Probabilistic reconstruction of ancestral protein sequences

Abstract

Abstract. Using a maximum-likelihood formalism, we have developed a method with which to reconstruct the sequences of ancestral proteins. Our approach allows the calculation of not only the most probable ancestral sequence but also of the probability of any amino acid at any given node in the evolutionary tree. Because we consider evolution on the amino acid level, we are better able to include effects of evolutionary pressure and take advantage of structural information about the protein through the use of mutation matrices that depend on secondary, structure and surface accessibility. The computational complexity of this method scales linearly with the number of homologous proteins used to reconstruct the ancestral sequence.

Accuracies of ancestral amino acid sequences inferred by the parsimony, likelihood, and distance methods

“…After completion of our simulation work, Koshi and Goldstein (1996) published a Bayesian method for ancestral sequence reconstruction, which is similar to Yang et al's (1995) method. Generally speaking, Koshi and Goldstein's method is expected to perform well, since it also takes into account both the branch lengths and the substitution pattern.…”

“…Although the simultaneous reconstruction gives the same results as those obtained node by node in most cases, they are not always the same (Yang et al 1995). Koshi and Goldstein (1996) also conducted some computer simulations to investigate the reliability of their method. The results obtained are generally consistent with our results.…”

“…Several statistical methods have been developed to infer the ancestral amino acid sequences from the sequences of present-day species when the phylogenetic relationship is known (Eck and Dayhoff 1966;Libertini and Di Donato 1994;Schluter 1995;Yang et al 1995;Koshi and Goldstein 1996). Among these methods, the maximum-parsimony (MP) method has been used most frequently.…”

Simple and accurate estimation of ancestral protein sequences

“…The most widely used methods of ancestral state reconstruction have been parsimony (12)(13)(14) and maximum likelihood (15)(16)(17)(18), although distance-based methods have also been described (19). MRBAYES (20) implements the Bayesian method of phylogenetic tree reconstruction, and the most recent version, version 3.1, now implements the reconstruction of ancestral states by the Bayesian method.…”

“…It is therefore essential to estimate the sequences of ancestral proteins by the most accurate means available. Several studies have used computer simulations to compare the accuracies with which different methods reconstruct ancestral sequences (17)(18)(19); however, the simulated sequences in those studies were not very realistic in that they did not include insertions and deletions. A recently described sequence-evolution simulation program, EVOLVEAGENE (21), is biologically realistic in that it separates mutation and selection and in that the mutations not only include base substitutions but also insertions and deletions.…”