work with guideSet object in python

[abearab]

Hi, I'm a big fan of crisprVerse and I found the GuideSet object very useful in my projects.

The GuideSet object is an extension of a GenomicRanges object that stores additional information about gRNAs.

However, I don't see an existing standard package to work with such data in python. I wasn't sure where to open this issue but I thought this might be a good place to start the conversation. Looking forward to hear your thoughts.

cc @Jfortin1

[jkanche]

Both @Jfortin1 and I have been thinking about this for a while. Providing the representation is probably straightforward, but supporting downstream analytical functionality might be a bit tricky. My understanding is we need some implementation or equivalent existing packages that provide similar functionality as R/Bioc's BioStrings: #3.

What methods from the guideset do you more frequently use? May be thats a good starting point for us to think of how we can port that functionality.

[Jfortin1]

@abearab The first thing we are planning to do is port over the GuideSet representation, and probably most of crisprBase as well, and some of the genomics arithmetic functions in crisprDesign

[abearab]

Hey @Jfortin1 – any updates in this regard? I'm hoping to launch a project and I would love to keep everything in python!

[abearab]

Thank you @jkanche and @Jfortin1 for your quick responses!

What methods from the guideset do you more frequently use? May be thats a good starting point for us to think of how we can port that functionality.

The way I think about this problem is this (all these are my naive understanding):

1. A user would pick a CRISPR technology, mainly the PAM sequence and spacer length are the key parameters to start.
2. Then user aims to search for potential guide sequences for a given genomics coordinates using a reference genome (potentially GenomicRanges object). Alternatively, a user may already have a list of guide sequences (for instance known pooled-library) so it's helpful to enable loading that into a guideset object for downstream tasks.
3. As @Jfortin1 mentioned, a tool like crisprDesign would help to calculate on on-target / off-target scores. I agree this might need fresh implementations.
4. For combinatorial platforms, guide pairs are also something I can imagine a clear use case on my end.