MOTIVATION: The prediction reliability of seven multiple alignment servers currently available on the Internet (ClustalW, MAP, PIMA, Block Maker, MSA, MEME and Match-Box) has been evaluated in terms of power (sensitivity) and confidence (selectivity). Therefore, the alignments obtained have been respectively compared to refined structural alignments for 20 families of related proteins with low levels of identity. RESULTS: Results clearly show that any powerful method remains reliable when the rate of identity falls. For some methods, power and confidence decrease linearly with the rate of identity, while other methods emphasize reliability at the cost of a lower power. Increasing the number of related sequences included in the alignment may either improve or decrease the quality of the predictions substantially. For some methods, the gain in power or in confidence is quite systematic; for others, the effect of the addition of homologous sequences is highly unpredictable. Extracting the consensus between two different methods may increase the overall confidence of the predictions tremendously. Our conclusions induce users of sequence alignment methods on the Internet to select the most suitable technique according to their requirements in terms of selectivity and sensitivity. AVAILABILITY: The aligned sequences of the 20 alignments of structure can be obtained automatically by sending the message 'send: cabios_tests.txt' by e-mail to '[email protected]'. CONTACT: [email protected]

MOTIVATION: The Match-Box software comprises protein sequence alignment tools based on strict statistical thresholds of similarity between protein segments. The method circumvents the gap penalty requirement: gaps being the result of the alignment and not a governing parameter of the procedure. The reliable conserved regions outlined by Match-Box are particularly relevant for homology modelling of protein structures, prediction of essential residues for site-directed mutagenesis and oligonucleotide design for cloning homologous genes by polymerase chain reaction (PCR). RESULTS: The method produces reliable results, as assessed by tests performed on protein families of known structures and of low sequence similarity. A reliability score is computed in relation to a threshold of similarity progressively raised to extend the aligned regions to their maximal length, up to the significance limit of matching segments. The score obtained at each position is printed below the sequences and allows a discriminant reading of each aligned region. AVAILABILITY: Sequences may be submitted to a Web server at http://www.fundp.ac.be/sciences/biologie/bms/+ ++matchbox_submit.html or sent by e-mail to matchbox/biq.fundp.ac.be (help available by just mailing help).

Original algorithms for simultaneous alignment of protein sequences are presented, including sequence clustering and within- or between-groups multiple alignment. The way of matching similar regions is fundamentally new. Complete matches are formed by segments more similar than expected by random, according to a given probability limit. Any classic or user-defined score matrix can be used to express the similarity between the residues. The algorithm seeks for complete matches common to all the sequences without performing pairwise alignment and regardless of gap weighting. An automatic screening delineates all the similar regions (boxes) that may be defined for a given maximal shift between the sequences. The shift can be large enough to allow the matching of any region of a sequence with any region of another one. It can also be short and used to refine the alignment around anchor points. The algorithm provides the most likely optimal alignment and a comprehensive list of the alignment dilemma. Duality between automatism and interactivity is provided. Depending on the problem complexity, a final alignment is obtained fully automatically or requires some interactive handling to discriminate alternative pathways.