Real-time PCR, Volume IV: Clinical qPCR

Table of Contents

Letter from the Editor
Index of Experts
Q1: What protocols do you have in place to ensure a sensitive and reproducible assay?
Q2: How do you verify, validate, and optimize test performance?
Q3: How do you establish the number and frequency of controls?
Q4: What kind of biosafety measures do you have in place for nuclear extraction of infectious materials?
Q5: What quality assurance measures do you use?
Q6: What measures do you have in place to avoid contamination, both specimen-to-specimen and amplified product?
List of Resources

Download the PDF version here

Letter from the Editor

When Genome Technology came out with its first technical guide for qPCR almost two years ago, little did we know that we'd have enough content to fill four guides. This month, we shift focus to real-time PCR for clinical applications, presenting valuable advice from another roster of esteemed experts.

While the previous guides dealt with real-time PCR for nucleic acid quantification in general, this one draws on expertise from clinicians and microbiologists who are turning to real-time PCR for the rapid, effective, and safe identification and quantification of viruses, bacteria, and other infectious agents.

While our questions attempt to cover the basics, we also thought it'd be a great resource for those whose labs are trying to make CLIA certification. Several questions deal with special biosafety issues surrounding extraction procedures, cross-contamination, and laboratory protocols.

So, for tested tips and tricks on how to optimize your protocols, ensure reproducibility, and avoid contamination for your clinical qPCR runs, we hope you'll keep this guide on hand.

— Jeanene Swanson

Index of Experts

Genome Technology would like to thank the following contributors for taking the time to respond to the questions in this tech guide.

Stephen Bustin
Queen Mary's School of Medicine and Dentistry
University of London

Philip Day
Centre for Integrated Genomic Medical Research
The University of Manchester

Ralf Gilsbach
Institute for Experimental and Clinical Pharmacology and Toxicology
University of Freiburg

Mikael Kubista
TATAA Biocenter, Sweden

Ian Mackay
Queensland Paediatric Infectious Diseases Laboratory
Royal Children's Hospital, Australia

Roisin McNeill
Clinical Science Institute
National University of Ireland, Galway

Claudio Orlando
Department of Clinical Physiopathology
University of Florence

Mark Pandori
San Francisco Department of PublicHealth Laboratory

Andrew Sails
Newcastle Health Protection Agency Laboratory
Newcastle General Hospital, England

Robert Sjoback
TATAA Biocenter, Sweden

Jo Vandesompele
Center for Medical Genetics
Ghent University Hospital, Belgium

Q1: What protocols do you have in place to ensure a sensitive and reproducible assay?

A sensitive assay depends to a large extent on primer performance. Hence we spend a considerable time designing and optimizing primers to ensure (1) as few primer dimers as possible are generated and (2) the amplification efficiency of the PCR approaches 100 percent. We always use specific primers, hence we use Mfold to locate the RT primer into a loop structure of the RNA, thus ensuring optimal hybridization of primer to RNA and consequent RT efficiency. Our RT is carried out at temperatures of 55ºC and above. In our hands SYBR Green I-based assays are a little more sensitive than probe-based ones; hence we generally use SYBR Green chemistry. An up-to-date standing operating procedure is essential for minimizing technical variability and delivering repeatable and reproducible results. We use mostly tissue biopsies and minimize biological variability by taking at least two biological replicates from that biopsy through RNA extraction and RT-PCR assay. Assays are performed in duplicate and, if the biological replicates do not agree, we repeat the assay (if technical replicates are discordant) or analyze a third biological replicate. In general, we also use laser capture microdissection to ensure that we are extracting RNA from the appropriate target cells. If using fresh biopsies, we combine the lysis/extract on/assay procedure to ensure we do not lose any RNA using Invitrogen's CellsDirect kit. For formalin-fixed archival material, we pool and analyze consecutive sections, again making sure we have at least two biological replicate pools for analysis.

