Data Analytics for High Throughput Drug Screens

点击翻译按钮获取中文摘要

This grant will contribute to the advancement of national prosperity and economic welfare by developing techniques for the design and implementation of precision drug treatments for cancer patients. A single tumor typically contains heterogenous cells that differ in their response to different drug treatments. Even small subpopulations of drug-resistant cells may cause tumor recurrence and therapeutic failure. The project will develop tools to identify the distinct sub-populations of cells present in the tumor based on its response to therapy, leading to personalized treatment plans for cancer patients and the development of new methodologies for deconvolution of mixed signals. The associated educational activities will include mentoring graduate and undergraduate students, developing course content related to the work, and short summer school for high-school students.This grant will use the differential response of the sub-populations to the drug under study to identify subpopulations and quantify their initial fraction as well as their individual dose response. Models assuming no interactions between the sub-populations as well as models with both linear and nonlinear interactions will be considered. This requires the formulation of statistical inference problems based upon approximations of the likelihood function of sample path trajectories of non-Markovian process models capturing the underlying cellular population dynamics. Identifying estimates of the parameters requires solving constrained optimization problems with highly nonlinear and non-convex objective functions.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.