Аннотация: Different subsets of protein-coding sequences may sometimes yield contradictory phylogenetic trees. These discrepancies may be based on different evolutionary histories of genes from a single genome: some sequences were acquired from a direct ancestor, some resulted from horizontal gene transfer, some could be present in a genome due to contaminations of its assembly. Changes in sequence evolutionary rates may also contribute to these effects. We have developed an automated approach to detect protein-coding sequences with regular and irregular evolutionary histories by comparing arrays of BLAST bit scores and hierarchical clustering of sequences with similar arrays of these scores. Our approach allows the discrimination of sequences from different organism sources in in silico metagenomic experiments. Also is surprisingly reveals the presence of large gene clusters in several sequenced genomes. The biological meaning of the latter finding is discussed.

Аннотация: We analyze recently proposed multimedia digital fingerprint- ing codes and derive first results on a new arising combinatorial problem.