How to combine biological replicates using a weighted average method?

Question

[Ashutosh]

Dear Community,

I want to combine my normalized log2 intensity ratios (score column in bed file) of biological replicates from tiling array data set. These biological replicates are processed together and similarly normalized. I have understood that a simple arithmetic mean of the log2 intensity is the wrong way of combining replicates. As described in the literature http://www.sciencedirect.com/science/article/pii/S0092867405006458,

I want to use weighted average method to combine log2 intensities from biological replicates. I have no idea how to use this method, I would appreciate, any kind of guidance or suggestion.

Thanks

Answer 1

[elenichri]

Hello,

I recently read the article http://www.ncbi.nlm.nih.gov/pubmed/12454644 and I think it responds to your question. What they apply is the geometric mean of replicates. For example, if you have A1, A2 and A3 biological replicates of A, you can combine them by sqrt(A1xA2xA3) (square root of their product)

Hope that helps!

Eleni