gene expression in log scale and qPCR miRNA expression data

[Sharmi]

Hi,

I am new to genomics and I have a few questions on gene and miRNA expression data -
1. to convert gene expression data to log scale we do - log(base2)(Gene expression +4). Why do we add a 4? I searched for answers online but did not get one.
2. I want to do a linear regression on gene expression data and miRNA data, i. e. if Y is gene expression data and A is miRNA expression data and X is the regulatory binary matrix between the two, then Y=X.A. But the gene expression I have is micro-array data and the miRNA is qPCR data. Is it possible to still use the regression equation? The papers I have read on mRNA-miRNA regression all use micro array data so I am not use how to use the data I have.

Thanks so much!

[diego diaz]

I think that 4 is a pseudo-count. It is used because the log of 0 is -inf. Then, when you transform your expression data to log, you have to add a pseudo-count to avoid those -inf values, generated for genes with 0 expression value.