Gubbins masking is incorrect currently

[watronfire]

See: mask_gubbins_aln.py

I thought that Gubbins just masked positions that were associated with recombination from all sequences of the alignment. But, I was digging into the above script that gubbins uses to mask the alingment and found that this probably isn't the case. Instead its using field #8 of the GFF, which lists the sequences that a recombinant regions is present in, and only masking the region from those sequences. This means that the current implementation in bacpage is incorrect, and will need to be replaced, but it means I can get rid of some cumbersome GFF to BED file conversion that I didn't trust anyway.

[watronfire]

I'll note that the current step of the pipeline uses the alignment in VCF format, so the linked script would have to be modified from running on a FASTA. This shouldn't be that hard but isn't as trivial as copy-pasting the script.