QTCAT Quantitative Trait Cluster Association Test problem with reading SNP data
1
0
Entering edit mode
2.5 years ago

Hello everyone, I'm running a GWAS analysis. First I used GAPIT and wanted to use PCs as population control covariates. To determine the appropriate number of PCs to include, I used BIC values which recommended 0 PCs (highest BIC value) which was consistent with PC plots as they also show no clear population stratification.

As my population seems unstructured, i wanted to used another association tool QTCAT:Quantitative Trait Cluster Association Test which is for unstructured populations and removes the need for adjustment for population structure in association analysis. It is also an R package just like GAPIT.

the genotype file requirement for this package is diploid Hapmap. now i have the diploid hapmap from TASSEL, but the problem is the file contains so many NN for missing values, which is not accepted my QTCAT, hence, all SNPs with NN are removed/not considered which removes 2/3rd of genotype file and considers only 1/3rd, for which it is highly unlikely to get any association.

Question is, How can i use whole Hapmap file without removing NN? is there any way to Impute it first? another requirement is the snp data should be of snp.matrix class which only has homozygous and heterozygous form of allels. NN tells it there's another form, hence not considered and removed.

GAPIT R QTCAT GWAS Hapmap • 736 views
0
Entering edit mode
23 months ago
BioRyder ▴ 200

You should mention the missing genotype format ('NN' f) while reading the genotype file in QTCAT. Below are sample code for considering NN while reading file.

gfile <- system.file("genotype_file.csv", package = "qtcat")

snp <- read.snpData(gfile, sep = ",",na.string = "NN")