IMGT seqence/germline build in ConvertDb doesn't seem right

[Jason Vander Heiden]

Look into the behavior of the -r argument to ConvertDb-airr and ConvertDb-changeo. It doesn't seem correct, at a glance.

[Jason Vander Heiden]

May be best to simply exclude the -r functionality from the 0.4.1 release.

[Jason Vander Heiden]

Schedule for v0.4.3.

And maybe move it out of ConvertDb-changeo and ConvertDb-airr into a separate subcommand. It's kind of clunky to trying to combine the gapping with flexible (airr -> airr and airr -> changeo) I/O.

[Jason Vander Heiden]

Maybe it would be better to make this a MakeDb-airr command, given the decision to switch the core format to the AIRR standard. The basic functionality would be to take in an AIRR Rearrangement file and do a little standardization within that scope. For now, I see this only entailing:

1. Adding IMGT gaps to sequence_alignment and germline_alignment and modifying the appropriate germline start/end fields.
2. Redefining productive and junction_aa based on the junction frame.
3. Filling in positional fields required by CreateGermlines from the _cigar fields.
4. Add the locus field if needed.

All as options, because the input AIRR file could come from any source, not just IgBLAST.

[Scott Christley]

Is this mainly dealing with different fields names, or does new functionality need to be written? I would also vote for the IMGT gapped alignments to be put in new fields instead of overwriting the existing field content.

[Jason Vander Heiden]

Some new functionality is needed, which is why it’s taking me so long to get around to it. The bottleneck testing the existing IMGT-gapping methods on different outputs. The field renaming stuff and conversion from _length to _end is all already in ConvertDb.