ИСТИНА

The authors describe a new approach to reconstruct reliable protein phylogenies. It allows one to refine phylogenetic signal contained in primary sequence data by removing informational noise from the alignment. Criteria implemented in the method allow one to predict the number of noisy columns to be removed without actually reconstructing intermediate phylogenies, which can be a considerable advantage in cases of computationally intensive analyses. The algorithm is implemented in a computer program, which was used to analyze bulk molecular data from the COG database (NCBI).