Replication of qRT-PCR experiments - (Mar/29/2006 )

Hi everyone

I´m rather new on this area, hope some of you can help me solve a practical problem concerning my experimental qRT-PCR setup
I have just started doing relative gene expression quantification using the Ct method. I work on a ABI 7500 workstation. I have allready validated the efficiency of my targets genes vs internal control and this worked well. At the moment I´m running my first plates - studying the actual gene expression between targets and my reference. I perform 6 replica of each gene per reaction plate. (ABI recommend at least 4 replica per gene). However one question hunts me...how many plate replications should i perform?

If you have comments on this - please let me know!

Mema

Until you get the results you want.... No just kidding. I'm not sure what method you use, and i'm also relatively new. First thing I look at is negative control, it should be as negative as possible, say at least 10 cycli later than your target genes (but preferably not at all of course). Furthermore your gene replicas should not differ too much from each other. If you can you can replicate your plates, but if youve already used 6 replica's chances are you'll find similar results. Hope it helps

David

Thanks a lot for your reply David
I agree about the negative controls. Just to ad...had a talk with the ABI support today. They recommended at least 3 replicate points per sample and assays, as well as 3 replicates for negative controls where the assays are present but you add water instead of cDNA.
If the Ct values are high the standard deviation is going to increase. Once you get Cts above somewhere around Ct 33 your std deviations will increase, and then the only way to decrease that is to run more replicates..... I do 6 replica per gene as my std has been quite high.
When I set up my replicates, I make a mix of assay, PCR Master Mix and water, dispense in the wells and then add my cDNA sample separately to each. I suppose this is it what one would call biological replicates.

The most rigorous experimental procedure is of course to start from the very beginning 3 times (or more) over with more cells, RNA extraction and cDNA conversion and then running on real-time PCR. Then it is truley 3 separate experiments. If they all show the same thing, it must be so!!

But so far I have come to the conclusion that to run more than one plate of the same assays and samples from the same RNA-prep should not be necessary when I have replicates on one plate!!!


Mema

Mema, I completely agree. I usually do 3 replicas per plate, and lately ive been getting some strange results: only one sample worked, and got a decent ct (30-35), but the other two didnt. I ran them another time: same results. So what do you do when such a thing (like only 1/2 of your samples working) happens? Cannot really analyse it for your SD is incorrect.

Also, I add my cDNA to the wells first, before adding the PCR Mix, which consists of Master Mix, water and primers. I dont think it really matters though.

David

Hi again david
I analyse my results using the ABI SDS1.2 vers (ABI 7500 system). I´haven´t done a ton of platestudies so far but of some 6 runs per gene I do see outliers. In the software it is possible to remove outliers manually or by default. What system are you working on -ABI?
With my relatively poor knowlegde ont his topic I suggest you should do more than tree replica per gene...(allthough it is a costly affair)
I´m an old school biologist, usely I allways do more then 5 replica...mostly ten. However, just last year I came to work with microarray technology...in this workfield usualy no more that 3 replica is made!

Sorry my mistake, wrote it wrongly...I also ad my template/cDNA before adding matermix. Actually I don´t know whether or not this matter.

David, about the missing amplification in some of your wells.....I´ve been told that nonuniformity in results could be due to evaporation of your sample during the PCR run. So mayby you should check the cover or seals you are using. However, the fact that you are getting no results at all are probably not due to this and something else most likely went wrong!
Mema