Skip to main content
Figure 6 | BMC Bioinformatics

Figure 6

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

Figure 6

Progressive alignment via "sequence fragments". Progressive alignment of input sequences. (a) Each sequence is initially its own fragment. All fragments are compared, pairwise, to find the best gapless local alignments. In this case, the grey boxes indicate two possible local alignments, and the underlined sequences show a third possible local alignment that is of lower significance, and also conicts with the first two. (b) The first local alignment is carried out by "fusing" two portions of seq1 and seq2 into one fragment. (c) The second local alignment is carried out, resulting in 9 sequence fragments at this point. However, the third local alignment, the underlined sequence in part (a), now conflicts with the existing alignments and cannot be performed. This is ensured by consistency conditions, enforced using a labelling scheme described in detail in [21] and illustrated here: each sequence in each fragment has a label (a string representation of a floating-point number) that is augmented whenever two fragments are "fused", in such a way that the numeric value of the label always increases from left to right along any sequence. Here, in (c), the fragment on seq3 labelled "0.2" can be aligned with the fragment on seq1 labelled "0.2", but not with any other fragment on seq1. (Figure reproduced from ref. [21])

Back to article page