— Stephen Bustin

An important part of assay optimization for PCR and qPCR is the availability of standard templates. Because we always desire to use short amplicons for optimal assay efficiency (high sensitivity), it is readily possible and affordable to synthesize template oligonucleotides (TOs) to 80 bases in length that serve as synthetic targets. These not only make available a template for any target sequence, but they permit a means to quantify target molecules because the amount of TO molecules can be readily determined to permit a standard titration plot to be constructed using a 10-fold dilution series down to sub-single molecule. By using a relatively dilute TO (yielding a Ct/p value in the high 20s to allow complete removal with uracil-N-glycosylase if they are synthesized with two or more dUTPs) as a standard, the assay can be verified for reproducibility by observing the PCR value achieved for the reactions using TOs. This controls for the condition following storage of both PCR reagents and TOs, and we permit a deviation of up to 0.5 of a Ct value. Following repeated use of TO titration curves over numerous reactions, the typical reaction efficiency is established and the replicate measurements of a single TO dilution can be used to predict the precise efficiency of the full standard curve, because the tangent of the slope is known and this is a constant.

— Philip Day

To achieve a reproducible and sensitive assay we find it important to first optimize and then highly standardize each experimental step. A very critical point is the preparation of high-quality DNA [or] RNA in case of gene expression studies. We prefer column-based extraction methods since this method ensures intact and highly pure preparations. In contrast to phenol-chloroform based extractions the impact of different operators is not evident with column-based extraction.

— Ralf Gilsbach

The most important aspect of designing a sensitive and reproducible assay is in the primer and probe design. We always use more than one software to design and test the designed primers. For the sensitivity it is especially important to avoid any primer dimer formation since this would limit the sensitivity even for probe-based assays, by competing with the formation of the desired PCR product. This will be most notably at low copy numbers where the primer dimer formation may completely inhibit the formation of the desired product. In addition, the formation of primer dimers can be somewhat random and can therefore affect also the reproducibility. Another factor is to avoid primers that bind to regions of the template with high degree of secondary structure. This may affect the efficiency of primer annealing and thereby reduce the sensitivity. The degree of secondary structure can be checked by software such as Mfold (http://frontend.bioinfo.rpi.edu/applications/mfold/cgi-bin/rna-form1.cgi).

— Mikael Kubista and Robert Sjoback

My ideal approach to determine analytical sensitivity is to use synthetic templates. In the past these have taken the form of amplification product (amplicon) ligated with a TA cloning vector. The copy number is calculated from Avogadro's number and the molarity (calculated using the spectrometer-derived mass and the calculated molecular weight of the plasmid backbone plus the amplicon insert). To improve the reliability of this approach, I have in the past made serial dilutions and amplified each in replicates of six, employing Poisson statistics to determine the dilution which contains one to two copies.

Since I currently work on respiratory viruses, which mostly incorporate an RNA genome, I now make a T7 RNA polymerase promoter tagged forward primer, chosen from the preferred diagnostic assay, and use the product obtained from this amplification as the template for in vitro transcription of RNA. It's important to DNase digest away any DNA template — which can take two or more rounds of treatment. Success is proven when a PCR negative dilution of the preparation remains positive by RT-PCR (using the same single-tube RTPCR mix for both; the PCR tube has the RT inactivated by heat).

Reproducibility is influenced by many things, and in the study of acute respiratory tract infections, newly identified viruses (NIVs) seem to be popping up on a regular basis. Identifying them and then characterizing them using PCR-based techniques has to be as complete as possible before the results of epidemiology studies can be considered to be comprehensive. Early assay developers can suffer from subsequent problems as new strains of each NIV are subsequently reported. These variants may confound early oligonucleotides which results in falsely reduced frequency of detection.

