Coordinate system BadRead

[TimoD73]

Describe the bug
BadRead reports in the header the coordinate of the read relative to the original genome.

@3d39fc0e-4092-3f08-05b3-1a216974b035 NC_000913.3,-strand,1264589-1277307 length=12713 error-free_length=12718 read_identity=99.898%

Often this coordinate is correct. However, when mapping all reads back to the reference genome, it becomes clear that the coordinates are often wrong.

When aligning the read with minimap, the true coordinates on the genome is: 3371966

3d39fc0e-4092-3f08-05b3-1a216974b035 16 NC_000913.3 3371966 60 6370M1I487M1D1772M1D1329M1I843M1D46M1D177M1D945M2D742M * 0 0

It took me 1,5 day to realize the coordinates were wrong.

[rrwick]

Yes, it's a common complaint about Badread! For reads on the negative strand, it's reporting the position in terms of the reverse complement reference sequence.

Since Badread is using a 0-based end-exclusive range (Python style), a negative-strand range of [start, end) would convert to a positive-strand range of [length - end, length - start), where length is the length of the reference.

For consistency, I don't want to change this behaviour unless I'm making a compatibility-breaking major release of Badread, but I'll leave this issue open as a reminder for if and when I do. Thanks!

Ryan