Houskeeping Gene not steady - (Mar/04/2009 )

Hello all,

I am running real time PCR on treated cells and my HKG, GAPDH (which is verified) is not staying steady. It seems to increase by about 1.5-2 cycles as the dose increases and then drops off at the highest dose back to control levels. The replicates are tight and this phenomenon is repeatable among 3 different experiments. What am I doing wrong? What can be done to have a HKG stay steady?

Thanks!
DrDrew

drdrew17 on Mar 4 2009, 02:29 PM said:

If you have a good repeat, you did nothing wrong because there is such gene that always consistently express under all experimental conditions or treatments; especially GAPDH. GAPDH seems to have most reports indicate that it is NOT a stay steady HKG for normalization under various conditions. It is very likely GAPDH just change expression level under your conditions/treatments.

I think your main task is NOT doing something that can make your housekeeping gene(s) stay steady, because you have experimental condition need to maintain. Your task is to find one HKG that fit your situation.

That is why many researchers begin using multi-internal controls for normalization or at least perform internal control selection before running all samples for real-time PCR.

Again, DO NOT even try to make GAPDH works for you, it is biological response and there is nothing you can do about. Just find another HKG that works best for you.

wuxx0153 on Mar 5 2009, 08:16 AM said:

I also came up with similar problem. I change the internal control from GAPDH to 18sRNA. 18sRNA are more steady in my condition than GAPDH. I also try beta-action, and similar problem as GAPDH. And I would try globulin. I found many paper use two internal control for qPCR. May it help you and good luck.

HKGs don't guarantee exactly the same readings all the time. Even the best HKGs only change very little, not none at all. Different HKGs are best for different tissues/cell population. Maybe if you are hardworking, you could try 4-5 HKGs for your sample, take the one which changes the least, or the average of several HKGs.

chrisbelle on Apr 6 2009, 12:35 AM said:

I usually do selection before run all genes.

The housekeeping genes I test at first round include:
β-actin, HPRT, 36B4, β-2 microglobulin

and if second riund needed they are:
GAPDH, PPIA, Histone, α-tubulin

We still have: S26, TATA box binding protein, β-glucuronidase, ubiquitin C
as third round back up.

The lab next door has another way.
After finished microarray first, they selected 3 genes that have smallest change from array data and test with real-time PCR. They end up used a non-housekeeping genes as internal control, but it worked perfectly well.
Of course they need some additional explanation in their manuscript.

wuxx0153 on Apr 6 2009, 10:17 PM said:

I didn't understand what you do.
Do you do the first round and if one is constant you use it?
Do you do all rounds in any case?

molgen on Apr 7 2009, 07:22 AM said:

We try 4 genes first, and usually it can get a relatively good result, and we use the best one.

If first round does not produce good result, we will try the second set of 4 genes.
So far only one person in our lab went to third set of genes.

The problem I have is since I did not use all potential samples during selection, only use 20% of samples, and after used all samples with a particular control genes, the consistency among all samples may differ from the consistency among those samples test for internal control selection. Very often, it looks good when fewer samples are used, but it turns less consistent when all samples used.

My boss does not want me use ("waste" in his word) all samples just for selection, is anyone have other suggestion?

There's a couple of good programs that do internal normalization of candidate HKGs by some kind of complicated algorithm. One is called Genorm and is available free online. At the very least, this could provide a way to validate your selection of a given gene as a housekeeper.

The website is:
http://medgen.ugent.be/~jvdesomp/genorm/