I Have download GSE21935.CEL Files.I have performed normalisation,preprocessing using R Package and when i am proceeding in identifying differentialy expressed genes,i am getting all adjusted p-values in range of 0.99.How to proceed with it?
I Have download GSE21935.CEL Files.I have performed normalisation,preprocessing using R Package and when i am proceeding in identifying differentialy expressed genes,i am getting all adjusted p-values in range of 0.99.How to proceed with it?
You have chosen a pretty tricky data set to analyze. Have you read the paper? There are some hints in there that you might try. They fit the first principal component as a coefficient in their model, which is quite similar to what you would do if you used the sva package (using sva(), not ComBat() - see the vignette for that package).
But there are lots of reasons why you might not get the results you expect. First, the samples are brain tissue from dead people, with a mean delay of 8 hours between death and sample excision. That's a really long time for mRNA to be sitting around, in an environment that is highly conducive to mRNA degradation. Second, in my experience with brain samples, the differences are often extremely subtle, so this study may well be underpowered. Like I said, tricky.
The original authors make no mention of multiplicity adjustment, so I assume they used unadjusted p-values. Otherwise I doubt they would have any genes to proceed with. They then did some hypergeometric tests using GO terms. You could try that, or maybe better, do GSEA using gene sets that you think might be perturbed in schizophrenia.
But I wouldn't recommend this data set for a beginner, so if you have a choice, you might choose something less demanding.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.