| The ability to recognize patterns that allow the distinction of genes and the formation of discrete family groups requires the generation and use of multiple classes of information. For any given family, these classes of knowledge would include (1) the sequence properties of the genes and its corresponding proteins, including its possession of defined motifs; (2) the phylogenetic history and pattern of conservation of the gene and the gene product; and (3) the structural features of the gene product. In order to gather this information and furnish the data needed to advance inferences on gene families of the monolignol biosynthetic pathway, Chapter 2 describes the three-dimensional structural models of 4-coumarate:CoA ligase, cinnamoyl-CoA reductase and cinnamyl alcohol dehydrogenase, that were built using a comparative modelling approach. The three proteins show a Rossman fold domain, but they were evolutionarily unrelated. The great benefit of recognizing homolog positions with discriminant capability poses the question whether these positions have a predictive ability? Could this signature positions that effectively distinguish subfamilies be applied with the purpose to predict subfamily structure in loblolly pine DNA sequences? Does loblolly pine subfamily structure for a given gene follows a similar organization as known in other plant species? Chapter 4 describes the application of the signature in this context using loblolly pine expressed sequence tag (EST) data. The rationale of the method involves using an alignment and phylogeny of translated sequences of known genes to detect the signature positions (scaffold alignment). Then use the scaffold to align translated EST sequences (obtained from a tBLASTn report). This EST-scaffold alignment allowed the determination of the amino acid occupancy in the corresponding signature position in the EST. This operation will allocate ESTs into the different subfamilies. This knowledge-based approach revealed that loblolly pine contained only one type of 4-coumarate:CoA ligase, whereas coumaroyl-CoA reductase and cinnamyl alcohol dehydrogenase had each members of two different classes. One by-product of the method was the assemblage of full-length, biologically-meaningful contigs from the same subfamily. (Abstract shortened by UMI.)... |
