sample size calculation - (May/23/2008 )
My project involves studing SNPs in a cancer. Now I will be having 1000 cancer samples. HOw do I dertermine teh apprpritae number of healthy samples to include in my study? Is there a minimum ratio? 1000 unhealthy sample to 500 healthy samples?
This is a complex subject, and you certainly need to talk with a biostatistician. But some pointers here:
If you have homogenous samples, you can reduce the healthy samples.
-cancer tissue/organ of origin (more organs --> bigger sample size)
-cancer type (more types --> bigger sample size)
-cancer aggresiveness (wider range of aggresiveness --> bigger sample size)
-Uniformly treated, treated under many different protocols, non-treated
-population from which the cancer samples were taken. Multinational, multiracial, multi-age, both sexes--each of these variable increases the requirement for healthy samples.
Look up some lectures here:
But for a precise number, you need a biostatistician. Don't fool around with the subject till it is too late!