AuberGene - a sensitive genome alignment tool

AuberGene is a genome alignment program which explores the idea of transitivity of local alignments. Even though it uses pairwise alignments generated by other genome alignment programs, it produces more sensitive and specific alignments by integrating the information from multiple genomes.

Example: We tested the program on the greater CFTR region. See the coverage of the region in UCSC Genome Browser.

The UCSC Genome Browser tracks show the collective alignment (alignment based on information from all the intermediate species) between human and mouse, with the help of 11 intermediate species: baboon, cat, chicken, chimp, cow, dog, fugu, pig, rat, tetraodon, zebrafish.

• Track Collective human-mouse alignment coverage (in blue). Each feature (fragment of sequence) is annotated with the number of transitive alignments supporting the collective alignment. This number ranges from 1 to 12: all 11 transitive alignments and the direct one (the darker the feature, the more evidence - transitive alignments - was available).
• Track Direct human-mouse alignment coverage (in black). The coverage of the human sequence with matches, according to the human-mouse alignment generated by BLASTZ (see Resources below for the alignment).
• Track Direct human-mouse alignment errors (in red). This track shows fragments of human sequence which were aligned incorrectly with the direct alignment. Features of this track point at fragments of sequence for which there are indications from at least three species (three transitive alignments) suggesting different matches with mouse sequence than indicated by the direct human-mouse alignment.

Reference:

• Szklarczyk, R., Heringa, J., AuberGene - a sensitive genome alignment tool, Bioinformatics, 15 June 2006; 22: 1431 - 1436. Preprint

Resources:

• source (aubergene.tar.gz, readme.txt).
• collective and transitive alignments generated by the program (tar.gz)
• data used in the program: Thomas JW et al. Nature 424, 788-793, 14 August 2003 (publication, supplementary information, alignments and sequences). See also Comparative Vertebrate Sequencing Project (web page)
• BLASTZ (web page)

Contact: ibivu [at] cs.vu.nl