recommendation for non-stranded reads?

[alexpan82]

Hello!

I am working with VDJ data generated from a non-stranded RNA-seq library.

Do you have any recommendations on running partis efficiently on such a data set besides running it again on the revcomp?

On that note, why does partis only accept plus-strand reads?

Thanks, Alex

[psathyrella]

Partis doesn't do any kind of preprocessing (i.e. paired read joining, primer masking, error correction, etc), and getting everything plus-stranded makes more sense to us as part of preprocessing, since it's a trivial alignment step that doesn't require the improved accuracy of partis's smith waterman or hmm methods. That said I did recently add bin/split-loci.py as a simple script to split apart h/k/l, and it'd be easy to make a similar helper script for converting to all plus-stranded.

[psathyrella]

I did it!

https://github.com/psathyrella/partis/blob/dev/bin/split-loci.py#L27

[psathyrella]

For the benefit of future searchers -- partis now also accepts the --reverse-negative-strands option (although it's just running split-loci.py)

https://github.com/psathyrella/partis/blob/dev/docs/quick-start.md