Single-end read compatibility

[LBWitt]

Hi, I am wondering if the pipeline is compatible with single-end read historical samples?
I have tried leaving the fq2 column empty in the units file and also deleting it entirely, but neither has worked.
If single-end reads are not supported, are there any workarounds, or is this planned for future development?

Thanks.

[zjnolen]

Hi there!

Currently, the mapping portion doesn't work with single-end reads, instead expecting historical samples to be given as paired end reads that then get merged.

As a workaround, if you're primarily interested in the population genetic analyses in PopGLen, the best option would be to first do the processing up to the bam file stage and input those. Something like mapache or grenepipe would be good ways to get this done in a snakemake workflow. The can be setup to do similar processing as PopGLen, but also has more customizable options for mapping.

I have been thinking of adding single-end libraries as an option in the future, but do not work with them myself, so wanted to do that with advice from people who do. I think that most of the steps could be easily adapted to just work with the single-end rather than paired-end reads, but duplicate removal would be my main uncertainty. Currently, PopGLen collapses historical reads into single fragments and uses Dedup to remove duplicates as it removes duplicates using coordinates from both fragment ends. Any that don't collapse go to Picard and are treated as paired end. In the case of single-end historical reads, I think combined adapter and quality trimming would mean the end of the DNA fragment can't be determined as with overlapping paired-end reads, so all reads with the same start position would have to be treated as duplicates (as Picard does). Is this in line with what you would expect, or is there another way you think these could be handled?