Walk me through the mathematical calculation of statistical power. How does sample size affect the power of a genomic validation study?