No paired-end reads were detected in paired-end read library about drop HOT 9 CLOSED

Hi Dina!
The way we perform the counting in the AS module is:

• we count all the split reads per sample, irrespective of an annotation file
• we merge them per cohort (drop group) and generate the splice sites
• on each splice site, we count the non-split reads
  From what I can see, in your case, the 2 first steps were completed, but the non-split reads did not work. Could it be that your data is not paired-end? For that step we use the featureCounts function from Rsubread. I found a similar error message from yours here: https://support.bioconductor.org/p/9141314/

from drop.

Hi Dina,
are you sure your bam files indeed contain paired-end reads as specified in your sample annotation? The error seems to indicate that no paired-end reads are found in the featureCounts call that Vicente described.

from drop.

Thanks for the response!
It's synthetic data, created as forward and reverse .fq files, then mapped to one file using STAR - so yes, definitely paired-end reads, but I'll check if something went wrong during this process.

from drop.

I've checked my data and it seems some are paired-end, and some are single-end sequencing. Is it possible to compare the two, or do all samples need to be one or the other? Will this cause any caveats I should be aware of?
Thank you @vyepez88 and @ischeller !

from drop.

Hi @dissakova ,
great that you found the problem!
We don't have experience ourselves for combining a dataset with paired-end and single-end sequencing data, so we can't give advice regarding this. The count produced by FRASER is a fragment count for paired-end data and read count for single-end data, so it's possible that it has an effect on the analysis. Consider splitting your analysis into one for the paired-end and one for the single-end reads if possible.

from drop.

and if you want to keep the samples together, simply specifying T/F in PAIRED_END column accordingly for each sample should make the pipeline run

from drop.

@vyepez88 Thank you! I ended up finding paired-end versions of the single-sequenced samples, and the pipeline works now. Thank you for your help!

from drop.

great to hear! for an overview of the samples, you could execute snakemake sampleAnnotation. Good luck!

from drop.

@vyepez88 Hello Vicente, I actually realized the pipeline wasn't fully completing but instead failing at 99% completion, but still outputting a results file. I've opened a new issue as this seems unrelated to the questions above.

from drop.