Estimation and Reliability of Molecular Sequence Alignments

Abstract

SummaryThe problem of estimating the relatedness of a pa1r of biological sequences IS addressed. A stochastic model of sequence evolution is described that allows insertion and deletion as well as replacement of amino acid residues (or substitution of nucleotides) over time. An expectation-maximization (EM) algorithm that obtains maximum likelihood estimates of the model parameters is introduced. The method assumes the sequences are related by descent from a common ancestor but the alignment (i.e., the precise evolutionary correspondence between residues in each sequence) is unknown. Results from the E-step of the EM algorithm are used to assess the likelihood that any pair of residues are related by direct descent from a common ancestor.

Neighboring base composition and transversion/transition bias in a comparison of rice and maize chloroplast noncoding regions.

“…The 18 regions were also aligned by an ML algorithm with a 95% arc probability (12,13). This method includes only those sections (or arcs on a directed graph) for which the sum of the probabilities of all alignments that contain that arc is >95% of the sum of probabilities of all possible alignments (13).…”

“…The substitutions remaining after the application of these rules to the alignments were then analyzed. The second method was to use a maximum likelihood (ML) alignment algorithm with an arc probability of 95% (12,13).…”

Surveying Saccharomyces Genomes to Identify Functional Elements by Comparative DNA Sequence Analysis

“…a conserved hexamer is expected to occur by chance about once in every eight promoters on average. We realize that these "back of the envelope" calculations must be interpreted cautiously because it is unlikely that even a probable optimal multiple sequence alignment will correspond precisely to the biological events that generated the sequences being aligned (Thorne and Churchill 1995), but in any case, four-way alignments seem to be the minimum required for efficient recognition of conserved regulatory sequences.…”

A phylogenetic study of the Lecanora rupicola group (Lecanoraceae, Ascomycota)

“…However, a maximum likelihood distance method that retains the full information of sequences is also available. The program Statalign (Thorne, Kishino & Felsenstein 1991, Thorne & Kishino 1992, Thorne & Churchill 1995 does not require an alignment, it calculates pairwise maximum likelihood distances among all possible pairs of sequences. The resulting distance matrix, containing the distances and their standard deviations, is subjected to tree inference using the program Modfitch35, which is a modified version of the Fitch program (Felsenstein 1989), included in the Statalign package.…”