%0 Journal Article
%T Pathological rate matrices: from primates to pathogens
%A Harold W Schranz
%A Von Yap
%A Simon Easteal
%A Rob Knight
%A Gavin A Huttley
%J BMC Bioinformatics
%D 2008
%I BioMed Central
%R 10.1186/1471-2105-9-550
%X We used concatenated protein coding gene alignments from microbial genomes, primate genomes and independent intron alignments from primate genomes. The Taylor series expansion and eigendecomposition matrix exponentiation algorithms were compared to the less widely employed, but more robust, Pad¨¦ with scaling and squaring algorithm for nucleotide, dinucleotide, codon and trinucleotide rate matrices. Pathological dinucleotide and trinucleotide matrices were evident in the microbial data set, affecting the eigendecomposition and Taylor algorithms respectively. Even using a conservative estimate of matrix error (occurrence of an invalid probability), both Taylor and eigendecomposition algorithms exhibited substantial error rates: ~100% of all exonic trinucleotide matrices were pathological to the Taylor algorithm while ~10% of codon positions 1 and 2 dinucleotide matrices and intronic trinucleotide matrices, and ~30% of codon matrices were pathological to eigendecomposition. The majority of Taylor algorithm errors derived from occurrence of multiple unobserved states. A small number of negative probabilities were detected from the Pad¨¦ algorithm on trinucleotide matrices that were attributable to machine precision. Although the Pad¨¦ algorithm does not facilitate caching of intermediate results, it was up to 3¡Á faster than eigendecomposition on the same matrices.Development of robust software for computing non-reversible dinucleotide, codon and higher evolutionary models requires implementation of the Pad¨¦ with scaling and squaring algorithm.The dynamics of genetic divergence are typically modelled as a Markov process where the rates of exchange between discrete sequence states are described by rate matrices. Discrete- or continuous-time Markov processes employ different, but related, rate matrices. The former involve a substitution matrix that specifies the probabilities of substitution between sequence states in a discrete period of time (P(t), [1]). The continuous-tim
%U http://www.biomedcentral.com/1471-2105/9/550