Quantifying the Displacement of Mismatches in Multiple Sequence Alignment Benchmarks - Fig 1

<p>(A): CS score compares columns in the reference to corresponding columns in the query; SP score compares all aligned pairs in the reference to the corresponding pairs in the query. (B): Toy alignments to illustrate the difference punishment for alignment shifts in CS, SP, and SPdist (<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0127431#pone.0127431.e009" target="_blank">Eq 8</a>) scores. All sequences in this contain a small sequence motif H-x-D-E which has to be aligned correctly. Both query A and B contain the same number of mismatches (indicated by the same SP score for both queries), the sole difference between these two queries is that sequence 2 in query B is shifted five positions further. In this situation one might argue that although both alignment are wrong, i.e. the sequence motifs are not aligned properly, query A is a slightly better alignment than B as the motifs in query A is positioned closer to each other than in query B.