Spatial thinking in genomic ecology or, Why are codons correlated?

Genomes from three prokaryotes were downloaded from The Institute of Genomic Research (TIGR) – a cyanobacteria (Synechococcus elongata), an enterobacter (Escherichia coli), and a radioresistant extremophile (Deinococcus radiodurans).  It was hypothesized that the phylogenies of these three represent the most, intermediate, and least ancient evolutionary histories, respectively, because of their environmental functionalities - ecological niches.  We further hypothesized that bioinformatic evenness along the DNA sequence should be indicative of the ecological complexity of the organism.  Consequently, we measured the 5’ – 3’DNA sequence (Shannon-Wiener) diversity among the proportions of the four nucleotide bases (A, G, C, T).  A coding unit (CU) was defined as a contiguous sequence of 1,000 codons and a diversity measure was calculated for each CU.  Data displays of per CU base proportions and diversity measurements show that the highest average diversity occurred in E. coli and the lowest occurred in D. radiodurans, inconsistent with the hypothesis.  We expected E. coli to have the least genomic diversity because of its niche in a homeotherm gut.  Variograms from spatial statistics illustrated spatial correlation among CUs and showed that diversity measures between “nearby CUs” in sequence are more like each other than diversity measures between CUs separated by some sequence distance.  As the separation distance increases, the spatial correlation between diversity measures decreases.  The minimum separation distance between CUs such that the spatial correlation has declined to zero varies –  1.8, 3.3, and 2.4, among the three species, respectively.  These findings suggest that examining the spatial characteristics of a genome can provide both evolutionary and ecological perspectives.  Our data indicates that perhaps E. coli with both commensal and free-living forms requires more genomic diversity than the other two species to survive in changing environments.  While the low diversity of S.elongata, and D. radiodurans suggest niche specialization and long evolutionary trajectories.