This is a test version of Biostars. For the public version, visit https://www.biostars.org.
microarray data analysis

I am working on microarray data analysis. to find out DE gene i have use p value <0.05. now i want to go with Benjamini-Hochberg correction, for this i have use samr package in R. for filtering significant genes delta value must be given by users. plz suggest me range of delta value and also give information about what is q value? how it is useful for Differential expression analysis of gene?

significance analysis of microarray
• 3,271 views • link
updated by arta arta • written by au.rinki.bio • au.rinki.bio

If you're working in R, are you using limma? and if so, why didn't you filter based on adjusted-p-value (the BH correction will have been done for you) instead of p-value.

• 0 views • link
written by russhh russhh

Hi au.rinki.bio,

Please chose a more descriptive title of your threads, "microarray data analysis" doesn't say a lot. "Multiple testing correction" is something you should have included, for example.

In addition, please give feedback on your previous question(s):
If an answer was helpful you should upvote it, if the answer resolved your question you should mark it as accepted. Upvote|Bookmark|Accept

Cheers,
Wouter

• 0 views • link
written by WouterDeCoster WouterDeCoster

1 answer

The reason why you are correcting the p values is multiple testing. In microarray data there are around 20.000 genes, as rows. So you do 20.000 times testing whether the gene is significantly DE. When you set your p value to 0.05, 1000 genes (20000 * 0.05) are significantly differentially expressed by chance. To avoid getting false positives or type-1 error, you correct your p values. Corrected p-values are generally called adjusted p values and q value is p value which is adjusted for False Discovery Rate (FDR).

To do that in R, you do not need to use any extra package. Following function will do the same job. 

p.adjust(p.values, method = "BH", n = length(p.values))
• 0 views • link
written by arta arta

i am using following command line and use affy package.

library(makecdfenv)
library(affyio)
library(affy)
ovarian<-make.cdf.env("HGU133A_Hs_ENSG.cdf")
data<-ReadAffy(cdfname='ovarian')
data
eset<-rma(data)
exprSet <- exprs(eset)
pvalue.exprSet = apply(exprSet, 1, function(x){t.test(x[1:4], x[5:45]) $p.value}) 
Combine.exprset.pvalue = cbind(exprSet,pvalue.exprSet)
write.table(Combine.exprset.pvalue,"NormalizedValues.xls",sep="\t",col.names = NA)
• 0 views • link
updated by WouterDeCoster WouterDeCoster • written by au.rinki.bio • au.rinki.bio

when i am using command suggested by you after p value calculation following error occur-

p.adjust(p.values, method = "BH", n = length(p.values))
Error in p.adjust(p.values, method = "BH", n = length(p.values)) : 
  object 'p.values' not found
Error during wrapup: cannot open the connection
• 0 views • link
updated by WouterDeCoster WouterDeCoster • written by au.rinki.bio • au.rinki.bio

I added code markup to your post for increased readability. You can do this by selecting the text and clicking the 101010 button. When you compose or edit a post that button is in your toolbar, see image below:

101010 Button

• 0 views • link
written by WouterDeCoster WouterDeCoster

thank you so much. but here is one more question what should be the range of thresold for agjusted p -value?

• 0 views • link
written by au.rinki.bio • au.rinki.bio

Based on your text above, the threshold should be 0.05. Can I strongly suggest that you read and work through the limma user's guide rather than doing the unmoderated t-tests that you've proposed. And by strongly suggest, I mean, really strongly suggest.

• 0 views • link
written by russhh russhh

when i am using limma package their are a lots of error coming and not running proprely. kindly suggest me any tutorial for analysis of microarray data( CEL file) using limma package.

• 0 views • link
written by au.rinki.bio • au.rinki.bio

In what way did it fail, and what have you done to work out why it failed? Here's a brusque tutorial I wrote: here

• 0 views • link
written by russhh russhh

thank you so much. but here is one more question what should be the range of thresold for agjusted p -value?

• 0 views • link
written by au.rinki.bio • au.rinki.bio

Log in to answer this question.

More posts like this