Small sample size

Hello,

In the eigenMT paper, sample sizes as small as 50 were tested. We have a sample size of 20 for a QTL analysis we are performing with RASQUAL. Is 20 possibly too small to work with eigenMT? Should we use an external, larger genotype source for LD estimation, or are our 20 samples sufficient for a "close enough" estimation?

Thanks,

Kevin
