Very nice Ira! What else are you considering addin...
# riboseq
j
Very nice Ira! What else are you considering adding?
RUST is another nice aspect of your pipeline but I think it will be better for our lab to handle that one as it was originally developed here. I wrote a better CLI interface for it and can try make a module once I submit my thesis but that is still a few weeks away
👍 1
🎉 1
i
also thinking it's way better to use the original RUST tool rather than the R implementation from my scripts - we could use an existing version and just update the container once you make a new release, unless you are making major changes to its usage in which case it's def worth waiting. apart from RUST was thinking ribocutter - it's a tool that finds contaminating sequences in sequencing experiments and even designs guides to deplete them. maybe it's not the most widely used rRNA depletion method, but it provides some nice QC also besides e.g. just quantifying rRNA, tRNA %, and it could just be optional https://www.biorxiv.org/content/10.1101/2021.07.14.451473v1.full apart from that not sure what to prioritise - if there are specific plots from riboseq-flow that you think would add value let me know - it is possible some of them might be redundant to QC we already have? PS good luck w your thesis!!
j
When Patrick wrote and published RUST it was released as a collection of scripts. I pretty much just added argparse. See here: https://github.com/JackCurragh/RUST. It is on PyPI so hopefully won't be a pain to get it into bioconda etc. I also wrote a Ribo-Seq QC package that I think will be easier to implement in nf-core but it has a slight different take on QC to RiboSeQC as it tries to describe the different properties of Ribo-Seq datasets in standalone metrics. These metrics make it easier to compare properties across 100s of samples but is less relevant for individual runs. https://github.com/JackCurragh/RiboMetric
i
RiboMetric sounds great to have - the outputs also could become part of the multiQC report given its capability of measuring features across many samples:) argparse will make RUST much easier to write a module for:) Are you referring to RUST 1.3.1? Is bioconda still needed given this really cool new tool https://seqera.io/containers/ is that right @Jonathan Manning or would getting it on bioconda still be necessary?
j
Bioconda will still be the requirement for the moment I think- there will need to be community discussion about when and how to switch away from that
👍 1
j
Yes! Containers is really exciting. I only just started using wave but this sounds like it is wave but easier?
j
It's basically doing the wave bit for you