How to choose BWA-MEM2 settings

Question

[mgranada3]

I am trying to run a BWA-MEM2 script to align DNA-seq data. But I am following this guide.

How did they come up with the -t, -T, and -R information? In the manual, -t is threads, -T says "Don’t output alignment with score lower than INT. This option only affects output," and -R read group header line.

#!/bin/bash
#SBATCH -J BWA-MEM
#SBATCH -A gts-xxxxxx
#SBATCH -N 1 --ntasks-per-node=4
#SBATCH --mem-per-cpu=8G
#SBATCH -t 4:00:00
#SBATCH -o Report-%j.out

cd $PBS_O_WORKDIR
ml anaconda3
conda activate bwa-mem2
ml samtools

bwa mem \
-t 8 \ 
-T 0 \ 
-R \
/C_glabrata_index_mg.fa /SC_0_L001.fastq.gz /SC_0_L002.fastq.gz \
| samtools sort -o SC_0_output.bam

[Ram]

You need to be clear about where you got this script - it's not in that guide. Also, your command is wrong and it won't work because the -R is empty. And you seem to have undeclared aliases which will not play well with slurm since slurm doesn't read necessarily know aliases while starting up a shell to run your job.

Answer 1

[Ram]

They did not "come up" with these recommendations, they're using sensible defaults. They're essentially asking to not ignore any alignments and use a sensible default of 8 threads while leaving the read group info up to you. You're free to change them and see what happens; generally, developers set conservative defaults on optional parameters and you can change only those that you need to change. In this case, you can feel free to ignore all 3 of those parameters except that you'll face some problems downstream if you don't have a read group, so make sure you set that.