search

churchill-lab / g2gtools

Personal diploid genome creation and coordinate conversion
http://churchill-lab.github.io/g2gtools
21 stars 9 forks source link

Unable to Generate Correct *.chain file for Inbred Mouse Strains #13

Closed rantingswede373 closed 6 years ago

rantingswede373 commented 6 years ago

I'm attempting to use g2gtools to convert coordinates between two inbred mouse strains (C57Bl/6J and CAST/EiJ) following the documentation online. If I ignore Indels it seems to work fine with the patched genome reflecting known variants in a few quick spot checks. The generation of the chain file for coordinate conversion does not seem to be working however, as the file is very small and a test run with a signal bed file returns the same coordinates as in the original reference despite there being millions of Indel variants between the two.

I'm using version 0.1.3 with the relevant command being: g2gtools vcf2chain --pass -f NCBIM37_um.fa -i mgp.v2.indels.annot.reformat.vcf.gz -s CASTEiJ -o mm9-to-CASTEiJ.chain

This returns a stdout as below: Chromosome: 1 STATISTICS Chromosome: 10 STATISTICS Chromosome: 11 STATISTICS ... ... complete Parsed 0 total lines Chain file created Execution complete: 00:00:10.13

So obviously something is off, although no explicit error messages come up. I confirmed that the strain is indeed "CASTEiJ" and not "CAST_EiJ" or some other variant as well as any other small issues that may have led to this. I'd like to try version 0.2.0, but haven't been able to successfully install it. Any help would be greatly appreciated!

kbchoi-jax commented 6 years ago

I think Sanger vcf files use CAST_EiJ as a strain name. Anyways I still highly recommend using v0.2.0. Please try my new documentation and let me know how the installation fails.

http://churchill-lab.github.io/g2gtools/#installation

rantingswede373 commented 6 years ago

If I change the --pass filter to --quality it appears to work. I'll have to double check the numbers to be sure they make sense, but it's at least generating statistics for each chromosome and the chain file is ~83 MB. Sorry, I thought I had checked that before posting.

At least from the source I have I think CASTEiJ is correct though: `curl ftp://ftp-mouse.sanger.ac.uk/REL-1211-SNPs_Indels/mgp.v2.indels.annot.reformat.vcf.gz --output mgp.v2.indels.annot.reformat.vcf.gz gunzip mgp.v2.indels.annot.reformat.vcf.gz grep -v '##' mgp.v2.indels.annot.reformat.vcf | head

CHROM POS ID REF ALT QUAL FILTER INFO FORMAT 129S1 AJ AKR BALBcJ C3HHeJ C57BL6NJ CASTEiJ CBAJ DBA2J FVB_NJ LPJ NODShiLtJ NZO PWKPhJ Spretus WSBEiJ`

kbchoi-jax commented 6 years ago

Sorry I never tested g2gtools with those old versions. I note they released REL-1807 recently.