t test Entering data and FDR

  • Iran Malavazi

    Iran Malavazi - 2011-09-12

    Dear all
    I have been using MEV for a while, but I'm now face with new needs that I'm not familiar
    I was using one class T test for the experimental design below using two color array and Agilent slides and protocols:

    WT grown medium A x WT grown medium B during 2 hours
    WT grown medium A x WT grown medium B during 4 hours
    WT grown medium A x WT grown medium B during 6 hours

    I have duplicate for each array (6 hybridization) and using the one class T test I got 3,700 differentially expressed genes

    I have two questions:
    Do you think the way I entered the data for analysis is correct. We know two replicates are not enough.

    The other question is:
    While performing the one class t test I used “p-value based on t distribution” in the menu p-value parameter and “just alpha no correction” in the menu p-value/false discovery correction. Therefore I don’t have information about the FDR.
    How should I perform this analysis in order to have de FDR in the one class t-test?
    Do I have to choose “p-value based on permutation”? How can I manage the permutation while setting the test?
    Which is more stringent: “Bonferoni correction” or “adjusted Bonferoni correction”?
    How can I set a FDR of 5% for example? Is it possible?

    Sorry for so many questions.
    I hope you could help me

  • Eleanor Howe

    Eleanor Howe - 2011-09-14

    Hi there,

    First of all, I think you want a time-course analysis rather than a one-sided t-test for this dataset. I recommend checking out the BETR module or the time-course functionality in the LIMMA module.

    Of course two replicates isn't ideal. But it's what you have; I would try using one of the appropriate packages and see how you do.



Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.

No, thanks