Bayesian Array Analysis To Assess the Effects of Nutrient Deficiencies in White Lupin

• Activity Number: 242
• Type: Contributed
• Date/Time: Tuesday, August 5, 2008 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Robert Norton*+ and Mousumi Rath and Bandita Parhy and Claudia Stone
• Companies: California State University, East Bay and California State University, East Bay and California State University, East Bay and California State University, East Bay
• Keywords: RNAi ; array analysis ; biometrics ; bioinformatics ; computational biology ; qPCR

Abstract

Lupinus albus L. has unique adaptations for extracting phosphorus (P) and iron (Fe) from soil when other plants cannot. The objective of this research is to identify genes up- or down-regulated in proteoid roots under P, Fe and nitrogen (N) deficiency. We begin with array data for two up-regulated genes, six replicates exist for each gene under each condition. Assume that the number of ratios of treatment to control expression intensities are>2; is a binomial variable with unknown proportion pi for each gene. This paper compares outcomes using a standard chi-square test; a flat prior on both unknown proportions; and a beta prior on pi. Since the number of replicates is small, it is reasonable to expect Bayesian results to be more useful in making decisions about the outcomes. The Bayesian analysis provides estimates for the proportion pi and the covariance of the two gene expressions.