— Ian Mackay

We use pre-designed and pre-validated TaqMan assays from Applied Biosystems. These assays are validated in silico using multiple primer design and evaluation softwares. We determine the efficiency of amplification and limit of detection (LOD) using a standard curve. We make a 10-fold dilution series spanning six to seven logs of a pool of two to three of the earliest products from an initial PCR run. After a post-PCR clean-up, product quantity is estimated from A260 readings and the dilution series is prepared starting with about 10 pg of amplicon. The assay is run in triplicate using the quantities for the standards. Once the run is complete, the PCR efficiency is determined from the slope of the plot of quantity versus mean Ct. The percent efficiency (E) is then E = [10(-1/slope)] - 1 x 100. I would then repeat the dilution series and assay using template cDNA and a semi-log plot in place of quantities of amplicon to check for PCR inhibitors co-purified during the extraction or RTstep. Inhibition can be more of an issue with RNA from FFPE tissues.

Intra-assay variation using the TaqMan assays is generally negligible and can be assessed by measuring the percent coefficient of variation (%CV) of the Cts of replicates of dilution standards in the same run. To assess inter-assay variation we include replicates of dilution standards in each 96-well plate. Generally the %CV of the Cts is less than 1 percent.

— Roisin McNeill

The real-time PCR assays we use are for the diagnosis and management of disease caused by viruses, bacteria, and fungi. To ensure an assay has sensitivity for the pathogen we are targeting we try to identify a gene target which is highly conserved in the organism. The availability of multiple genome sequences of the target pathogen in GenBank or other sequence databases facilitates the identification of suitably conserved target sequences, therefore ensuring maximum assay sensitivity. This helps to ensure that the assay has maximum sensitivity for all possible variants within the target strain, species, or genus. To ensure maximum sensitivity in terms of the limit of detection, new assays need to be thoroughly optimized and evaluated prior to introduction into routine use. To maximize reproducibility of the assay we ensure that all tests are performed according to very strict standard operating procedures.

— Andrew Sails

One of the most important aspects in the entire workflow from patient sample to a molecular diagnosis is the availability of a highly validated and robust assay. The performance of primers is evaluated at two levels: a first in silico evaluation using our established pipeline integrating 4 bioinformatics tools (http://medgen.ugent.be/rtprimerdb), followed by thorough experimental validation using a serial dilution series, one time gel electrophoresis and melting curve (if SYBR Green I detection chemistry is used).

— Jo Vandesompele

Q2: How do you verify, validate, and optimize test performance?

We always run standard curves to determine consistent PCR efficiency. We optimize our relative primer and Mg++ concentrations and establish melt curve profiles for amplicons. Initially, these are checked on gels and, if they contain appropriate restriction sites, are digested. If there is any doubt about an amplicon, we will sequence it.

— Stephen Bustin

By using the same design software and design parameters for each assay we usually get primers working well with similar conditions or at least a good starting point for optimizations if needed. We usually first verify the assays using non-specific DNA binding dyes, even when developing probed assays, to be able to check the degree of primer dimer formation by the dissociation curve. An alternative is to use our dye BOXTO as indicator for primer-dimer formation in probed based assays (Biotechniques 2006). We also always check new assays by gel electrophoresis to validate that we generate a PCR product of one defined length and that this length corresponds to the expected length of the amplicon. The second step is to check the assay efficiency and linearity by making a dilution series of the template. If optimization is required we usually first focus on primer and probe concentration, Mg2+ concentration and the annealing temperature.

— Mikael Kubista and Robert Sjoback

When looking for a new virus or the first instance of a NIV at our site, we start with two discrete, first-generation assays and then try to further optimize those assays using any positives we obtain from screening patient extracts. We can then make controls as described in answer 1, and employ the improved assay to search for more positives. The design of second-generation, real-time PCR assays, usually for use by a routine clinical diagnostic laboratory, is based on the sequences of more strains to comprehensively define, and hopefully avoid, nucleotide sequence variation, which reduces assay efficiency. Since some NIVs do not occur at high frequencies, even during their peak epidemic season, obtaining enough positives to satisfactorily evaluate the assay can take time. If that is likely, collaboration with other laboratories is especially important, otherwise the process of virus-positive specimen collection must continue over long periods, possibly limiting the application of the assay(s).

