[Bioperl-l] BPlite using SimilarityPair

hilmar.lapp@pharma.Novartis.com hilmar.lapp@pharma.Novartis.com
Thu, 3 Aug 2000 20:55:25 +0100

There I have one question: why is "frac_identical" only in Similarity
not also in SimilarityPair where it sets both Similarity objects like
score, bits, etc. ??

     The reason is that score, bit-score etc refer to the whole alignment,
     whereas the fraction of identical residues depends on the denominator.
     You can use the total length of the alignment, or the number of query
     bases in the alignment, or the number of subject bases in the
     alignment, which are not necessarily the same in alignments with gaps.
     Depending on the penalty, gaps may account for a significant part of
     the total alignment length so that it wouldn't make sense to use it as
     the denominator, and if there are only short or no gaps, the choice of
     denominator doesn't really matter. That's why I left out the attribute
     for the total length as denominator.