search

innate2adaptive / decombinator

Decombinator v5: fast, error-correcting analysis of TCR repertoires
https://pypi.org/project/decombinator/
MIT License
22 stars 8 forks source link

Downstream analysis #30

Closed benjytan88 closed 9 months ago

benjytan88 commented 9 months ago

Hi, I recently got a dataset from our collaborator who processed bulk TCR-Seq using Decombinator. I read in your publication that the data is AIRR-compatible but when I tried loading it in R using immunarch, I couldn't do so as some columns are missing. Can you recommend any packages / programs to perform downstream analysis for your data (i.e. calculating clonality, diversity, repertoire similarities, etc...)?

JamieHeather commented 9 months ago

I don't use any third party tools downstream so I can't recommend any - perhaps the other authors can.

However I just checked the current AIRR-C schema and all the required fields are present in Decombinator output, assuming your collaborator used an up-to-date version. I can't get immunarch installed on the current version of R, so I can't test this myself: what fields does it say are missing? Have you tried adding those columns yourself?

benjytan88 commented 9 months ago

Hi @JamieHeather , Thanks for your quick reply! I never thought of adding the columns myself as I have the assumption that empty columns won't work. But, I tried it as per your suggestion and apparently it works! FYI, the missing columns are cdr1, cdr2, fwr1, fwr2, fwr3, fwr4, fwr1_aa, fwr2_aa, fwr3_aaand fwr4_aa. I'm going to close this now. Thanks again!

JamieHeather commented 9 months ago

Happy to help, and thanks for letting us know!

In case it's helpful in future, it's worth knowing that the AIRR-C schema often basically forces a bunch of empty columns, as a lot of the 'required' columns (which are supposed to always be present) aren't relevant to all analyses. There's also a bunch of additional 'optional' columns (which includes all of those you list) which don't need to be present - so this is maybe just a case of immunarch being a little over-zealous in its expectations.