It is an age old problem:

How can you determine whether a gene is stable?

Compare it to an established stable gene!

How do you know that one is stable?

Compare it to an established...oh wait.

There is no real correct answer to this question, given that not only do levels of individual mRNAs fluctuate substantially, but the entire transcriptional milieu of individual cells can vary. For the most part, we simply accept that the reference genes we choose will vary, but will vary in a random fashion, and at a level substantially lower than the (non-random) transcriptional changes we are hoping to measure.

Also, the appropriate set of reference genes for a given comparison may vary: comparing changes in muscle gene expression will use different reference genes than comparing expression in liver.

So how does one determine appropriate reference genes a priori?

With maths!

And a large panel of gene candidates, and a representative collection of samples of interest.

There are several established methods for this, each using slightly different criteria to determine 'good' reference gene candidates.

There's the geNorm method of Jo Vandesompele:

Article Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, D...

There's the bestkeeper method of Michale Pfaffl:

Article Determination of stable housekeeping genes, differentially r...

And there's the Normfinder method of Andersen et al:

Article Normalization of Real-Time Quantitative Reverse Transcriptio...

All three can use the same data, and if you get similar answers from all three analyses, you can be pretty confident your reference gene candidates are strong.

interested

I suggest you refer to the article:

Rocha-Martins M, Njaine B, Silveira MS. Avoiding pitfalls of internal controls: validation of reference genes for analysis by qRT-PCR and Western blot throughout rat retinal development. PLoS One. 2012;7(8):e43028. doi: 10.1371/journal.pone.0043028.

Rafael Linden Thank you for your response - very helpful !!

Housekeeping genes by definition are genes that are more or less constitutively expressed in all tissues because they are responsible for basic continuous functions of cells, in contrast to genes that are regulated based on the type of tissue they are in or the environmental conditions or stimuli such as hormones. The housekeeping genes typically used with the expectation that their expression will not differ between experimental condition and control condition such as beta globin were found in early microarray experiments in the 90s to have little or no differences in their expression under varying conditions. In reality, all housekeeping genes also have some variability in RNA expression levels, so you must find out what is best for your experimental system, and ideally choose several. Also, the typical housekeeping genes were defined based on mammalian cell systems, so, for any other organism you should look for the appropriate genes (e.g zebrafish, insects, plants...).

It is an age old problem:

How can you determine whether a gene is stable?

Compare it to an established stable gene!

How do you know that one is stable?

Compare it to an established...oh wait.

There is no real correct answer to this question, given that not only do levels of individual mRNAs fluctuate substantially, but the entire transcriptional milieu of individual cells can vary. For the most part, we simply accept that the reference genes we choose will vary, but will vary in a random fashion, and at a level substantially lower than the (non-random) transcriptional changes we are hoping to measure.

Also, the appropriate set of reference genes for a given comparison may vary: comparing changes in muscle gene expression will use different reference genes than comparing expression in liver.

So how does one determine appropriate reference genes a priori?

With maths!

And a large panel of gene candidates, and a representative collection of samples of interest.

There are several established methods for this, each using slightly different criteria to determine 'good' reference gene candidates.

There's the geNorm method of Jo Vandesompele:

Article Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, D...

There's the bestkeeper method of Michale Pfaffl:

Article Determination of stable housekeeping genes, differentially r...

And there's the Normfinder method of Andersen et al:

Article Normalization of Real-Time Quantitative Reverse Transcriptio...

All three can use the same data, and if you get similar answers from all three analyses, you can be pretty confident your reference gene candidates are strong.

Greetings,

I know it's a little late, but if you decided to follow what John Hildard proposed, I'd kindly recommend you to use the 3 above mentioned software together, to finally try to find a correlation among the results by using Pearson's correlation. It's been reported more than twice that, in some cirumstances, you happen to obtain 2 Vs 1 results, by contrasting out the 3 methods. So pearson's would help you out discussing and finding a plausible results.

Good luck