— Ian Mackay

In working with clinical samples it can be difficult, if not sometimes impossible, to standardize sample acquisition procedures, but all efforts should be made to minimize chances of RNA degradation. Fresh tissues should be snap frozen in liquid nitrogen or processed for microdissection as soon as possible after recovery and stored at -80°C.

RNA yields and extraction efficiencies vary depending on the source material and may be difficult to control for. At all stages of the extraction, precautions must be taken to ensure the activity of endogenous RNases is minimized and that exogenous RNases are not introduced. Following RNA isolation we assess total RNA integrity using the Agilent Bioanalyzer. MicroRNA can also now be analyzed using this system. For total RNA we have defined threshold RIN values for inclusion of RNA samples in qPCR analysis depending on the source of and if necessary after extraction also. We use a calibrated Nanodrop spectrophotometer to assess RNA concentration and purity, the latter being verified by determination of the A260/A280 and A260/A230 ratios. RNA is aliquoted and its integrity periodically checked.

In optimizing our cDNA synthesis reactions we assessed various enzymes and priming strategies for efficiency of cDNA synthesis. We amplify multiple targets so gene-specific primers are not an option unless the gene is very low abundance. We use random hexamers with the Superscript III enzyme. All samples are DNase digested during RNA extraction and we test for carry-over DNA using exon-spanning multiplex primers on a no-RT control sample included in each batch of samples, adding a post-extraction DNase treatment if necessary.

— Roisin McNeill

For absolute quantification, the optimization of standard curve is the initial requirement. Curve features (slope, y-intercept, and r2 values) should be as close as possible to the optimal expected performances and stable among different runs. When large differences among plates are observed, results cannot be accepted. For this reason a cut-off of acceptability should be preliminarily defined for each assay, and consequently the assays out of range discarded. The inclusion of control samples, possibly in a wide range of expected results (high, intermediate, and low dose), is mandatory for validation.

The kinetics of amplification for each unknown samples should be observed carefully, to exclude samples with abnormal amplifications (not parallel curves, absence of normal plateau effect, etc.)

For relative quantification, the use of a reference sample is mandatory for the comparison of different runs. Once again, the use of two to three controls serves to evaluate the reproducibility.

When intercalating dyes are used, melting curves should be performed for each assay and for each sample to prove the presence of a single peak and the absence of unspecific amplicons and primer-dimer peaks.

— Claudio Orlando

Like all other cases when a laboratory switches tests, we validate by comparing real-time PCR to a gold standard, aiming towards achieving a comparison of 70 specimens (50 positives and 20 negatives). However, as many of us know, PCR outruns the sensitivity of many older "gold standard" assays and this can complicate matters. In a shift from HSV culture to PCR, we found several specimens that were culture- negative but PCR-positive. How did we deal with that? The idea we utilized, was to use what I refer to as "second gene" confirmation. In the case of HSV, that meant actually confirming the real-time PCR with (instead of the culture gold standard) a second PCR which amplified a distinct HSV gene than the first real-time PCR. Hence, if I can find two HSV genes in a specimen, I can call it a real positive. This helps to get around the sensitivity gap commonly recognized between older assays and NAATs.

— Mark Pandori

Q3: How do you establish the number and frequency of controls?

We run two duplicate NTC, one -RT and two positive controls (in addition to the standard curves). One set of NTCs is dispensed and sealed prior to the assay, the second set at the end. If the NTC is positive, we check the melt curve and if they differ from the amplicon, we accept the result if Ct>5. If <5, repeat assay using twice the amount of template. For probes, if NTC is positive, we repeat the assay, unless the DCt is >8.

— Stephen Bustin

We integrate standard curves of diluted samples and/or known positive and negative controls in our reaction setups for each conducted gene depending if a quantitative or qualitative analysis is demanded.

