So I have some Affy microarray data for miRNAs and I'm wondering what normalization techniques to apply to it. As I understand it the challenges inherent in miRNA nomalization are, unlike mRNA:
1) the smaller number of miRNAs as opposed to mRNA means there is less to work with
2) tissue specific miRNA expression means you are working with even less
I'm currently looking at a histogram of miRNA expression values that have a LOT of outliers, there is a sharp peak near the low-expression end of the curve then a very VERY long, low tail trailing into the higher levels of expression.
I have not yet removed low expression probes from the data and, being new to miRNA microarray analysis, I was wondering if this might be due to unexpressed miRNAs only registering at background levels for the arrays?
Also, is there a consensus on which normalization strategy works best for miRNA? Ie. centering to median values, quantile normalization, variance stabilizing normalization etc.?
I'd be very grateful for any advice you can offer.