I have had a similar problem with significant differences in GAPDH expression between treatment groups. I don't think that any significant difference in your housekeeping gene is acceptable, because it makes any differences in your target genes difficult to interpret. Therefore, the best thing is to find a housekeeping gene that does not differ by treatment. My new favorite housekeeping gene is OAZ1, which I discovered after reading a great paper by de Jonge et al. (2007), "Evidence based selection of housekeeping genes" in PloS one. This paper compared a vast array of genes in human and mouse samples and provides a list of the most stably expressed genes--I would start there to find the one that works best for your experimental conditions.

Hi there, I think its recommemded that it should be no more than 0.5 Ct cycle of a difference between your housekeeping gene

If the change in GAPDH has been observed only by looking at the cT values, that would not be enough to make a correct statement. For the same reason, reference genes(RG) are used to make relative quantification of genes of interest (GOI).

In case your RG-GOI dcT values are big enough, this 1cT variation in RG should not be a problem for you.

If you suspect that the RG that you use is not stable between conditions and you achieve a lower significance due to this, I would recommend using an additional one like ß2m, HPRT etc.

Please keep in mind that GAPDH has a huge pseudogene problem, if the primers you use are not selected properly they would also amplify gDNA.

Thank you!

I suggest that you should use more than one RG. Comparing their stability you'll find the better one!

And yes I'm agree with Mr. Ayata!

Good luck!

Because you cannot be absolutely sure that the amount of template you add is the same in each tube (due to tips, handling etc.), I agree that only the Ct is not a good reason to accept or deny the RG. Perform your expreriments in triplicates, and use at least two different dilutions of the same sample for two genes (one is the RG). Do it for both treated-untreated cells. The ΔCt will give you a first idea of the RG. If you can use more than one. I recomend 18SrRNA, which is very stable, and gives high Ct.

Hi,

You should always validate the reference genes for the conditions (cells, stimuli, time og stimulation etc…) that you are using.

I suggest that you run a plate with all your experimental conditions and test several possible reference genes (18s, GAPDH, HPRT1, TATA box, Cyclophilin, Actin, etc…)

Then you can experimentally determinate the best reference gene or pair of reference genes for your experimental conditions by using free software’s such as GeNorm or NormFinder.

You can find these software’s and a lot of information about qRT-PCR technique in this link:

http://www.gene-quantification.de/hkg.html

Hope it help,

Well may be 1 Cq unit it's not so big difference, that depends on a lot of things. For example, was your cDNA coming from the same amount and quality of RNA (I mean, as much as you can be sure about that from the systems implemented in your lab)? The idea is that, at the same conditions, we should have comparable amounts of cDNA, thus comparable Cq of our RG. However, as we know, that is only happening in an ideal world. Thus, RGs at the end alow us to compare the relative expression - basically as correction factor for the loading/synthesis/amplification efficiency - asuming that our RG is not altered by our treatments. Tips to handle that, as other people said, test more than one RGs to find the better for your experiments, increase your replicates and try to keep the most conditions for the RT-qPCR steps (RNA extraction, amounts, purification, cDNA synthesis, PCR reaction, etc etc etc) as constant as you could.

Good luck.

Hello Enrique,

in fact there are maybe three ways of you dealing with the situation.

The easier would be for you to compare stimulated cells with its GAPDH expression and control non-stimulated cells with basal GADPH.

A more elegant way would be if test more than one house keeping gene (HKG) when you run your qPCR.

Nevertheless you should always check which HKG is not really regulated in your experimental setup. Than you should test low copy constitutive non-regulated HKG. You may try our lab. favorites 18S, HPRT, B2M or Bact.

Hope this helped.

Good luck

Thanks for those criteria, Tiago.

B2M is now among our favorite candidates. It is very stable in T cells in different activation states, and has low variance between subjects. So we'll see if its efficiency is as beautiful!

I have had a similar problem with significant differences in GAPDH expression between treatment groups. I don't think that any significant difference in your housekeeping gene is acceptable, because it makes any differences in your target genes difficult to interpret. Therefore, the best thing is to find a housekeeping gene that does not differ by treatment. My new favorite housekeeping gene is OAZ1, which I discovered after reading a great paper by de Jonge et al. (2007), "Evidence based selection of housekeeping genes" in PloS one. This paper compared a vast array of genes in human and mouse samples and provides a list of the most stably expressed genes--I would start there to find the one that works best for your experimental conditions.

Hi Enrique, the best way to be sure that your control is valid is to try your control gene under the conditions in your experiment. If you have enough RNA, you don't need to perform the experiment again. As suggested above, just run a separate plate with several candidate control genes. The one with the lowest Ct difference between treated and untreated (control versus experimental) is your best control gene. Ideally, you can use the results from an RNAseq experiment to find the best possible gene with the lowest difference between treated and untreated. These experiments are expensive, so it is understandable if you don't have these results readily available. Hope that helps!

Hi I just ran my samples and I got a huge variation btw Ct values of cyclophilin. They range from 14 to 19. It's probably due to poor RNA integrity then?