Sigma-2: Multiple sequence alignment of non-coding DNA via an evolutionary model

BMC Bioinformatics

Table 1 Performance in aligning yeast sequence

Program	dataset	Matches per base^a	dataset	Matches per base^a	Difference
Sigma-2	orthologous	1.9893	shuffled	0.0031	1.9862
Sigma-1.1.3	orthologous	1.8688	shuffled	0.0050	1.8638
FSA	orthologous	2.4695	shuffled	0.1465	2.3230
Dialign-TX	orthologous	2.7498	shuffled	0.4539	2.2959
Pecan	orthologous	3.0234	shuffled	0.4430	2.5804
Mavid	orthologous	3.3181	shuffled	2.8248	0.4933
T-Coffee	orthologous	3.5582	shuffled	3.3495	0.2487
Clustal-W	orthologous	3.6202	shuffled	3.7517	-0.1315
KAlign	orthologous	3.7480	shuffled	3.8434	-0.0954
MLagan	orthologous	3.2956	shuffled	2.7082	0.5874
Muscle	orthologous	3.4541	shuffled	3.1901	0.2670
PCMA	orthologous	3.4822	shuffled	2.8941	0.5881

Performance in the yeast benchmark, described in the text, of 12 programs (including 2 versions of Sigma). 947 genes were selected, each of which had 1000 bp of non-coding upstream sequence in S. cerevisiae and four other species. Each upstream sequence and its four orthologues were aligned (dataset "orthologous"). In addition, 947 "scrambled" files were prepared each of which contained sequence from each of the five species, including no orthologous sequences, and these were aligned (dataset "scrambled"). "Matches per base" indicates the average number of nucleotides in other species that each nucleotide in the input data was aligned with (so its theoretical maximum is four). The difference between the "orthologous" and "scrambled" numbers is a measure of how discriminative the program is to genuine orthology.

ISSN: 1471-2105