A Study Using Sequence Comparison to Investigate the Molecular Evolution of Mitochondrial tRNA Genes

In investigating the origins of life and molecular evolution, it is very important to study structure and functions that are common to distantly related species and to analyze the corresponding gene sequences, which are called “molecular fossils.” We therefore developed a computational method [1] based on principal component analysis (PCA) and multidimensional scaling analysis (MDS), and in the present study have used it to detect bases characterizing specific sequences mitochondrial transfer RNA. Our methods first classify the sequence in a genomic database into groups by PCA of multiple sequence alignment: Gene sequences are represented as vectors in a generalized sequence space, and groups of similar sequences are revealed when these vectors are projected onto a lower-dimensional sequence space. The distribution of bases is then compared with the distribution of sequences by using MDS, in which the bases of each sequence are projected individually onto the same sequence space. This makes it possible to identify bases characteristic of each group. Applying this method to the sequences of all the mitochondrial tRNA genes, we expect to detect not only bases that are always conserved but also bases that are often conserved.