I'm using GWASTools for my GWAS analysis. It turned out that there are some genotyping problems with a few hundred SNPs on the X-chromosome. I already have all my data stored in gds files for use in GWASTools and I want to change the genotypes of the erroneous SNPs directly in the gds file. I was hoping there is kind of a "setGenotype" function with works exactly like the "getGenotype" function that I use to obtain specific SNP genotypes, but couldn't find anything like that.

Did I miss anything? How can genotype changes of specific SNPs be done in gds files without going back to raw data file and creating a new gds file based on them?

Thank you for your help!

You need to use the gdsfmt package to modify GDS files directly.

library(gdsfmt)

file <- system.file("extdata", "illumina_geno.gds", package="GWASdata")
gdsfile <- tempfile()
file.copy(file, gdsfile)

# open file in write mode
gdsobj <- openfn.gds(gdsfile, readonly=FALSE)

# get node of the genotype variable
node <- index.gdsn(gdsobj, "genotype")

# uncompress node so we can write to it
compression.gdsn(node, "")

# change some values for 100 SNPs
nsnp <- 100
nsamp <- 77
newsnps <- matrix(sample(c(0,1,2), size=nsnp*nsamp, replace=TRUE), nrow=nsnp, ncol=nsamp)

# write new values for SNPs 101:200
write.gdsn(node, newsnps, start=c(101,1), count=c(100,77))

# close
closefn.gds(gdsobj)
cleanup.gds(gdsfile)

file.remove(gdsfile)

Thank you! That worked fine!