johnros / selectiveci Goto Github PK

Hi, I have been looking for functions to perform multiple corrections for confidence intervals and found your SelectiveCI scripts. I have bootstrap estimates of mean gene measurements, have constructed 95% confidence intervals, and have compared intervals from an experimental treatment to a control treatment, considering those genes with non-overlapping intervals to be statistically significant.

If I understand the BY2005.R script correctly, I would input x=mean estimate, sigsq=sd of the population squared?, n.rejections= number of of genes with overlapping intervals?, n.tests=number of genes tested?, alpha=0.05

Is the above an appropriate use of this function for generating FCR adjusted multiple confidence intervals?

Thanks,
Jeremy