In my targeted sequencing of ~500 case and ~500 controls (same population) I have identified a some rare variants which are present in higher density in cases than in control. Now, two of the SNPs are having very high frequency in one another population according to Exac data-set. No, what is the best way of proving that there is no population contamination in my case samples?
"Now, two of the SNPs are having very high frequency in one another population according to Exac data-set. "
What populations are you comparing? Between e.g. Europe and Finnish is a smaller difference than comparing Asians and Europe so that is relevant to know.
The best way would be to genotype all 1000 samples on a normal genotyping chip and do a PCA and show that there is no overall difference between cases and controls, but likely you are not wanting to do this. So it depends a bit how many kb you resequenced? Also mining other datasets could be useful... Finally if you can prove that these SNPs are likely increasing risk to the cases that is evidence, but nowadays replication is wanted in an additional cohort. So you might want just to genotype these snps in a separate cohort.
