Network Paradigm for Molecular Evolutionary Analysis of Metabolic Pathways

In order to understand the role of evolution in biological function, it is insufficient to study genes in isolation. Previous studies restricted their analyses to one or two genes, often due to data paucity and sequencing costs. Now, inexpensive high-throughput sequencing data permits systems biologists to study evolution of entire protein/gene networks in parallel. To this end, we conducted an evolutionary analysis of several human metabolic pathways. We hypothesized that selection pressures distribute randomly in the pathway. We drew our pathway maps from the EHMN pathway database, and we downloaded our sequences and orthology predictions for primate and rodent genomes from the Ensembl database. We then analyzed multiple sequence alignments of orthologous coding sequences with nucleotide substitution rates (dN/dS). Significant rates were mapped onto network representations of our pathways. In our network analysis, we studied the rate variance with our network measures (centrality, betweenness, etc). Preliminary results fail to refute the hypothesis; we observe no general distribution of selection pressures, and nonrandom distributions are particular to individual pathways. Our results question previous studies claiming a general pattern of evolutionary pressure in genetic networks and suggest the need for a more sophisticated graph-theoretic toolbox.