Is there a working pipeline to use vg for such human genome project? (including augmentation and call)

MehmetGoktay commented 5 years ago

I m planning to use vg to call snps for A. thaliana (more then 1200 samples).

I understand for such projects I need to use toil-vg call ( because augmentation and call does not scale up for such big projects)

Does anyone have simple pipeline that I can test for a few samples starting from raw vcf file till calling snps against graph (including augmentation and call steps )

Best, Mehmet

adamnovak commented 5 years ago

@glennhickey Do you have a whole genome pipeline example handy? I feel like either you or @cmarkello has done this most recently. You ought to be able to do the whole thing with toil-vg run, but it has a jillion options and the only way to figure out which to actually use is to look at an example.

And if so can we maybe put it on the wiki? We definitely ought to have an official example for this case.

We do have https://github.com/vgteam/vg/wiki/Working-with-a-whole-genome-variation-graph but that just covers up through alignment and doesn't use toil-vg.

glennhickey commented 5 years ago

Thanks for the reminder. I will put something in a README or WIKI soon and link to it here soon. A couple things in the meantime:

toil-vg call --help will show the argments and options. Of the options, --gams and --chroms are the most important. You should be able to make a VCF in the directory "output" using the directory "js" as a working directory with something like this command

toil-vg call ./js ./graph.xg SAMPLE ./output --gams sample.gam --chroms chr1 chr2

Some things to watch out for

toil-vg call can only call one (diploid) sample at a time.
it is usually helpful to normalize the output vcf (bcftools and vt can do this)
you can use the --genotype option to write genotype qualities but it will be even slower. otherwise the QUAL field as well as the AD fields can be used to get an idea of the confidence in a call

Padraic8 commented 5 years ago

Hi, any update regarding the README or WIKI example for this? Would really appreciate a template we could work off. Thanks

vgteam / vg

Is there a working pipeline to use vg for such human genome project? (including augmentation and call) #2122