Computational genomics of developmental gene regulation

Abstract

The development of multicellular organisms requires the precise execution of complex transcriptional programs. The demands posed by development, coupled with the relatively late evolution of multicellularity, could have led to a separate mode of gene regulation for gene involved in, and regulated throughout development. I investigated the regulation of genes by enhancers using histone modifications coupled to gene expression, based on the observa- tion that developmental genes are surrounded by dense clusters of conserved enhancers which act in concert. Genes regulated by enhancers are much more likely to be developmentally regulated genes, and many enhancers at each loci co-ordinate to direct transcription across multiple tissues. CAGE-seq is a powerful tool for determining the structure of promot- ers. I analysed promoters in Amphioxus using CAGE-seq to determine if the diverse promoter architectures observed in vertebrates had ancestral ori- gins. Promoters in amphioxus can be divided into developmental and house- keeping promoters, which each have characteristic patterns of dinucleotide enrichment. Housekeeping promoters in Amphioxus have a novel promoter architecture, and a contain a high frequency of bidirectional promoters, which represents the ancestral vertebrate state. This set of genes highlight the mal- leability of promoter architecture during evolution. I developed a package in R/Bioconductor ‘heatmaps’ to enable effective visualisation of this, and other, data. Taken together, these results suggest a second mode of regulation in ver- tebrates governing the regulation of developmental genes.