Question: Retrieve list of kept SNP's after LD pruning
1
gravatar for leeandroid
6 months ago by
leeandroid80
leeandroid80 wrote:

Hi everyone!

I followed the SNPRelate package tutorial in order to filter a VCF by LD and finally pursued PCA. My problem is that by converting the VCF into GDS SNP's and Chromosomes ID's are also converted into a different nomenclature.

sample.id, a unique identifier for each sample.

snp.id, a unique identifier for each SNP.

How can I get the original SNP ID's that are kept after LD pruning?

Thanks in advance.

setwd("~/Desktop/snprelate")
    library(gdsfmt)
    library(SNPRelate)

    cap.vcf <- "all_148.vcf"
    snpgdsVCF2GDS(cap.vcf, "cap.gds", method = "biallelic.only")
    snpgdsSummary("cap.gds")

    genofile <- snpgdsOpen("cap.gds")

    pop_code <- scan("pop.txt", sep = "\t", what = list("pop.txt"))
    pop_code

    set.seed(1000)
    cap.LD <- snpgdsLDpruning(genofile, remove.monosnp = TRUE, ld.threshold = .1, maf = .1, missing.rate = 0.1, method = "composite", slide.max.bp = 500000, verbose = TRUE)

    snp_id_list <- unlist(cap.LD)
snprelate snp gds R vcf • 376 views
ADD COMMENTlink modified 6 months ago by zx87545.3k • written 6 months ago by leeandroid80
2

Thank you for adding the code.

A few things to note:

  • if the chromosomes in your VCF are labelled with a 'chr' prefix, then you should use ignore.chr.prefix = "chr" with the snpgdsVCF2GDS() function
  • SNPRelate retains both the row index of each variant and also the ID (presumably whatever was set in the VCF ID field). The documentation states: "the variable snp.id stores the original the row index of variants, and the variable snp.rs.id stores the rs id". So, if you want to map back from the LD pruning object to the variant ID, then go by the row index.
  • for samples, I can only assume that they are taken from left to right as they appear in the VCF. You should run rigorous checks to confirm this by looking at variants in the GDS object and how they match up to your VCF
ADD REPLYlink written 6 months ago by Kevin Blighe30k
1

Thank you for replying, Kevin.

I was aware of the importance of chromosome nomenclature so I previously changed it to numeric values. Regarding variant ID's you're right. I was able to retrieve its original name by corresponding the row index in the snp_id_list to the row with the same index of the snp.rs.id stored in the GDS file. (In this case sample name was not an issue, I copied it by mistake when I wrote the question. But, its stored the same way as the snps). Any of GDS fields can be checked by read.gdsn(index.gdsn(genofile, "field_name"))

Thank you!

ADD REPLYlink written 5 months ago by leeandroid80
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1192 users visited in the last hour