For an exact comparison of mRNA transcription in different samples or tissues it is crucial to choose the appropriate reference gene. The optimal reference gene should be constantly transcribed in all types of cells at any time in cell cycle and differentiation. Moreover the transcription of such a gene should not be regulated by internal or external influences, at least not more than the general variation in RNA synthesis. The reference gene used for normalisation of gene expression in real-time qPCR studies should also pass through the same steps of analysis as the gene to be quantified. However, such a perfect reference gene does probably not exist. The stability in expression of often used reference genes such as GAPDH and ACTB has been shown to vary considerably and are consequently unsuitable as reference genes for normalisation of gene expression analysis in some cases [5, 10–12]. Also the low expressed reference gene TBP is highly regulated when comparing normal and tumour tissues .
Numerous studies have been carried out in order to evaluate reference genes in specific tissues in several species. The majority of these studies are directed towards specific tissues [1, 3, 7, 14]. They clearly demonstrate that it is very difficult to find a 'universal' reference gene having stable expression in all cell types and tissues, and in particular to find reference genes that remain stable between samples taken at different time points under different experimental conditions. The first priority, however, is to identify genes with stable expression preferably across cell types since many real-time qPCR studies are performed on cDNA isolated from tissues with a mixed cell population.
There have been limited experiments of reference gene selection for use in other livestock and production animal species. Bogaert et al.  proposed ubiquitin (UBB), ACTB and B2M as reference genes for normalisation of qPCR data for normal equine skin. In cattle studies, De Ketelaere et al.  selected SDHA, YWHAZ, and 18S rRNA as being the most stable genes for accurate normalisation of qPCR of bovine polymorphonuclear leukocytes and Goossens et al.  found YWHAZ, GAPDH and SDHA to be the most stable reference genes across different preimplantation embryonic stages. Common for the bovine studies are the target of the studies, which are single cell populations. Garcia-Crespo et al.  compared expression of six potential reference genes in sheep tissues and found GAPDH with the lowest variation among the panel of six tissues, whereas ACTB and YWHAZ showed the worst score in variability. This is not in agreement with our findings but can be a result of the tissue composition. At present three studies have examined reference genes in pig i.e. Foss et al.  describes the use of GAPDH, ACTB and HPRT in immune cells and tissues using northern blot and PCR. In this study GAPDH was shown to be more stable than ACTB. Erkens et al  validated mRNA expression stability of 10 reference genes in porcine backfat and longissimus dorsi muscle. The most stable reference genes suitable for normalisation of qPCR data of backfat and longissimus dorsi muscle in the pig was ACTB, TBP and TOP2B. Kuijk et al.  investigated transcription levels of seven frequently used reference genes (inclusive B2M, ACTB, GAPDH) at different stages of early porcine embryonic development. Our study is the first detailed study of the stability and level of pig reference genes across a large number of tissues. Our findings confirm studies demonstrating tissue specific regulation of some of the commonly used housekeeping genes. Furthermore, it provides recommendations for choice of reference genes in studies where high and low abundant transcripts are under investigation. In agreement with the findings of Erkens et al.  both TBP and ACTB are suitable reference genes. However, our data show that ACTB is most relevant for high abundant transcripts, while TPB is most relevant for low abundant transcripts. It is clear from both Erkens et al.  and our study that GAPDH is quite unstable and not suitable as a reference gene.