abPOA puts big gap in middle of sequence

[glennhickey]

I'm aligning a short sequence to a very long one, and they look like

ATATA
ATATATATATA<...>A

When I align them with abpoa, it puts the gap in the middle

ATAT------------<...>A
ATATATATATA<...>A

But I'd much rather have an alignment like

ATATA-----------<...>
ATATATATATA<....>

I can understand that, strictly speaking, with the global alignment scoring model, there isn't much difference between these two scenarios, in practice I think the second one is always preferred. Do you think it would be possible to add some way to left shift or favour gaps at the ends in global alignment mode?

Thanks, as always.

to reproduce

wget http://public.gi.ucsc.edu/~hickey/debug/abpoa_fail_mar23.fa
abpoa abpoa_fail_mar23.fa -m 0 -r 1

[yangao07]

A quick fix has just been pushed. Try it out @glennhickey

[glennhickey]

Thanks!! this does indeed solve some of the simpler cases I've found. Will test it on a larger dataset now.