How To Calculate The P-Value Of Mirna Between Two Samples

Question

[2011101101]

I have known the expression(normalized) of miRNA between two samples ,I want to calculate the p-value and if the p-value <0.01.Thank you for your advice first !

[Nicolas Rosewick]

could you give more informations (which tool ?, which type of data ? , etc...)

Answer 1

[Sean Davis]

If you have only two samples (and each sample has only one measurement per gene, I am assuming), getting a statistical measure (p-value) of differential expression is not possible. Perhaps the best you can do is to calculate a fold change between the two samples.

[2011101101]

I don't understand why calculate a fold change ? log2(c2/c1)?

[ancient_learner]

refer this (especially page:17) http://bioconductor.org/packages/2.12/bioc/vignettes/edgeR/inst/doc/edgeRUsersGuide.pdf

[Sean Davis]

I am not sure that the data are from miRNA-seq. Even if they are, they are not raw counts, so edgeR will not likely be the correct toolset. That said, the general guidance in the edgeR user guide is still somewhat applicable.

[ff.cc.cc]

It is just a common way to see how many times a probe is over/under expressed in a group vs the other.

Since the mean of your 2 populations is the observation itself (N=1), It doesn't have much sense (even if technically feasible) to use T-test based statistics (e.g. limma)

So look at biologically interesting differences.

[Sean Davis]

I cannot comment on how to calculate fold change without knowing the scale of your data (linear, log2, log10, etc.). I suggest that you contact a local bioinformatics person to work through this with you.

[2011101101]

thank you for your reply