In gene expression studies the stable expression of the assessed reference genes is a critical point. Therefore we introduce alien cRNAs and/or RNA amount as an independent normalization approach to confirm the results of our reference genes. In addition these results give us information about the quality of the starting material, i.e. are the Ct-values in a reasonable order of magnitude.

To detect occurring contamination of reagents with target DNA we perform routinely no template controls (NTC) and in case of gene expression studies also no reverse transcription controls (NoRT).

— Ralf Gilsbach

Each extraction batch should contain at least one negative extraction control consisting of known negative samples that are processed in parallel with the unknown samples to validate that no crosscontamination occurs between the samples. If the extraction is performed manually we usually include a negative extraction control for about eight to 10 samples. If an automated system is used, the number of controls will depend on the number of samples processed in parallel. For a 96-well based system we normally include three to four negative amplification controls spread out across the plate. Each sample should contain an extraction control to validate that the extraction was successful and that no inhibition of the PCR occurs. Preferably this should be in the form of a multiplexed internal amplification control that is co-amplified in each PCR tube to validate that PCR amplification has occurred in each specific tube. Also, for each assay and PCR amplification, no template controls (NTC) are used in at least duplicate to control that the reagents are not contaminated.

— Mikael Kubista and Robert Sjoback

Generally speaking, a weak to mid-range positive is a good idea to ensure that the assay is functioning within its predetermined limits of detection and that changes in reagents or operator error have not altered the sensitivity. One negative control for every 10 specimen extracts to be tested is a good rule of thumb.

— Ian Mackay

Apart from the regular controls for cDNA synthesis (e.g., RT-negative) and PCR (NTC, positive controls, etc.) we have paid considerable attention to the validation of multiple endogenous controls for the normalization of our relative gene expression data. Since the precision of the estimate of change in target gene expression is dependent on the stability of the endogenous control, the variability associated with the target gene and any covariance between these two terms it is essential that endogenous control genes be validated. We analyzed a large panel of the most commonly used EC genes for relative qPCR in breast cancer studies using the geNorm and NormFinder programs to identify the two most appropriate EC genes for use in our breast tissues and found that choice of EC can have a considerable effect on expression data. ECs validated for fresh frozen tissue gene expression analysis are not necessarily suitable for FFPE tissue analysis so multiple, small amplicons should be tested. Some groups persist in using either single or non-validated EC genes but I would strongly recommend careful, systematic validation of candidate endogenous controls genes before embarking on relative quantitation studies. It is particularly important when using endogenous control genes that they and the target genes of interest have similar amplification efficiencies and your relative quantity calculation should include a correction for differences in amplification efficiencies. We use the qBASE software to simplify the calculations involved in using multiple reference genes.

— Roisin McNeill

Conventionally, No Template Controls (NTC) are used to evaluate possible cross-contaminations in any scheme of real-time PCR applications. In the initial setup of an mRNA assay, we are used to evaluate some random RNA samples with and without reverse transcription (i.e, RT reaction withoutenzyme) to exclude interferences from residual DNA occasionally present in RNA extracts.

For positive controls we normally adopt two (less frequently three controls) covering the expected range of concentrations in unknown samples: high and low dose (intermediate in some cases). The use of a quality control chart (see question 5) is advisable.

— Claudio Orlando

Appropriate positive controls are required to demonstrate that the assay is performing correctly. The most basic control format may be positive control material in the form of nucleic acid extracted from a culture of the target organism. This only acts as a control to demonstrate that the amplification and detection process is working adequately. This type of control is now being supplemented or replaced by a more sophisticated approach in which a positive control virus or bacteria is "spiked" into the clinical sample prior to testing.

