Hi Megan. Are the BAMs you received actual alignments or is the BAM format being used to store unmapped reads?

Thank you for your response! The company returned BAM files that contains a header and the aligned reads (read name, position, and alignment quality). Reads in this file that aligned to repetitive elements were filtered out and PCR duplicates were removed. Multimapping reads are also not included.

@Megan Rasmussen Sorry for the delay! What analysis are you trying to do with the data then?

I'm hoping to do the riboseq specific tools and downstream analyses including for ORF predictions and translational efficiencies. We want to observe some usual QC features, like periodicity and footprint length, first but differential expression with translational efficiencies is our first main goal.

I think you will need to use the

--skip_trimming

and

--skip_alignment

arguments and set up the sample sheet with the paths to the BAMS instead (But I have never tried this so I don't know if itll work smoothly)

Thank you! We tried running with the BAM samplesheet and despite the skip commands it returned with an error "-- Entry 1 - fastq_1: FastQ file for reads 1 must be provided, cannot contain spaces and must have extension '.fq.gz' or '.fastq.gz' ". Is there a spot in the code where we can ignore the input looking for *.fastq.gz and search for the *.bam instead?

If you are comfortable changing the code yourself you could populate these channels manually. I think realistically I was naive to think my last suggestion would be enough. Apologies

Hi @Megan Rasmussen the current version of the workflow doesn’t support multiple entry points or using bam files as input. Thus if you want to use nf-core/riboseq you would need to either make multiple changes to the code on your own fork/copy (not ideal because you lose the benefit of version control) or better, ask the sequencing company for the fastq files and settings for UMIs, trimming and alignment and you run nf-core/riboseq starting from the raw/trimmed fastq reads. This would overall make your analysis end-to-end and more reproducible.