Amino Acid Distributions and the Effect of Optimal Growth Temperature
We perform an exhaustive analysis of genome statistics for organisms, particularly extremophiles, growing in a wide range of physicochemical conditions. Specifically, we demonstrate how the correlation between the frequency of amino acids and their molecular weight, preserved on average, typically decreases as optimal growth temperature increases. We show how the relation between codon degeneracy and amino acid mass is enforced across these organisms. We assess the occurrence of contiguous amino acids, finding several significant short words, often containing cysteine, histidine or proline. Typically, the significance of these words is independent of growth temperature. In a novel approach, first-passage distributions are used to capture correlations between discontiguous residues. We find a nearly universal exponential background that we relate to properties of the aforementioned individual amino acid frequencies. We find this approach reliably extracts correlations that depend on growth temperature, some of which have not been previously characterized.