We also include working standards to control for inter-assay variability. These consist of either previously extracted nucleic acid from the target organism or an aliquot of simulated sample containing inactivated virus or bacteria. The samples are prepared by diluting the material to give a crossing point or Ct value of between 32 and 35 cycles. Significant inter-assay variation results in the Ct value of these samples changing between each individual assay run. Therefore plotting the Ct value of the working standards over time allows one to identify if an assay is "going off" with the Ct values being seen to rise over time. This can identify if there is a change in sensitivity of the assay perhaps due to a change in reagent batch.

— Andrew Sails

Q4: What kind of biosafety measures do you have in place for nuclear extraction of infectious materials?

We occasionally extract nucleic acids from human respiratory tract specimens. These arrive in our laboratory in screw-capped tubes. Wearing a lab coat, gloves, protective eyewear, and a surgical mask, we transfer an aliquot of each to a 1.5mL tube containing a guanidinium-based chaotropic agent which is expected to render any virus non-infectious. The process of extraction then proceeds according to standard laboratory chemical safety protocols, and as per the nucleic acid kit's manufacturer instructions. Vortexing is avoided and centrifugation is followed by an aerosol-settling step.

— Ian Mackay

We adhere to the University College Hospital, Galway, and National University of Ireland, Galway, health and safety guidelines for the handling of clinical specimens. Measures include individual risk assessment of specimens; laboratory chemicals and procedures; the use and proper maintenance of appropriate biosafety cabinets in the processing of clinical samples; vaccination and training of personnel; and the use of personal protective equipment. Key personnel are trained in emergency response procedures and any incidents are logged and reported.

— Roisin McNeill

All infectious materials are manipulated in a class II biosafety cabinet until they are placed into a lysis buffer or sealed into an amplification capillary (in the case of the LightCycler). We have assessed the various lysis buffers we use in the laboratory to see if they inactivate the infectious organisms we are working with and we have found that all of the lysis buffers are effective in that regard. Most lysis buffers contain either SDS or guanidinium, so this is not a surprise. For Mycobacterium tuberculosis, we first boil specimens for 25 minutes as the first step in the extraction.

— Mark Pandori

Our real-time PCR assays are integrated into our diagnostic clinical microbiology services, therefore we already had biosafety measures in place to protect laboratory workers when handling samples. In the UK pathogens are classified into different levels (1-4) of safety, each having their work practices, etc., to ensure safe laboratory working and to limit exposure to the laboratory workers. We strictly follow the guidelines published by the UK's Advisory Committee on Dangerous Pathogens in "The management, design and operation of microbiological containment laboratories."Handling samples which may contain category level 3 organisms can be a challenge; however, we have developed work practices which limit the risk to laboratory workers. For example we render such samples non-infective by lysing them using a guanidinium-based buffer prior to removing them from a biological safety cabinet. A good example of this is our influenza A (H5) assay in which patient samples are extracted using an automated instrument in our category 2 laboratory. Prior to them being loaded onto the instrument, the samples are rendered non-infective by the addition of the lysis buffer in the total nucleic acid extraction kit. The samples can then be transferred from the safety cabinet in the category 3 laboratory into the category 2 laboratory and the extraction completed.

— Andrew Sails

Q5: What quality assurance measures do you use?

We use the Agilent Bioanalyzer to assess total RNA quality, our 3':5' assay for RNA integrity, and our SPUD assay for the presence of assay inhibitors. We are also in the process of developing a comprehensive set of tools to allow a more accurate assessment of RNA integrity based on 3':5' assays carried out on 10 to 15 reference genes.

— Stephen Bustin

QA in a front-line research environment is not thought of in quite the same manner as it is for a clinical microbiology laboratory. Due to the paucity of standards for respiratory virus PCR, there is little that can be done to improve this situation. However, panels of amplicon at suitable copy numbers could be created by laboratories and sent to key collaborators to at least improve the overall reliability of PCR assays for NIVs; something we are contemplating once we determine the best target sequences to amplify. Incomplete genes would be preferable from a genetic manipulation regulatory point of view, but agreement must be reached on the best target for a given NIV.

