Potential Kex2 cut site annotation

[darcyabjones]

We have internally been discussing including a Kex2 cut-site prediction component to Predector.
We're not sure that it's necessarily useful as a predictor of effectors, but it's a useful thing to have if you are looking at candidates.

This was proposed based on by the recent work by Megan Outram at ANU.

• https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010000
• https://nph.onlinelibrary.wiley.com/doi/full/10.1111/nph.17516

In that paper the cut-sites are predicted using this script: https://github.com/JanaSperschneider/Publications_Code/blob/master/2020_04_Tox3_LxxR_Paper/LxxR_Motif.py

We've been proposing a simpler regex solution that users will need to filter out themselves based on need. As far as we can tell, there aren't great methods for predicting cut-sites currently, so the general motif without is probably the best for manual evaluation.

We're proposing the three following regular expressions.

"[LIJVAP][A-Z][KRTPEI]R"
"L[A-Z][A-Z]R"
"[KR]R"

Output would be in the GFF, but also as a summarised version in the main table, with the format <match>:<start>-<stop>;....

I'd be interested to hear anyones thoughts on what we can do to make this better.
I've noted that the three regular expressions yield a lot of matches, so some filtering might be desirable, but we haven't yet decided on how best to do that.