Developmental transcriptomes of Zea mays and Arabidopsis thaliana: from coexpression modules to genome evolution
By varying transcriptional profiles both spatially and temporally, plants create a variety of developmental units (such as leaves or roots or flowers), and patterns of gene expression evolve over generations. By examining gene transcript abundances across the breadth of development, we can obtain a genome-wide view of gene regulation. These patterns of coexpression can be used to discover the conserved non-coding sequences that act as transcription factor binding sites that coordinate expression. In the agriculturally significant crop species Zea mays (maize), I examined the transcriptome within 50 different tissue types at different developmental stages, including leaf (nine stages), root (eight stages), tassel (seven stages), anthers, ear, embryo, endosperm, floret, husk, ovule, pericarp, silk and stalk (one to four stages each). I used hierarchical clustering of 34,876 gene models to form 24 gene coexpression modules of genes with similar expression profiles. Sixteen modules revealed patterns of tissue-specific expression. Some modules correlated with the developmental age of tissues. Gene Ontology (GO) enrichment analysis confirmed that many modules consist of genes with similar biological functions. Three of the modules consist of genes that are physically proximal within the genome. Six modules contain an over-representation of known promoter motifs upstream of module genes and a search for de novo promoter motifs revealed novel sequences. In the model plant species Arabidopsis thaliana, I examined a set of 30 tissues and compared the sequence and expression divergence of pairs of genes that were duplicated via an ancient polyploidy event. Gene pairs that are constrained in sequence divergence are also constrained in expression divergence, indicating that purifying selection is the primary mode of gene evolution. Constraint is greater among genes that are expressed in several tissues than genes expressed in fewer tissues. Genes with different biological functions (GO terms) follow different patterns of divergence, with transcription factors evolving quickly and structural molecules diverging slowly.