Fast and Sound Two-Step Algorithms for Multiple Alignment of Nucleic Sequences

Abstract

We report here on our work on multiple alignment of biological sequences. The observation of actual alignments has lead us to formulate heuristics from which we have derived new and efficient algorithms. These two algorithms are very fast and give reasonably good results. They work in two steps: blocks are found in the sequences, and then the sequences are aligned between the blocks. Strong and sound hypotheses made at every step, as well as a new representation of blocks, yield high efficiency of the resulting algorithms.