conditions to apply BH correction

0

Entering edit mode

Loredana Le Pera ▴ 10

@loredana-le-pera-4645

Last seen 9.6 years ago

Dear List, I would like to understand if/when Benjamini-Hochberg correction can be applied and if not, what kind of correction is possible to apply. From the oneChannelGUI package vignette: "The function raw p-value distribution is implemented to evaluate if the BH type I error correction methods could be used. To apply BH correction two conditions should be satisfied: 1. The gene expressions are independent from each other. 2. The raw distribution of p-values should be uniform in the non significant range." My doubts mainly concern the second condition: - what does it mean exactly? - where does it come from? If this condition is not true, what I can do to consider multiple testing correction? Thanks for any reply or suggestion, Loredana.

oneChannelGUI oneChannelGUI • 1.5k views

ADD COMMENT • link updated 13.0 years ago by Pavel Goldstein ▴ 70 • written 13.0 years ago by Loredana Le Pera ▴ 10

0

Entering edit mode

Pavel Goldstein ▴ 70

@pavel-goldstein-4502

Last seen 9.6 years ago

Hi Loredana , 1. Positive dependence is also Ok. 2. under the null hypothesis, the distribution of the p-values is Uniform[0,1] - it can be proved. So this is assumption of the method, you do not have to check it. Pavel On Thu, May 12, 2011 at 5:11 PM, Loredana Le Pera <loredanalepera@gmail.com>wrote: > Dear List, > > I would like to understand if/when Benjamini-Hochberg correction can be > applied and if not, what kind of correction is possible to apply. > > From the oneChannelGUI package vignette: > > "The function raw p-value distribution is implemented to evaluate if the BH > type I error correction methods could be used. To apply BH correction two > conditions should be satisfied: > 1. The gene expressions are independent from each other. > 2. The raw distribution of p-values should be uniform in the non > significant range." > > > My doubts mainly concern the second condition: > - what does it mean exactly? > - where does it come from? > > If this condition is not true, what I can do to consider multiple testing > correction? > > > Thanks for any reply or suggestion, > Loredana. > > _______________________________________________ > Bioconductor mailing list > Bioconductor@r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]

ADD COMMENT • link 13.0 years ago Pavel Goldstein ▴ 70

0

Entering edit mode

Dear Pavel, thanks a lot for your reply, but to be sure my question was clear enough, I would like if you gave a look at this link: [1]http://rss.acs.unt.edu/Rdoc/library/oneChannelGUI/html/rawpCheck .html What do you think about it? Loredana -----bioconductor-bounces at r-project.org ha scritto: ----- Per: Loredana Le Pera [2]<loredanalepera at="" gmail.com=""> Da: Pavel Goldstein Inviato da: [3]bioconductor-bounces at r-project.org Data: 12/05/2011 06.16PM Cc: [4]bioconductor at r-project.org Oggetto: Re: [BioC] conditions to apply BH correction Hi Loredana , 1. Positive dependence is also Ok. 2. under the null hypothesis, the distribution of the p-values is Uniform[0,1] - it can be proved. So this is assumption of the method, you do not have to check it. Pavel On Thu, May 12, 2011 at 5:11 PM, Loredana Le Pera [5]<loredanalepera at="" gmail.com="">wrote: > Dear List, > > I would like to understand if/when Benjamini-Hochberg correction can be > applied and if not, what kind of correction is possible to apply. > > From the oneChannelGUI package vignette: > > "The function raw p-value distribution is implemented to evaluate if the BH > type I error correction methods could be used. To apply BH correction two > conditions should be satisfied: > 1. The gene expressions are independent from each other. > 2. The raw distribution of p-values should be uniform in the non > significant range." > > > My doubts mainly concern the second condition: > - what does it mean exactly? > - where does it come from? > > If this condition is not true, what I can do to consider multiple testing > correction? > > > Thanks for any reply or suggestion, > Loredana. > > _______________________________________________ > Bioconductor mailing list > [6]Bioconductor at r-project.org > [7]https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > [8]http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]] _______________________________________________ Bioconductor mailing list [9]Bioconductor at r-project.org [10]https://stat.ethz.ch/mailman/listinfo/bioconductor Search the archives: [11]http://news.gmane.org/gmane.science.biology.informatics.conductor References 1. http://rss.acs.unt.edu/Rdoc/library/oneChannelGUI/html/rawpCheck.html 2. mailto:loredanalepera at gmail.com 3. mailto:bioconductor-bounces at r-project.org 4. mailto:bioconductor at r-project.org 5. mailto:loredanalepera at gmail.com 6. mailto:Bioconductor at r-project.org 7. https://stat.ethz.ch/mailman/listinfo/bioconductor 8. http://news.gmane.org/gmane.science.biology.informatics.conductor 9. mailto:Bioconductor at r-project.org 10. https://stat.ethz.ch/mailman/listinfo/bioconductor 11. http://news.gmane.org/gmane.science.biology.informatics.conductor

ADD REPLY • link 13.0 years ago loredana.lepera@uniroma1.it ▴ 10

Login before adding your answer.