I want to go straight from bwa mem alignment to BAM format as I don't need the SAM file and it takes up too much space. How do I achieve this?
$\begingroup$
$\endgroup$
0
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
1
For directly outputting a sorted bam file you can use the following:
bwa mem genome.fa reads.fastq | samtools sort -o output.bam -
Optionally using multiple threads:
bwa mem -t 8 genome.fa reads.fastq | samtools sort -@8 -o output.bam -
-
$\begingroup$ Ah, great solution if sorting is needed! I had tried this in the past but I was trying to pipe it though
samtools viewand then intosort. Didn't realise you could go straight intosort. $\endgroup$ Commented May 12, 2017 at 23:48
$\begingroup$
$\endgroup$
1
If you don't need sorting, you can avoid the computational overhead by piping the output from bwa mem into samtools view like so
bwa mem ref.fa in.fq | samtools view -b - > out.bam
The - in samtools view tells it to read from stdin and -b indicates BAM output.
You can also just allow samtools to infer the output format by using the -o option, which will guess the required format based on the file extension given
bwa mem ref.fa in.fq | samtools view -o out.bam -
answered May 12, 2017 at 5:08
-
$\begingroup$ In my workflow, BWA output goes to MergeBamAlignment, so
samtools viewseemed lower overhead thansamtools sort. (Directly piping from BWA to MergeBamAlignment, as suggested here, failed for me.) $\endgroup$ Commented Feb 1, 2023 at 18:53