4

I would like to perform a samtools mpileup from a single file that contains thousands of read groups with different SM tags. I could split the bam by read group using samtools split, and then perform the mpileup, but splitting into thousands of files and then performing an mpileup from thousands of files is crazy slow on my machine.

I could imagine an implementation using pysam, but I was hoping someone has a solution already.

Bcftools mpileup already has this feature, but the feature was never implemented for the original mpileup format, which is what I need.

An example

Contents of input.sam:

@HD VN:1.6  SO:unknown
@SQ SN:chr1 LN:1000000
@RG ID:0    SM:sample_0
@RG ID:1    SM:sample_1
r0  0   chr1    24  0   1M  *   0   0   G   I   RG:Z:0
r1  0   chr1    24  0   1M  *   0   0   G   I   RG:Z:1

When I run samtools mpileup I get this behavior:

> samtools mpileup input.sam
[mpileup] 2 samples in 1 input files
chr1    24  N   2   ^!G$^!G$    II

Whereas, what I actually want is this output:

[mpileup] 2 samples in 1 input files
chr1    24  N   1   ^!G$    I   1   ^!G$    I

For a long time I thought this was a bug in samtools mpileup, but it appears that it is a feature that was described/hinted at in the samtools documentation, but was never implemented: https://github.com/samtools/samtools/issues/599#issuecomment-604941884

winni2k
  • 2,266
  • 11
  • 28
  • been a long time since i use that.. you need the RG tag field the reads right? mpileup uses this to separate into different columns (i.e samples) – StupidWolf May 03 '20 at 20:13
  • Won't GATK handle this? At the very least the old unified caller required input exactly like this. – Devon Ryan May 05 '20 at 07:01
  • That's a really good point! I did not realize gatk has a Pileup command. Unfortunately, that command (I tested v4.1.7.0) is also not read group aware :( – winni2k May 05 '20 at 08:39

1 Answers1

2

I broke down and finally scratched this itch. I have implemented a tool that performs read-group-aware text-pileup from a single SAM/BAM file. The tool is called streaming_pileupy and it's available from pypi for installation with pip. After installation the command would be:

spileup input.bam sample_names.txt

The tool is pre-alpha and missing some features. Contributions are welcome.

winni2k
  • 2,266
  • 11
  • 28