Skip to main content
Figure 1 | Algorithms for Molecular Biology

Figure 1

From: Multiple sequence alignment with user-defined anchor points

Figure 1

Possible mis-alignments caused by tandem duplications in the segment-based alignment approach (DIALIGN). We assume that various instances of a motif are contained in the input sequence set and that the degree of similarity among the different instances is approximately equal. For simplicity, we also assume that the sequences do not share any similarity outside the conserved motif. Lines connecting the sequences denote fragments identified by DIALIGN in the respective pairwise alignment procedures. (A) If a tandem duplication occurs in two sequences, the correct alignment will be found since the algorithm identifies a chain of local alignments with maximum total score. (B) If a motif is duplicated in one sequence but only one instance M2 is contained in the second sequence, it may happen that M2 is split up and aligned to different instances of the motif in the first sequence. (C) If the motif is duplicated in the first sequence but only one instance of it is contained in sequences two and three, respectively, consistency conflicts can occur. In this case, local similarities identified in the respective pairwise alignments cannot be integrated into one single output alignment. To select a consistent subset of these pairwise similarities, DIALIGN uses a greedy heuristic. Depending on the degree of similarity among the instances of the motif, the greedy approach may lead to serious mis-alignments (D).

Back to article page
\