Hi there, 

It came about a situation which I have more samples that I can afford to proceed with the qPCR experiments. I have three clinically separated groups with more than 20 samples. So, the idea would be to select 7-8 sample from each group. For that, I have three numerical variable such as age, left ventricle diastolic diameter, left ventricle excretion fraction and perhaps other variables. Does it make sense to select samples according to a heatmap and/or PCA? Am I in danger of incur in any error?

Best,

Fred