This is a test version of Biostars. For the public version, visit https://www.biostars.org.
file formatting for SNPRelate

Hello!

I am trying to use the R package SNPRelate to calculate pairwise relatedness between 8 males and 8 females. I have a VCF file with ~13,000 SNPs. I am able to use SNPRelate to convert the VCF file to a GDS file with snpgdsVCF2GDS(). However, when I try to use the functions that act on the GDS file, such as snpgdsLDpruning() or snpgdsIBDMLE(), I get an error: "There is no SNP!" because all ~13,000 SNPs were removed as non-autosomal. However, almost all my SNPs are autosomal. So I'm wondering if my VCF was improperly formatted or missing information and how I might add it back in. Suggestions? Thx.

#open vcf
parents.vcf <- "~/Data/parents.vcf"

#convert vcf to gds
snpgdsVCF2GDS(parents.vcf,"parents.gds")

#open gds
parents.gds <- snpgdsOpen("parents.gds")

#calculate relatedness
snpgdsIBDMLE(parents.gds)

>Identity-By-Descent analysis (MLE) on SNP genotypes:
Removing 13354 SNP(s) on non-autosomes
Error in .InitFile2(cmd = "Identity-By-Descent analysis (MLE) on SNP genotypes:",  :
  There is no SNP!
vcf relatedness snprelate gds
• 5,528 views • link
updated by Ram Ram • written by lovestowell • lovestowell

1 answer

In case anyone else is having the same problem and also didn't read the man pages carefully...

Both snpgdsVCF2GDS() and snpgdsIBDMLE() have some extra arguments that make things work more smoothly. I wasn't sure if the sex chromosome in my original VCF file was part of the trouble, so I removed it and then proceeded to use SNPRelate.

In bash shell:

grep -v '^chr_Sex' parents.vcf > parents.autosomal.vcf
vcftools --vcf parents.autosomal.vcf --recode --recode-INFO-all --out parents.autosomal.vcf

In R:

parents.vcf <- "~/Data/parents.autosomal.recode.vcf"
snpgdsVCF2GDS(parents.vcf,"parents.gds",ignore.chr.prefix="chr_")
parents.gds <- snpgdsOpen("parents.gds")
IBDMLE <- snpgdsIBDMLE(parents.gds)
• 0 views • link
updated by Ram Ram • written by lovestowell • lovestowell

Log in to answer this question.

More posts like this