I'm new to nf-core/fastquorum and have a question ...
# fastquorum
e
I'm new to nf-core/fastquorum and have a question about the alignment strategy. It seems that all "bwa-mem" commands in the ".command.sh" files come with the "-p" tag. Given that I explicitly run the pipeline with "--duplex-seq false" since I need only single-end reads for my experiment, doesn't that defeat the whole purpose? I do not wnat the aligner to think that my data is pair-ended at all...
s
What do you mean by "I need only single-end reads"?
e
The experiment was designed in a way that essentially makes it single-end. From what I saw in the documentation of the "fastquorum" pipeline that it handles single-end just fine. And the output of it on my data also seems reasonable. Yet in the ".command.sh" files that document the executed commands (inside the work directory) I saw the "bwa-mem" command together with the "-p" flag which is meant for interleaved paired-end files. Now I know that BWA documents are a mess at times, so maybe I'm misunderstanding something here. Still I'm reluctant to fully trust the results without fully understanding how it worked...
A separate point that I would like to make, after looking at the commands, is that the pipeline would be much more flexible if it could alternate between "bwa mem" and "bwa aln" depending on provided template length. Or if it allowed to change seed length for "bwa mem".
s
I don't think there is much call for bwa aln to be honest. I think the
-p
is because the input is an unmapped bam, which has been samtools fastq, so that generates interleaved format in the case of paired end data. I don't think it will make any difference in your single ended case. You can change any of the arguments to bwa mem via
ext.args
e
Thank you, I will check this opportunity. From what I gather "bwa aln" is somewhat superior to "bwa mem" when it comes to shorter reads, even if such reads are not common, though playing with seed size may provide similar results. Combing over the manual for "bwa" I found the source of my confusion about
-p
. At the beginning of the manual it is presented as an option specifically for interleaved format. Later on it is denoted as "smart format" option that decides whether adjacent reads in the input file are paired or not based on their names, on account that reads of same pair share same name. This allows processing of files with both single- and paired-end reads.