rnaseq icon indicating copy to clipboard operation
rnaseq copied to clipboard

Add pseudo-alignment via Kallisto

Open drpatelh opened this issue 5 years ago • 3 comments

We already have an option to perform pseudo-alignment with Salmon but it would be great to have a similar workflow that uses Kallisto instead. I am hoping we can recycle alot of the logic I wrote when writing the Salmon DSL2 sub-workflow. If so, it should be relatively straightforward. Will need some experienced input in order to provide sensible defaults for the indexing and quantification steps. Contributions, comments and general feedback welcome :)

drpatelh avatar Dec 01 '20 23:12 drpatelh

I'm happy to give it a try if it's alright?

BatoolMM avatar Dec 03 '20 17:12 BatoolMM

Absolutely! That would be great @BatoolMM 😎

drpatelh avatar Dec 03 '20 17:12 drpatelh

Hi @drpatelh

Please check this - it might come in handy to add kallisto quant modules

https://github.com/tamuanand/rnaseq/tree/master/modules/nf-core/modules/kallisto

Please let me know if you have any questions.

Some assumptions:

  1. This assumes that the kallisto/index has been built with this
  2. Please look at the comments here about how I have my sample_ids named- feel free to change the code to make it more generic https://github.com/tamuanand/rnaseq/blob/master/modules/nf-core/modules/kallisto/tximport/bin/kallisto_tximport.R#L15-L26
  3. The R script needs further enhancement to account for lack of tx2gene file
  4. I then collect all the individual stderr files from Kallisto to feed into MultiQC
  5. The module could be improved further to include bootstrapping - https://pachterlab.github.io/kallisto/starting - this might help - https://github.com/cbcrg/kallisto-nf/blob/master/kallisto.nf
ch_kallisto_multiqc = KALLISTO.out.kallisto_stderr_mqc

MULTIQC (
        ch_multiqc_files.collect(),
        ...............
        ch_kallisto_multiqc.collect{it[1]}.ifEmpty([]),
       ..............
    )

tamuanand avatar Mar 19 '22 09:03 tamuanand