[Seqan-dev] Cannonical Alignment Row Ordering

[Brett Bowman <bnbowman@gmail.com>]

Is there a standardized way to order sequences in Alignment objects in SeqAn?

Reason being that the code in "align/evalulate_alignment.h" appears to take Row #0 as the reference:

"""

    // Get iterators.

    TGapsIter it0 = begin(row(align, 0));

    TGapsIter itEnd0 = end(row(align, 0));

    TGapsIter it1 = begin(row(align, 1));

    TGapsIter itEnd1 = end(row(align, 1));

    for (; it0 != itEnd0 && it1 != itEnd1; ++it0, ++it1)

    {

        if (isGap(it0))

        {

            stats.numDeletions += 1;

        }

        if (isGap(it1))

        {

            stats.numInsertions += 1;

        }

   }

"""

While the code for writing alignments in "align/align_base.h" prints rows in numerical order, which puts the query below the reference, in an inversion of standard practice:

"""

  Reference ATCTCTCTC-A-ACAACAA--CAACGGAGGAGG-AGGAAAAGAGAGAGAT

  Align        |||||||||*|*|||||||**||||||||||||*||||||||||||||||

  Query      ATCTCTCTCAACACAACAACGCAACGGAGGAGGAAGGAAAAGAGAGAGAT

"""

I think this also runs counter to how a number of SDP algorithms are traditional written, where seqan puts the Horizontal sequence first, which IIRC is traditionally the query not the reference (e.g. Baker & Giancarlo (2002), http://www.sciencedirect.com/science/article/pii/S0196677402912149)
"""

    assignSource(row(alignment, 0), sequenceH);
    assignSource(row(alignment, 1), sequenceV);

    Score<int, Simple> scoringScheme(2, -1, -2);

    int result = bandedChainAlignment(alignment, seedChain, scoringScheme, 2);
"""

-Brett