Hi all, I'm new to this channel 🙂 and I wanted to know if there is anything I can adjust in this command to get the featureCounts table: nextflow run nf-core/rnaseq -r 3.26.0 -profile docker --input /home/stefanie/Biggie/A549_RNAseq_GSE215771/samplesheet_test.csv --outdir /home/stefanie/Biggie/nf_output --fasta /home/shared/GRCh38/GRCh38.primary_assembly.genome.fa --gtf /home/shared/GRCh38/gencode.v43.primary_assembly.annotation.gtf For context, my lab is switching over from in-house pipelines to nextflow because package dependencies have become a pain. In our in-house pipeline, we would do FastQC (of input fastq files) ---> cutadapt to trim ---> alignment with STAR --> index resulting bam files via samtools --> filter aligned bam files for quality via samtools --> index filtered bams --> qualimap QC the mapping --> featureCounts for the counts file. But I noticed that with nextflow, a plethora of tools are deployed and it's taking up a lot of computational space. There isn't an option to run things one at a time just so I can see what the output is at each step, but perhaps it would be greatly appreciated if someone here can help me modify my command so that it doesn't run tools I did not use before?

Take a look at the process skipping options, so you can e.g. skip QC tools you do not like to run. If the number of containers bugs you, you can also

--use_rustqc

to replace many QC tools with one. While experimenting with the best combination, using the CLI is probably fine,m but for later, I recommend using a

-params-file

in order to provide the parameters to Nextflow. Like so it is easier to document and re-run. Since Nextflow has a

-resume

option, you can interrupt a pipeline run anytime in between, if you feel like something is off. You can then alter a parameter or add a new one and resume your previous run. Nextflow will not recompute what doesn’t have to be recomputed in that case.

Sorry @Matthias Zepper, that's not quite right. featureCounts is NEVER used for quantification in the pipeline, even with prokaryotic data (which by default does bowtie2 -> Salmon)

Thanks for the correction. I love how Brits politely say “not quite right” instead of “plain wrong” 😂 I did indeed not corroborate that statement, just recalled that we discussed about it being useful for prokaryotic quantification and assumed it was likely added when you created that profile.

I'm wondering if you could simply run featureCounts separately, outside of the nf-core pipeline? All it would need is the gtf and bam as inputs.