## Proper controls

No template controls

Part of deciding where to set the threshold depends upon the negative controls. No template controls (NTCs) should always be included on every plate in every experiment! These controls assure that what is being

Figure 2.9

False amplification. A. Amplification view showing an acceptable amplification curve (a) and unacceptable curves (b, c) plotted using a logarithmic scale for the y axis (ARn). B. The same curves plotted using a linear scale for the y axis. C. Multicomponent view of sample b. D. Multicomponent view representative of one of the c sample wells.

measured is in the sample and not a contaminant. A good recommendation (Bustin and Nolan, 2004a) is to disperse NTCs over the plate. For example, prepare 2-3 NTCs in the first few wells on the plate and seal these wells before adding template to any of the other wells. Then prepare 2-3 NTC wells after all the templates have been added. This technique should help determine how the contamination is being introduced if positive NTCs become an issue. What if the NTC is reported as a Ct value less than 40? Inspection of the shape of the amplification curves can be informative. Sometimes a curve is seen that is gradually increasing, but not exponentially, and crosses the threshold at some point, as shown in Figure 2.9A.b. In other cases one might see curves that look like very low amplification curves (Figure 2.9A.c). Observing the shape of the curves with ARn expressed on a linear scale (Figure 2.9B) emphasizes that these curves are far from optimal. One cause is a decreasing ROX signal as the FAM signal remains the same, aptly called ROX drop. Another cause is a slight increase in FAM signal while the ROX remains the same, known as FAM creep. Both can occur to varying degrees and at the same time. Since most software divides the FAM signal by the ROX signal to normalize each well, the software interprets this change as amplification. Examination of the Multi-component view allows viewing of all the fluorescent reagents in an individual well so the relationship between the reporter signal and the reference dye can be observed as demonstrated in Figure 2.8C and 8D. Extreme cases can often be caused by evaporation resulting from improper sealing of that well. What is the solution to non-exponential or false positive NTC? In many cases, upward adjustment of the threshold is sufficient to eliminate the positive NTC. In some instances, the rogue well can be removed from the analysis if it is technically faulty. Careful attention to technique should minimize this type of problem.

What if the NTC show real exponential amplification? NTC contamination can be of several varieties. The contaminant can be a PCR product from a previous real-time PCR experiment. Theoretically, if the tubes from a realtime PCR run are never opened and disposed in an area far removed from the set-up area, then this type of contamination should not occur. However, in reality there might be occasions when the tubes are opened. For example, when an assay is being validated and the amplicon is run on a gel to check for a single product or purified to sequence the amplicon to check for specificity. At least one manufacturer of master mixes has addressed this problem by using dUTP (2' deoxyurindine 5' triphosphate) instead of dTTP and including the enzyme uracil-N-glycosylase (UNG). The UNG should digest any carry-over real-time PCR products that have incorporated dUTP. This is why there is a 2 minute step at 50°C step included in the thermocycling protocol if there is UNG in the master mix. The 2 minute step at 50°C inactivates the UNG so it does not affect subsequent PCR products being generated in the current run. A down side of using this enzyme is that if a little bit of residual activity is left, the UNG will begin to digest the amplicons in the current PCR reaction. This only becomes important if it is necessary to sequence the amplicon or run it on a gel. In this case, remove the samples from the machine immediately after the run and store them in the freezer until time to run them on a gel. How to determine if contamination is from carryover from previous real-time PCR runs? If a master mix containing dUTP but no UNG is used, add UNG to the master mix and see if this eliminates the positive NTC. Consider using UNG routinely. Of course, the ideal situation is to prevent the carryover contamination from happening, but this is not always practical, especially in multi-user situations. If dTTP is used in the real-time PCR reactions, then it will not be possible to determine the source of contamination by this method. What if dUTP and UNG are used in the reactions and the NTCs are still less than 40? Examination of the Raw Spectra and/or Multicomponent views shows the beginning of genuine exponential amplification. Primary contamination (as opposed to carry over contamination) is especially common when plasmids, PCR products and/or synthetic templates are used for standard curves. The standard becomes a contaminant unless stringent technique is exercised. Replacing all reagents, cleaning preparation sites, and moving to another building are all methods by which to escape from contamination. These measures are not always totally successful and the positive NTC situation must be dealt with. Likewise, if the samples are small and irreplaceable, there may be no choice but to deal with a positive NTC. Here is where some judgment much be carefully exercised. Here are some recommended suggestions (Bustin and Nolan, 2004a) that should be considered. If the NTC is between 35 and 40 cycles and the sample Ct is 25 or greater than 10 cycles lower than the NTC, then it is reasonable to think that the sample amplification is real. If the sample Ct is 30 and the NTC is 35 this should give serious pause. If the NTC is less than 30 cycles, then there is most likely a serious contamination problem. Another example of when NTCs need to be carefully examined is when the sample Ct values fall in the 35-39 cycle range. In this case, it is very important to verify that the sample is truly above background. One way to accomplish this is to add more template to the reaction. The NTCs should remain the same and the sample Ct should decrease 1 Ct for every two-fold increase in template. If the template is a limiting factor, as is frequently the case when using a small number of cells or micro-dissected tissue, then it is important to increase the number of cycles in the run to verify that there is greater than 10 Ct difference between the sample and the NTC.

### No reverse transcription control

If using a real-time PCR assay to quantify mRNA, then an important control is a reaction that allows one to assess the amount of genomic contamination in the sample. This control is commonly referred to as the minus reverse transcription (-RT) control. It is a sample of starting RNA that is treated identically to all the samples, but has no reverse transcriptase added during the reverse transcription step. How important this control is to the experiment and whether it is needed to include this control in every experiment is dependent upon the genes being examined and how the primer/probes were designed. What if the -RT control is positive? There are several strategies for eliminating or minimizing genomic detection. Some advocate designing primer/probed sets that span introns. Try to span a large intron rather than a small one. Some introns are less than 100 base pairs and could compete with the designated amplicon. Spanning an intron-exon border is not always possible, as with single exon genes and those with pseudogenes. Treating the samples with DNase is a good method to remove the majority of genomic DNA contamination from a RNA sample and is generally recommended. DNase treatment, however, usually reduces RNA yields, and this may be undesirable in the case where small numbers of cells are analyzed. If DNase treatment is not possible and the primer/probe set does not span an intron, a -RT control should always be included.

### Positive control

There are several types of positive controls, but one of the best is a standard curve. The merits of using a standard curve have been discussed earlier in this chapter. If using a standard curve for quantification, check the slope and the y-intercept to be sure that the standard is not deteriorating. If it is not possible to include an artificial standard curve for each primer/probe set, then a standard curve over a smaller range using genomic DNA or a universal RNA or a known positive sample as a template should be included to monitor the efficiency of the assay. Include a single positive sample if a standard curve isn't possible. If the situation arises where there is no amplification in the test samples, then having a positive control will aid in troubleshooting the problem and narrowing down the possibilities to reagent versus template problems.

Figure 2.10

Two differently labeled probes. A. Amplification view showing a probe labeled with FAM/BHQ1 (a) and a MGB probe from an Applied Biosystems Gene Expression Assay (b). B. The same in the Raw Spectra view. C. Multicomponent view of sample a. D. Multicomponent view representative of one of the b sample wells.

0 0