This move is dangerous as SAM printing is very complex, but it will benefit in
the long run. The planned change will reduce the redundancy, improves clarity
and most importantly makes it much easier to output multiple primary hits in an
optional tag.
The changes after r317 aim to improve the performance and accuracy for very
long query alignment. The short-read alignment should not be affected. The
changes include:
1) Z-dropoff. This is a variant of blast's X-dropoff. I orginally thought this
heuristic only improves speed, but now I realize it also reduces poor
alignment with long good flanking alignments. The difference from blast's
X-dropoff is that Z-dropoff allows big gaps, but X-dropoff does not.
2) Band width doubling. When band width is too small, we will get a poor
alignment in the middle. Sometimes such alignments cannot be fully excluded
with Z-dropoff. Band width doubling is an alternative heuristic. It is based
on the observation that the existing of close-to-boundary high score
possibly implies inadequate band width. When we see such a signal, we double
the band width.
The old version wastefully extends many seeds contained in an aligned region
found before. While this wastes little time for short reads, it becomes a
serious defect for long query sequences.
This is an attempt to fix this problem, but more tuning are needed.
I really hate this: complex and twisted logic for a nasty scenario that almost
never happens to short reads - but it may become serious when the reference
genome consists of many contigs.
On toy examples, the code seems to work. Don't know if it really works...
This is important for bwa-mem which does local alignment. A short exact match
is worse than a long inexact match. Also fixed a bug in approximating mapping
quality.
Heng Li