Hi everyone!
I have used BINGO -- a Cytoscape plugin that performs gene set enrichment analysis -- for ages and only recently the following question has arisen in my mind: do the reported "corrected p-values" correspond to the false discovery rates as determined by the Benjamini & Hochberg FDR correction?
Hi!
Are think you are absolutely correct that with multiple variables e.g. genes, polymorphisms, SNPs, that chance occurrence of rejecting the null hypothesis is artificially too high and leads to False Positive conclusions.
There are a variety of ways to deal with multiple variables statistically to correct the p values. In general , it depend upon whether the variables are considered independent or more likely dependent. The Bonferroni correction is conservative, while the two modifications are similar in concept but actually analyze the data in opposite directions.
I would suggest you contact Marc Hochberg directly at University of Maryland where he is rheumatologist epidemiologist and likely very willing to provide you with more definitive answer. I would google U of M medical school rheumatology and you should find the email link. Let me know if you have any glitches and I will look online in my directory of ACR (American College of Rheumatology).
- Badges
- Science topic
- Similar topics
- Psychology