Help! ignoring bad results in standard curve - about the pcr efficiency... (Apr/03/2008 )

Hi guys!

I'm doing real time pcr with the standard curve method. I'm using gDNA to construct it, and the efficiencies are sometimes good (never excellent, actually) and some times very bad... I dont have the money and the time to try to improve this with further optimization, primer re-designing, etc..

So, when a standard curve doesnt look quite well, sometimes the efficiency can be improved erasing some of the dilutions of gDNA, so the program doesnt consider them and the efficiency is recalculated. But is this acceptable? I mean, if I worry about keeping all my unknown samples into the standard curve, is this 'improvement' valid?

Please let me know what you think or if you have done this..

Thanks!!

---

How many replicates do you use for each standard point in your curve? Just one sample for each standard point may not be accurate since any error in that specific sample may lead to a big deviation in your efficiency.

If you're not already doing that, let me suggest you something. It is better if you use triplicates for each point, and take the median (not mean) value. You triplicate the cost of building the standard curve but it is much more reliable.

Deleting points from your curve is NOT good. The less points you have, the higher correlation (R square) you will get, but it is not a real correlation. (Suppose that you just have 2 points, then the correlation will be 100 %, since a perfect line passes through any two points.). So, don't erase your points

Good luck

---