Immune repertoire reconstruction from bulk RNAseq data

[npechl]

Description of feature

Hi everyone,

I am not sure if there is already a discussion going on regarding this issue (I didn't locate one).

We've been utilizing nf-core/rnaseq for our analyses, and I believe it would be incredibly valuable to have a new module (or maybe a subworkflow) for immune repertoire reconstruction from bulk RNA-seq data.

In previous projects, I've worked with TRUST4, which takes as input alignment files (.bam) and corresponding reference files and generates very useful outputs for immunity analysis.

Best,
Nikos

[MatthiasZepper]

Wouldn't Airrflow be the more appropriate pipeline here?

I believe, this is out of scope for the general rna-seq pipeline, but I suggest discussing with the airrflow developers, if that tool would be a good fit for the pipeline.

[npechl]

Hi @MatthiasZepper,

Airrflow takes B-cell or T-cell repertoire sequencing data as input, whereas TRUST4 uses alignment files from bulk or single-cell RNAseq. This is the main reason I proposed to include it here. To be honest, I am not sure where it fits exactly.

There is also a complete pipeline (I think it is written in SnakeMake but not so versatile) based on RNAseq for tumor immunity analysis: https://liulab-dfci.github.io/RIMA/.

[MatthiasZepper]

I see. I recommend asking about this rather on the nf-core Slack, as it is more straightforward discussing there, the #rnaseq-dev channel is probably the best place then.

Personally, I am somewhat cautious adding further specialized functionality to the pipeline, since it is already a quite complex one and a further pipeline sprawl to Sarekian dimensions is probably not too advisable. But if there is indeed no special preprocessing required for that tool and you volunteer to implement it as a subworkflow that can be easily integrated, I also have no hard objections against it either.

Based on the RIMA pipeline that you give as an example, Oncoanalyser could also be suitable. It shares some tools with RIMA, but lacks the immunity component, so Trust4 could be a welcome addition?

[npechl]

Perfect! Thank you for the answer!

I would recommend closing this issue and continuing the discussion on Slack.