Evolutionary Transcriptome Analysis Based on Differentially Expressed (DE) Genes

Abstract

To address how gene regulation plays a key role in phenotypic innovations through high throughput transcriptomes, it is desirable to develop statistically-sound methods that enable researchers to study the pattern of transcriptome evolution. Most methods currently available are based on the Ornstein-Uhlenbeck (OU) model that considers the stabilizing selection as the baseline model of transcriptome evolution. In this paper, we developed a new evolutionary approach, based on the genome-wide p-value profile arising from statistical testing of differentially expressed (DE) genes between species. Our current approach is focused on the estimation of transcriptome distance between species. We first establish the relationship between the evolutionary model (the Markov-chain or Poisson model) and the proportion of null hypothesis (u0), which can be used to estimate the transcriptome distance. Further, we calculate the posterior probability of a gene being DE when a p-value is given, denoted by Q=P(DE|p), and develop a simple algorithm to estimate the transcriptome distance for any number of genes in the genome. Our compute simulations showed the statistical performance of these new methods are generally satisfactory.