spurious strand swap of introduced errors

[dlaehnemann]

In a test run with mdasim 2.1.1 on the example input data and using the testing pipeline, I got the following example_input.example_mdasim_out_prefix_Amplicons.example_mdasim_errors.cnt_comparison.tsv:

#substitution   log_cnt     aln_cnt
A to T      0       0
A to C      0       0
A to G      1       1
T to A      1       1
T to C      0       0
T to G      0       0
C to A      2       2
C to T      1       1
C to G      1       0
G to A      0       0
G to T      0       0
G to C      1       2

#ref_pos    mdasim_ref  mdasim_alt  aln_pos     aln_ref     aln_alt     ref=aln?
28241   C   G   28241   G   C    swapped 
43784   G   C   43784   G   C    true 
176217  T   A   176217  T   A    true 
181153  C   T   181153  C   T    true 
189621  C   A   189621  C   A    true 
196027  C   A   196027  C   A    true 
205366  A   G   205366  A   G    true

The position is correct, but strands seem to swapped between what mdasim reports itself and what the alignment based testing pipeline finds.

But right now our focus should be on epyscope, so I'm only documenting this here for the future and as a caveat for epyscope output. But I think we should only fix this once we have come further with epyscope. Also, this is a further point in favour of reimplementing mdasim with proper unit tests....

[dlaehnemann]

Also, we should first try to reproduce this independently before tackling it as a real issue, to make sure this is not some mdasim version conflicts on my machine. And we should double check it's a problem of mdasim and not of the alignment based testing pipeline...