Quality Assessment of Multiple Alignment Programs

ArenewP interest in the multiple sequence alignment problem has given rise to severalnew algorithms. In contrast to traditional progressive methods, computationally expensive score optimization strategies arenow predominantly employed. We systematically tested four methods (Poa, Dialign, T-Coee and ClustalW) for the speed and quality of their alignments. As test sequencesw used structurally derived alignments from BAliBASE and synthetic alignments generated by Rose. The tests included alignments of variable numbers of domains embedded in random spacer sequences. Overall, Dialignwa the most accurate in caseswse low sequence identity,went T-Coffee we in cases wes high sequence identity. The fast Poa algorithmwg almost as accurate,wcur ClustalW could compete only in strictly global caseswse high sequence similarity. # 2002 Published by Elsevier Science B.V. on behalf of the Federation of European Biochemical Societies