Ken Boucher Group

Boucher Group Research

Ken Boucher

Figure 1 Oncogenetic tree model for colon cancer

Ken Boucher has varied research interests in mathematical modeling and statistics.  Three of these are described below

Oncogenetic trees.
We described a simple algorithm for fitting a model of carcinogenesis in which the genetic abnormalities causing the cancer are partially ordered, resulting in a labeled, oriented tree structure. We later applied oncogentic trees to colon tumors (This is joint work with A. Szabo, M. Slattery and C Sweeney, among others).

Analysis of high throughput sequencing data.
Dr. Boucher has an ongoing collaboration with David Nix and Brett Milash to biostatistical and bioinformatics tools for high throughput sequencing data. Examples include existing methods for analysis of Chromatin immunoprecipitation followed by high throughput sequencing (ChIP-Seq) data. Current efforts are underway to take the correlation structure present in sequencing data into account in study design. The average correlation in expression between a random pair of genes may exceed 0.50 in a typical experiment.

Analysis of longitudinal outcomes in cohorts with high rates of attrition due to death or other clinical events.
Analyses relating treatments and exposures to longitudinal measures of biomarkers, quality-of-life, and other outcomes are often complicated in cancer populations by high rates of attrition due to death or other clinical events. This problem, referred to as “truncation by death” in the statistical literature, often produces biased comparisons of the longitudinal outcomes, particularly when the proportions of subjects with events that terminate follow-up differ between the groups being compared. Dr. Boucher is part of a team that is developing and applying the recently formulated framework of principal stratification to develop improved statistical methodologies to address this problem in collaboration with Tom Greene at the University of Utah and collaborators at MD Anderson Cancer Center, the Cleveland Clinic, and the University of Pennsylvania.