Hi Karolina, the best option is to test several genes of comparable size and try to identify stable gene. A starting point would be to see what ref genes people have used for their FFPE tissues and try first these to see if they are stably expressed. Hope this help

Did you performed miRNA microarray profile between control & disease?

If yes, then you can choose one or more miRNAs as internal control for TaqMan, those showed steady (similar) expressions in both conditions.

Hi Vlady, thank you. That is a good option. I will have a look.

Harsh, in my studies I did not look at the control samples. I only had access to ovarian cancer samples. So I am really comparing miRs across tumours - so unnormalised data with U6 endogenous control. So in this case it will be difficult to use miR as a control. But that is a good idea. I did not think about that.

Thank you Marcus! Are you saying that they are good for ovarian tissue as well?

Karolina, Vlady's is the most practical approach. Look for what people have used before in PCR, particularly in ovarian tissue and then test them in your conditions. I have never heard on mir16 and 26b being used as endogenous controls. In our human kidney and liver tissue samples we use RNU44, RNUI48 and U6. Selecting something from microarray data is just not practical. There is too much data to sort through and the sensitivities of the techniques are different, so just because it was unchanged in the array it does not guarantee that you will not see a change by real time pcr.

If you use Expression Suite (can be freely downloaded from Life Technology) with your arrays/Taqman data, you can select "all" of your miRs as potential controls, and the program assigns them all a score. The lower the score, the more stably expressed your miR across samples-it is very helpful in selecting invariants from arrays. You can then simply re-assign the ones you are clearly not intersted in based on score as "targets" then take a closer look at the top miRs for invariance-I've looked at differences between contol and treated mean Cts and standard deviations in addition to the "score" assigned by Expression Suite to make my selections. Additionally, you can select miRs or snRNAs that you would typically used as endogenous controls and see how they perform. I just became aware of Expression Suite and it's saved me a lot of time and allowed me to look at my data from more angles, so I wanted to pass this along. Best Wishes!

In most of our profiling experiments we used multiple normalization procedures and take as bona fide results, those that score concordant regulation. If you run a profile usually you can normalize with the median expression of all the miRNA expressed in the samples or with the average expression of all the housekeeping included (U6, U5A, RNU44, RNU48). This is the safest way.

Alternatively you can use U5A and sno135 or sno202 in a single Q-PCR reactions and check if results vary with the two normalizations. All these genes also work in FFPE. Also check which HK are used in hypoxia experiments in the literature.

Dear Roberta thank you for the great advice! It is all well explained.

Markus, I will have a look at miR16 an miR26b. I might even do 2 different controls and compare results.

Mariano, I see in the papers that even researchers exploring hypoxia are still using U6. I am very interested in how that will develop in the future (to use U6, since it changes in hypoxia).

Thank you all for your help.

May be we can using the marker standart DNA

Dear Karolina,

It depends on your cells type but I have used miR-16 and Y1 with brain cells and tissue.

Karin