— Ian Mackay

We have various QA measures in place, some of which are a statutory requirement to maintain laboratory accreditation. For example, we record the batch numbers and expiry dates of all of the reagents used in each assay run. We use reagents from a reputable supplier. We get another scientist to validate or confirm assay test results prior to them being reported. We also re-test samples in a "blinded" fashion to provide additional confidence in our test procedures.

External quality control schemes also play a very crucial role. The first external quality control scheme to be developed was the European Union Quality Control Concerted Action for Nucleic Acid Amplification in Diagnostic Virology. This temporary entity has been superseded by Quality Control for Molecular Diagnostics, a nonprofit organization for the standardization and quality control of molecular diagnostics and genomic technologies (www.qcmd.org). This organization sends out proficiency panels of simulated clinical samples containing a wide range of viral and bacterial pathogens for molecular diagnostic assays. Over 100 laboratories from more than 60 countries regularly participate in the program, which is endorsed by the European Society for Clinical Virology and the European Society for Microbiology and Infectious Disease. Laboratories providing molecular diagnostic testing should participate in this scheme to ensure quality of testing.

— Andrew Sails

Nucleic acids extracts are tested for presence of enzymatic inhibitors using the SPUD assay in which the amplification of a potato gene spiked into the patient template is evaluated. For RNA extracts, we also perform integrity controls such as a capillary gel electrophoresis and 5'-3' PCR amplification of a universally expressed reference gene. For gene expression-based assays, the stability of the applied reference genes is monitored in each experiment using proven methods and software (http://www.biogazelle.com).

— Jo Vandesompele

Q6: What measures do you have in place to avoid contamination, both specimen-to-specimen and amplified product?

PCR is always set up in a separate room in a class II cabinet. Separate pipettes are used for dispensing reagents and samples. All tips are filter-tipped. If possible, we use our Corbett CAS robot to dispense reagents. Following amplification, tubes are not opened. If they need to be, say for running a gel, that is done in a different building.

— Stephen Bustin

The most important is to employ a good laboratory practice and to have a defined working procedure. Separate rooms or at least areas for the pre-PCR handling where no template nucleic acids are handled and the template is added, and most importantly separate rooms for the post-PCR handling and extraction which both handle samples with high template concentration. It is also important to have a good sample work flow so that no samples are taken from a later stage in the procedure to an earlier, for example from the post-PCR to the extraction room, and to use different lab coats in the pre-PCR, extraction and post-PCR areas. To validate that no contamination occurs we include a set of controls. Negative extraction controls are included to check for cross-contamination between samples, during the extraction and handling of the samples, and no template controls are included to check for contaminations of the reagents. To minimize the risk of carry-over contamination we use the UNG system to degrade any potential contamination of amplified products.

— Mikael Kubista and Robert Sjoback

We employ four separated rooms (not separate air conditioning) and employ uni-directional workflow practices each day for (i) template extraction, (ii) reaction mix and template addition, (iii) thermal cycling and amplicon detection, and (iv) cloning.

— Ian Mackay

One of the major concerns, given the sensitivity of qPCR, is sample-to-sample carry-over during the initial stages of sample preparation, especially when extracting RNA from frozen tissues. Where microdissection is either not required or possible, it is generally necessary to fragment tissue into smaller pieces before homogenization, and it is important at this stage to use disposable plastics and frequently change gloves.

Another potential source of sample-to-sample carry-over is the homogenization step. We dismantle the homogenizer bit after use and clean it thoroughly. After sonication it is autoclaved twice and baked at 80°C. Tissues are homogenized in groups according to histopathological parameters and between each sample the bit is dismantled and cleaned thoroughly. Before homogenizing the next group of samples the bit is dismantled, washed, sonicated, rinsed, and autoclaved again. We use two bits to reduce downtime.

Throughout the remainder of the extraction procedure the typical precautions are followed, such as changing pipette tips between samples. RNA is prepared and processed under laminar flow using a dedicated set of pipettes in a separate room from the thermal cycler and post-run analysis equipment and amplified products are confined to that area before disposal.

