Absolute quantification and normalization
Posted 17 January 2010 - 10:38 AM
I'm quite familiar with real-time RT-PCR, but have never done any absolute quantification before and need some clarification regarding what's rattling around in my brain.
RNA was isolated from multiple tissues and developmental stages of my organism using standard techniques including DNase treatment and cleanup. I then used a NanoDrop to quantify my RNA, and used precisely 2 ug RNA to synthesize cDNA.
I have a target gene and primers that give me excellent amplification with efficiency = 100% (using a pool of cDNA from multiple tissues/dev stages) and a single peak in the dissociation curve. I have also sequenced the amplicon to confirm specificity.
The target gene has been cloned into a plasmid, which I will linearize and use for the standard curve for absolute quantification.
In the past, with relative quantification, I used internal reference genes (i.e. b-actin, 18S, eF1a, etc.). Now, with absolute quantification, do I need a reference gene, particularly since I've essentially normalized everything to total input RNA?
My confusion here is that if I use the delta-delta Ct method with a reference gene, once I subtract the Ct of my ref gene (say, 15) from my target gene (say, 19), I'd get a normalized Ct (in this case, 4). If I use this normalized Ct on my external standard curve (linearized plasmid), the copy number of my target would be vastly under-represented, right??
If I do need a reference gene for normalization, how do perform the absolute quantification???!
I've read a whack of a papers on the subject (many from Bustin or Pfaffl) but need some confirmation that I'm doing this correctly...
Thanks in advance!!
-- Needs Clarification
Posted 18 January 2010 - 03:05 PM
Here's a quote from Leong et al. 2007. Absolute quantification of gene expression in biomaterials research using real-time PCR. Biomaterials 28:203-210:
aqPCR can be more robust than relative quantification because aqPCR is independent of any normalizing gene. In aqPCR, the only normalization done between samples was the use of starting amount of total RNA that represented the entire transcriptome, thus decreasing the error.
Since you've normalized to total input RNA, I think you're fine.
Hopefully someone will correct me if I'm incorrect.