These two programs implement ideas described by Z. Zhang, P. Berman, T. Wiehe and W. Miller in the paper "Post-processing long pairwise alignments", Bioinformatics 15, 1012-1019. In particular, all experiments described therein where performed using them. However, be warned that they are in no way intended to be "production quality" software. Rather, they seemed (just barely) good enough to validate some theoretical conclusions. The program called "decom" simply converts one alignment file to another, by removing alignment segments of very low score. The program called "decomT" produces the tree of "all decompositions", as descibed in the paper. The programs assume the alignment-scoring parameters match=10, mismatch=-10, gap of length k = -60 - 2k but these can be reset on the command line, as in decom alignments I=-8 (I is the penalty for transitions, V is for transversions). To compile them, you will need to obtain and install the "blastlib" library that accompanies our sim4 program. Also, alignments need to be in the format used by our alignment tools, such as the sim (AKA sim96) program. One set of test data is included here, consisting of the files: elegans - genomic sequence data from C. elegans briggsae - genomic sequence data from C. briggsae alignment - an alignment of the elegans and briggsae sequences computed using the Smith-Waterman criterion for local alignments decom.out - file computed by the command decom alignment Y=2000 > decom.out The Y parameter limits the amount that a segment of the alignment can score, in this case to -2000. Alignments here are scored match=10, mismatch=-10, gap of length k = -60 - 2k. dotplot.pdf - PDF file giving a dotplot view of the files called "alignment" and "decom.out". decomT.out - file computed buy the command decomT alignment T=1 > decomT.out The T=1 flag asks for a particular format of output. Here, we get a file starting with 8146 12331-58917 ---> 12331-30146 31489-32158 32284-32359 33208-33333 34960-35167 36856-58917 2858 12331-30146 ---> 12331-17556 19702-19806 20344-30146 1996 12331-17556 ---> 12331-15559 16204-17556 1110 12331-15559 ---> 12331-13182 13768-15559 This shows the vertices at the top of the tree. For instance, the root corresponds to score 8146, where five subalignments are joined to give an alignment covering positions 12331-58917 of the first sequence. The left-most child of the root, corresponding to positions 12331-30146, is formed when three subalignments coalesce at score 2858, and so on. Thus, each node of the tree is represented by a line in the decomT output. The X paramter can be set to a high value of one wants only the top part of the tree; nodes of score < X are not printed.