Combinatorial Polyfunctionality Analysis of Single-Cell Genomic Data

• Activity Number: 169
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Lynn Lin*+
• Companies: Fred Hutchinson Cancer Research Center
• Keywords: Single-cell genomic data ; Reversible Jump MCMC ; Bayesian Analysis

Abstract

Advances in single-cell technologies have enabled high-dimensional, high-throughput gene expression measurements on single cell level. Using these single-cell assays, generic variations about cell population heterogeneity can now be studied. In this vaccine clinical trial study, the expressions of 96 genes are measured in about 90 single cells from each 16 individuals under two conditions (antigen stimulation vs. non-stimulation). We develop a computational framework that enables unbiased polyfunctionality analysis of antigen-specific cell subsets. Within each subset, the cells are simultaneously expressing multiple genes. We use reversible jump Markov chain Monte Carlo method to develop strategies for selecting and modeling important functional cell-subsets and select the most likely gene expression patterns to be antigen-specific for high-dimensional count data. Under the identified most probable model, the specificities are quantified by a set of posterior probabilities, which can be used to profile a subject's immune response to external stimuli.