MOSAIK fails to align to ambiguity codes in reference

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Align reads to reference genome with ambiguity codes
2.
3.

What is the expected output? What do you see instead?

The reads I'm attempting to align have been generated directly from the 
reference genome for each alternative defined by the ambiguity codes (for 
example, if in the reference there is "S", I've generated two reads with both a 
"G" and "C" in that position).

What version of the product are you using? On what operating system?

Latest git clone 2.1.52, on Mac OS X.

Please provide any additional information below.

I used the following parameters: -om -hs 14 -mm 0 -m unique

Original issue reported on code.google.com by krai...@gmail.com on 16 Feb 2012 at 5:01

Attachments:

• toy_amb.fa
• toy_single_bp50_error0_tiled.fa

[GoogleCodeExporter]

The reads are named with metadata from where they were generated from, e.g. 
">HWI_2L_toy_0_50_refPos" implies that it originated from chromosome 2L (only 
first 10kb of this chromosome is included in toy_amb.fa), the numbers are 
BED-formatted start and stop points for the read according to it's origin, and 
refPos implies it's from the reference allele in the posititive orientation 
(altPos --> alternative allele in positive orientation; refNeg --> reference 
allele in reverse complemented orientation; and altNeg --> alternative allele 
in reverse complemented orientation).

Original comment by krai...@gmail.com on 17 Feb 2012 at 3:45

[GoogleCodeExporter]

Changing the mismatch parameter to -mm 1 seemed to fix this, but according to 
the manual, "if the reference genome has an ‘M’ allele, a read position 
with an ‘A’ or ‘C’ at that location will be scored as a match", but 
clearly that is not the case.

Original comment by krai...@gmail.com on 17 Feb 2012 at 7:26