A Two-Level Luria-Delbruck Mutation Model: Improving the Accuracy of Mutation Rate Estimates

• Activity Number: 25
• Type: Contributed
• Date/Time: Sunday, July 31, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Qi Zheng*+
• Companies: Texas A & M University
• Keywords: mutation rate ; hierarchical model

Abstract

The fluctuation experiment is the method of choice for estimating microbial mutation rates. In a fluctuation experiment a number of test tubes are each seeded with a small number of normal cells; during an incubation period, mutation may occur in each tube and the mutant cells multiply. The numbers of mutant cells in the tubes at the end of the experiment are regarded as identically distributed random variables and are used to estimate the mutation rate. This practice relies on the assumption that mutation rates are the same across all test tubes. Experimental evidence suggests that this assumption is not always tenable. We present a two-level mutation model to relax this assumption, thereby improving the accuracy of mutation rate estimates. Under the two-level mutation model the tubes in an experiment are considered exchangeable environments, and each tube has its own inherent mutation rate drawn from a gamma prior distribution. Individual mutation rates are estimated within the standard hierarchical model framework. We present a theoretical finding that simplifies the simulation of the posterior distributions of the individual mutation rates.