Co-expression analysis of miRNAs and target mRNAs

Question

[baibhu1234]

Hello everyone, I was working on identification and characterisation of miRNAs using sRNA dataset. I have identified the miRNAs using ShortStack program. I have identified the potential mRNA targets using psRNAtarget program.

Now i need to perform the co-expression analysis of miRNAs and corresponding targets. which normalisation i should perform in order to make the comparison unbiased.

Do i need to perform normalisation for library size, mRNA size, and log2 (to control variance) all?

Any suggestion is appreciated. Thanks

Answer 1

[i.sudbery]

My vote in these situations is generally rlog or vst. They implement both library size normalisation and variance stabilisation. I wouldn't control fo rmRNA size for co-expression analysis, as things are compared to themselves, differences in mRNA size are cancelled out anyway.