— Roisin McNeill

To reduce the likelihood of specimen-to-specimen contamination good laboratory practice should be followed. The use of automated extraction instruments and PCR set-up robots can reduce the possibility of manual pipetting errors. The likelihood of specimen-to-specimen contamination is particularly increased if samples containing high numbers of template molecules (such as a sample with a high viral load) are processed in the same run as samples which are most likely to contain very low or negative titres of the same target. An example of this would be the processing of a genital swab for herpes simplex virus, which may have a very high viral titre in the same run as a set of cerebrospinal fluid specimens from patients with possible viral meningitis where the viral titre may be very low. To try and reduce the possibility of specimen-to-specimen contamination between such samples, we run them on different instruments on different runs to try and reduce the likelihood of contamination.

— Andrew Sails

The use of a close-tube system (never opening the vial with amplified PCR products) significantly reduces the chance of carry-over contamination of amplified product. Specimen-to-specimen contamination is reduced by using a validated automated nucleic acids extraction system.

— Jo Vandesompele

List of resources

Our panel of experts referred to a number of publications and online tools that may be able to help you get a handle on running clinical qPCR. Whether you're a novice or a pro, these resources are sure to come in handy.

Publications

Nolan, T., Hands, R.E., Bustin, S.A. Quantification of mRNA using real-time RT-PCR. Nature Protocols 1, 1559-1582 (2006).

Nolan, T., Hands, R.E., Ogunkolade, B.W., Bustin, S.A. SPUD: a qPCR assay for the detection of inhibitors in nucleic acid preparations. Anal Biochem 351, 308-310 (2006).

Gilsbach, R., Kouta, M., Bönisch, H., Brüss, M. Comparison of in vitro and in vivo reference genes for internal standardization of real-time PCR data. Biotechniques Feb;40(2), 173-7 (2006).

Espy, M.J., Uhl, J.R., Sloan, L.M., Buckwalter, S.P., Jones, M.F., Vetter, E.A., Yao, J.D, Wengenack, N.L., Rosenblatt, J.E., Cockerill, F.R., Smith, T.F. Real-time PCR in clinical microbiology: applications for routine laboratory testing. Clin Microbiol Rev Jan;19(1), 165-256 (2006).

Vandesompele J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., Speleman, F. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 3(7), RESEARCH0034 (2002).

Andersen, C.L., Jensen, J.L., Orntoft, T.F. Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res 64(15), 5245-5250 (2004).

Hellemans, J., Mortier, G., De Paepe, A., Speleman, F., Vandesompele, J. qBase relative quantification framework and software for management and automated analysis of real-time quantitative PCR data. Genome Biol 8(2), R19 (2007).

Pattyn, F., Robbrecht, P., De Paepe, A., Speleman, F., Vandesompele, J. RTPrimerDB: the real-time PCR primer and probe database, major update 2006. Nucleic Acids Res 34, D684-688 (2006).

Longo, M.C., Berninger, M.S., Hartley, J. L. Use of uracil DNA glycosylase to control carry-over contamination in polymerase chain reactions. Gene 93, 125-128 (1990).

Real-Time PCR in Microbiology: From Diagnosis to Characterization. Ian M. Mackay, Ed. Caister Academic Press (2007).

PCR Troubleshooting: The Essential Guide. Michael L. Altshuler. Caister Academic Press (2006).

Real-Time PCR: An Essential Guide. Kirstin Edwards, Julie Logan and Nick Saunders (eds). Horizon Bioscience (2004).

The management, design and operation of microbiological containment laboratories. HSC, Advisory Committee on Dangerous Pathogens. HSE Books (2001).

Websites

www.qcmd.org
http://frontend.bioinfo.rpi.edu/applications/mfold/cgibin/rna-form1.cgi
http://medgen.ugent.be/rtprimerdb/
www.biogazelle.com