WITHDRAWN: Structured Prior Modeling for Bayesian Analysis of Complex Mixtures in Combinatorial Marker-Based Flow Cytometry Studies

• Author(s): Lin Lin and Cliburn Chan and Mike West
• Keywords: Identification ; Bayesian mixture modeling ; MCMC

Abstract

In study of immunology,flow cytometry measures each level of marker protein using single-colour fluorescently labeled tags.One important recent development is using combinatorial marker assays in which each marker is targeted with a probe labeled with two or more fluorescent tags. Thus, using K colours enables the identification of up to(K choose m)subtypes of cells using m colours per type, advancing the ability to characterize variation in immune responses involving larger numbers of cell subtypes. We describe novel classes of Bayesian mixture models with highly structured priors that reflect the design and nature of resulting data under combinatorial encoding setting. We develop Bayesian computation using MCMC, discuss the challenging issue of identification and a novel, general approach to its solution, provide evaluation examples with simulated data and application in immunology.