Regression analysis of RNA-Seq data and qPCR data

[wilson90]

I have been trying to check the correlation between the RNA-Seq RPKM and qPCR expr value.

Is it possible that the linear correlation between log2(RPKM) and log2(qPCR) value is worse than the linear correlation between RPKM and qPCR expression value consistently?

*qPCR expression value here is referring to normalized ct value by comparing with one gene (as shown in ncbi GSE5350)

Thank you.

[Simon Anders]

Why would you compare a cycle number (log scale) with an RPKM value (linear scale), or take the log of a ct value (log-of-log scale)? Maybe remind yourself what a ct value actually means before proceeding.

[frcop]

You should indeed not use "expression" when you refer to Ct values. To go from Ct to expression you need to raise the efficiency to the power -Ct (presuming you've calculated the efficiency, otherwise you probably assume it's 2). It's also best to use multiple reference genes to normalize because otherwise you're putting a lot of fate in the stability of that one gene.