Statistical Analysis of Serial Dilution Assays Using Estimating Functions and Data Cloning

• Activity Number: 471
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Computing
• Author(s): Subhash Lele*+
• Companies: University of Alberta
• Keywords: Hierarchical models ; Bayesian inference ; Pseudolikelihood ; Non linear models ; MCMC convergence

Abstract

Serial dilution assay data are commonly analyzed using non-linear regression analysis of the standard samples alone. Recently hierarchical models are proposed to combine data across patients. These hierarchical models are usually analyzed using non-informative Bayesian approach. The non-linear regression method suffers from the instability of the solutions and the Bayesian approach suffers from difficult MCMC convergence. In this paper, we use data cloning and estimating functions to obtain maximum likelihood estimators and best predictors of the individual concentration while removing the instability of the solutions for non-linear models and non-convergence issues associated with the MCMC.