A bioinformatic study on the evolution of bamboos and other graminoid poales

For the past decade, next-generation sequencing (NGS) has been used to undertake genomics-scale projects in molecular biology. This method of sequencing involves randomly fragmenting a sample of nucleic acid and randomly generating millions of short reads. The large number of reads moves the assembly and other analyses to be performed by computer algorithms. As well as sequencing full chromosomes, full RNA extracts can be sequenced to determine levels of gene expression and exon boundaries.;In this study, NGS is used to examine the evolution of bamboos (Bambusoideae), which are a subfamily of grasses (Poaceae). Bamboos are divided into two main phenotypes: woody and herbaceous. Woody bamboos are characterized by lignified culms, bisexual florets and undergo gregarious monocarpy, while herbaceous bamboos have less lignified shoots, unisexual florets and flower annually.;The evolution of this subfamily was examined in a phylogenomic framework using full chloroplast genome (plastome) sequences. First, a set of methods for plastome assembly was developed using automatable scripts and accuracy-testing steps. These methods were then used to generate full plastomes from bamboos. Full plastomes were then analyzed phylogenomically under a maximum-likelihood and Bayesian framework. This analysis revealed paraphyly between the temperate woody and tropical woody clades. Previously established tribal and subtribal relationships were also confirmed.;Nuclear transcripts were assembled from four bamboo species by sequencing RNA from floral tissue using NGS. Two assembly methods were performed: a de novo assembly used only overlapping reads to produce transcripts and a reference-based assembly used a previously sequenced nuclear genome from a bamboo to place reads and assemble them into transcripts. After quality assessment, the reference-based produced about double the number of transcripts than then de novo assembly, but may be biased toward the number of nuclear genes found in the reference genome. Floral development transcripts were examined as gene trees. Several preliminary correlations to phenotype were determined using putative gene function. The nuclear transcripts were then used to perform a phylogenetic analysis. After 3,878 transcripts were determined to be single-copy, they were concatenated and analyzed under a maximum-likelihood framework. Each of the 3,878 transcripts were also analyzed individually. This analysis strongly supported a monophyletic relationship between the woody bamboos, which contrasted with the plastome analysis.;Finally, a plastome was generated for a near-grass Joinvillea ascendens (Joinvilleaceae). The plastome was generated using completely de novo methods and was found to have undergone two large-scale inversions. These inversions occurred after two other large-scale inversions in the Poaceae-Joinvilleaceae lineage, which had been previously documented using PCR amplification. The two previously documented inversions were verified by comparing grasses to a plastome with an ancestral gene arrangement. Several gene and intron losses were also verified in grasses.