[BioC] help for BH - FDR and Fold Change filtering question,

[James MacDonald]

Hi Saurin,

Saurin D. Jani wrote:
> Hi BioC,
> 
> For affymetrix microarrays, I have two groups: WT (wild type) and KO (knock out). I normalize data by some method, and then I do following steps for all my GeneChips.
> 
> - Perform t-test 
> - generate p-value
> - get Fold Change 
> - compute adjusted p-value (BH method)
> 
> Now, I do filter genes based on two options, please let me know which is better.
> 
> OptionsA:
> - filter genes which satisfies p-value and fold change threshold and 
> get number genes and get max. adjusted p-value. as FDR.
> 
> OptionsB:
> - filter genes which satisfies BH-FDR (fdr x) and I get N number of genes, so, then I do Fold Change threshold so, I get N-1 genes and  report N-1 genes with x fdr. 
> 
> is my x fdr is still true for N-1 genes or no? because have done extra step after I filter genes based on fdr. 

No, the fdr will not be accurate any more.

Best,

Jim


> 
> Thank you so much in advance,
> Saurin
>        
> ---------------------------------
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, MS
Biostatistician
UMCCC cDNA and Affymetrix Core
University of Michigan
1500 E Medical Center Drive
7410 CCGC
Ann Arbor MI 48109
734-647-5623