220.127.116.11. Internal reference standards
Unfortunately, external standards cannot correct for factors unique to each sample that affect the RT and/or PCR reactions, such as RNA quality and quantity, enzyme inhibitors, sample degradation, internal fluorescence etc. To correct for these factors, internal reference standards are used. These come in two forms:
“Exogenous internal reference standard”, which is a pure, unrelated RNA of known concentration that is added to the RT mastermix prior to RT-qPCR (Tentcheva et al., 2006). The amount used should be <1% of the amount sample RNA, so as not to affect the RT-qPCR reaction efficiencies. These are used to calculate cDNA reaction efficiencies of individual samples (correcting for RT inhibitors).
“Endogenous internal reference standards” (commonly called ‘housekeeping genes’), are relatively invariant host mRNAs present in every sample. These can be used to normalize quantitative data for differences between samples in RNA degradation, as well as inhibitors (Bustin et al., 2009; Radonić et al., 2004) and to guard against ‘false-negative’ data (due to RNA degradation).
There are a couple of practical difficulties with endogenous internal reference standards. First, one can never be certain that they are truly invariant (Radonić et al., 2004). The current recommendations are therefore to use an index of 3 or 4 endogenous reference standards for data correction (Bustin, 2000). Second, contaminating genomic DNA in an RNA sample can interfere with accurate quantification of the endogenous gene mRNA. This can be avoided by digesting the RNA sample with DNAse prior to RT-PCR, or more elegantly by designing intron-spanning primers for the endogenous reference gene (Bustin, 2000; Yañez et al., 2012; Locke et al., 2012), such that only cDNA to the mRNA can be amplified.
Internal reference standards are costly, since they are run for all samples. Their inclusion should therefore be evaluated in relation to their importance to the project; more for fully-quantitative experiments and fewer for semi-quantitative surveys.