You can clone with
I have a read at position 0 and cigar 1S100M. So the soft-clip base is actually at -1. In the soft clip condition in AlleleParser.cpp (around line 1627), the third argument in makeallele becomes the largest unsigned int and causes a substring function to fail. I wouldn't think a soft-clip base should be counted towards the allele at a position.
This sounds right. I'm working on a solution.
improve scaling of probabilities, resolve #42
When using --allele-balance-priors, --binomial-obs-priors, scale
probabilities according to the number of possible observation
Resolve #42 by preventing use of soft-clipped sequence at the beginning
of the reference.