I have about 4200 array genotyping VCFs (from the Illumina Infinium CoreExome-24 Kit) and I have merged them using bcftools
merge. The chip has 500K exonic SNPs. These are trio data - which means 1700 of them are probands, mother and father respectively.
Following that, I converted them to Plink format using plinkv2
--make-bed function. I included the fam information later on by generating a fam file.
Finally, I ran KING
--kinship but am getting a very high error rate for within family relatedness, which I know is not correct.
Are there other QC steps I should have been running at the merge step? I have removed all monomorphic SNPs and multi allelic sites, too.