Pre-processing alignments with bamshrink before genotyping with `--no_bamshrink`

[malthesr]

For technical reasons, I'd like to pre-process alignments with graphtyper bamshrink and then running graphtyper genotype --no_bamshrink afterwards.

Is there a set of options I can provide to graphtyper bamshrink such that running graphtyper genotype --no_bamshrink on the output would yield equivalent results to just running graphtyper genotype?

I was playing around with settings in graphtyper bamshrink and comparing to the tmp BAM output from graphtyper genotype -a --no_cleanup, but I end up with different reads in the two outputs.

[KH-NN]

Hi,

is it possible to provide a command so that we can shrink the bam/crams before graphtyper. We have a problem with disk size in cloud so it would be awesome to do this before genotype calling.

[hannespetur]

Hi, I think typically the --avg-cov-by-readlen=D --interval=chrA:B-C options should be set in graphtyper bamshrink. "chrA:B-C" should be the region being genotyped padded by 100bp.

Best, Hannes

[KH-NN]

Thanks for getting back to us.

With this command would we get the same as the bamshrink part does in graphtyper genotype?

[hannespetur]

Yes, you should maybe make sure as well, but in my test I got 100% identical results.

You can test it by running with the --no_cleanup option in graphtyper genotype.

graphtyper genotype <ref> --sam=<BAM> --avg_cov_by_readlen=<(echo 0.2) --region=chr20:20000001-20001000 --no_cleanup

and check if the file in /tmp/graphtyper*/bams is the same as the bamshrinked file with 100bp padded region

graphtyper bamshrink <BAM> --avg-cov-by-readlen=0.2 --interval=chr20:19999901-20001100 --output=test.bam

Both files had the same md5sum

Best,
Hannes

[KH-NN]

fantastic. we will try. thanks a lot.