to associate professor at Stanford in 2011, when she moved to the department of Statistics in the Wharton School at University of Pennsylvania.  Currently, she is working in the area of applying statistical concepts to modeling and inference problems in computational genomics.  In particular, she is developing computational techniques to study heterogeneous tissues through bulk and single-cell sequencing data.