Online Program Home
My Program

Abstract Details

Activity Number: 237 - SPEED:Statistical Methods for GWAs, Genetics, Genomics, and Other Omics Studies, Part 1
Type: Contributed
Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #306406
Title: A Feature Allocation Model for Cytometry by Time-Of-Flight Data
Author(s): Arthur Lui* and Juhee Lee and Peter Thall and Katy Rezvani
Companies: University of California - Santa Cruz and University of California, Santa Cruz and U.T. M.D. Anderson Cancer Center and M.D. Anderson Cancer Center
Keywords: Indian buffet process; Bayesian nonparametrics; cytometry; feature allocation model; missing value imputation; clustering

A Bayesian feature allocation model embedded with clustering capabilities is developed to analyze mass cytometry data and characterize their underlying heterogeneous cell repertoire structures. Each repertoire consists of a collection of cells possessing different phenotypes that can be characterized by differences in expression levels of cell surface markers. In particular, mass cytometry data collected to study the clinical efficacy of natural killer (NK) cells as immunotherapeutic agents against leukemia are considered. The data of interest includes expression levels of 32 surface markers on thousands of cells from multiple samples. The proposed model simultaneously characterizes NK cell phenotypes based on the expression of surface markers, and estimates the composition of samples based on the identified phenotypes. The Indian buffet process is utilized to model cell phenotypes. Non-ignorable missing data present due to technical artifacts in mass cytometry instruments are imputed. We present simulation studies and a study of cord-blood data from MD Anderson Cancer Center.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program