The Ultimate qPCR Experiment: Producing Publication Quality, Reproducible Data the First Time

qPCR is more complex than perceived by many scientists. The production of an amplification curve and an associated quantitative cycle value does not necessarily mean interpretable data. The MIQE guidelines and associated methodology articles published thereafter, underline the ongoing drive to help scientists produce reproducible data from qPCR, culminating in a simple, stepwise methodology to ensure high-quality, reproducible data from qPCR experiments. The concept of data normalization has led to the ongoing publication of articles solely focused on this subject for various sample types and experimental parameters. The analysis of qPCR data can be challenging, especially as experiments grow in sample number and complexity of biological groups. A defined approach to qPCR data analysis is necessary to clarify gene expression analysis. Quantitative PCR (qPCR) is one of the most common techniques for quantification of nucleic acid molecules in biological and environmental samples. Although the methodology is perceived to be relatively simple, there are a number of steps and reagents that require optimization and validation to ensure reproducible data that accurately reflect the biological question(s) being posed. This review article describes and illustrates the critical pitfalls and sources of error in qPCR experiments, along with a rigorous, stepwise process to minimize variability, time, and cost in generating reproducible, publication quality data every time. Finally, an approach to make an informed choice between qPCR and digital PCR technologies is described. Quantitative PCR (qPCR) is one of the most common techniques for quantification of nucleic acid molecules in biological and environmental samples. Although the methodology is perceived to be relatively simple, there are a number of steps and reagents that require optimization and validation to ensure reproducible data that accurately reflect the biological question(s) being posed. This review article describes and illustrates the critical pitfalls and sources of error in qPCR experiments, along with a rigorous, stepwise process to minimize variability, time, and cost in generating reproducible, publication quality data every time. Finally, an approach to make an informed choice between qPCR and digital PCR technologies is described. qPCR (see Glossary) is generally viewed by researchers as a powerful technique that can provide precise and quantitative data reflecting the biology of the tested experimental parameters. However, without following strict guidelines, validation and data analysis procedures, the results can be far from valid [1Bustin S. Nolan T. Talking the talk, but not walking the walk: RT-qPCR as a paradigm for the lack of reproducibility in molecular research.Eur. J. Clin. Invest. 2017; 47: 756-774Crossref PubMed Scopus (63) Google Scholar, 2Bustin S.A. Huggett J.F. Reproducibility of biomedical research — the importance of editorial vigilance.Biomol. Detect. Quantif. 2017; 11: 1-3Crossref PubMed Scopus (10) Google Scholar]. Unfortunately, the adoption and transfer of inadequate and varied protocols between individual laboratory members and laboratories throughout the scientific community have led to frustration in reproducing data [3Hayden R.T. et al.Multicenter comparison of different real-time PCR assays for quantitative detection of Epstein–Barr virus.J. Clin. Microbiol. 2008; 46: 157-163Crossref PubMed Scopus (100) Google Scholar, 4Zhang T. et al.Inter-laboratory comparison of chronic myeloid leukemia minimal residual disease monitoring.J. Mol. Diagn. 2007; 9: 421-430Abstract Full Text Full Text PDF PubMed Scopus (49) Google Scholar, 5Jaworski J.P. et al.Interlaboratory comparison of six real-time PCR assays for detection of bovine leukemia virus proviral DNA.J. Clin. Microbiol. 2018; 56: 617-628Crossref Scopus (16) Google Scholar]. This has driven the production of the minimum information for publication of quantitative real-time PCR experiments (MIQE) guidelines and related methodology articles to help the scientific community in augmenting experimental rigor and uniformity to produce more reliable and consistent data [6Bustin S.A. et al.The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments.Clin. Chem. 2009; 55: 611-622Crossref PubMed Scopus (10347) Google Scholar, 7Sanders R. et al.Improving the standardization of mRNA measurement by RT-qPCR.Biomol. Detect. Quantif. 2018; 15: 13-17Crossref PubMed Scopus (14) Google Scholar, 8Taylor S.C. Mrkusich E.M. The state of RT-quantitative PCR: firsthand observations of implementation of minimum information for the publication of quantitative real-time PCR experiments (MIQE).J. Mol. Microbiol. Biotechnol. 2014; 24: 46-52Crossref PubMed Scopus (58) Google Scholar]. Nevertheless, there remain concerns regarding the quality of qPCR results in the published literature [1Bustin S. Nolan T. Talking the talk, but not walking the walk: RT-qPCR as a paradigm for the lack of reproducibility in molecular research.Eur. J. Clin. Invest. 2017; 47: 756-774Crossref PubMed Scopus (63) Google Scholar, 2Bustin S.A. Huggett J.F. Reproducibility of biomedical research — the importance of editorial vigilance.Biomol. Detect. Quantif. 2017; 11: 1-3Crossref PubMed Scopus (10) Google Scholar]. When designing experiments for qPCR, all protocols, such as sample handling, harvesting, nucleic acid extraction, reverse transcription, and qPCR should be described and vetted in detail. Mistakes or assumptions can be made in the planning process, resulting in a flawed experimental design with results and conclusions based on artefacts of pre and/or post sample handling procedures as opposed to the true effect of the tested experimental parameters [7Sanders R. et al.Improving the standardization of mRNA measurement by RT-qPCR.Biomol. Detect. Quantif. 2018; 15: 13-17Crossref PubMed Scopus (14) Google Scholar]. Poorly optimized reactions can result in data that are consequent to a combination of sample contaminants and/or poor annealing temperature, leading to misinterpreted results and conclusions that are difficult or even impossible to reproduce [9Dijkstra J.R. et al.Critical appraisal of quantitative PCR results in colorectal cancer research: can we rely on published qPCR results?.Mol. Oncol. 2014; 8: 813-818Crossref PubMed Scopus (42) Google Scholar, 10Bustin S.A. et al.The need for transparency and good practices in the qPCR literature.Nat. Method. 2013; 10: 1063-1067Crossref PubMed Scopus (206) Google Scholar]. Despite the MIQE guidelines and other methodology articles, the variability and reproducibility pitfalls associated with qPCR remain elusive for many laboratories [7Sanders R. et al.Improving the standardization of mRNA measurement by RT-qPCR.Biomol. Detect. Quantif. 2018; 15: 13-17Crossref PubMed Scopus (14) Google Scholar, 11Sanders R. et al.Considerations for accurate gene expression measurement by reverse transcription quantitative PCR when analysing clinical samples.Anal. Bioanal. Chem. 2014; 406: 6471-6483Crossref PubMed Scopus (59) Google Scholar]. This review article describes the major sources of error associated with a qPCR experiment and strategies for their minimization, along with a rigorous, stepwise approach to producing accurate and precise results. Finally, a simple decision tree is proposed to choose between digital PCR (dPCR) and qPCR technologies to minimize cost and time to publication for any type of sample and target abundance. There are two major sources of error in any life science experiment that can lead to a high level of variability between test conditions: random and systemic error [12Althubaiti A. Information bias in health research: definition, pitfalls, and adjustment methods.J. Multidiscip. Healthc. 2016; 9: 211-217Crossref PubMed Scopus (967) Google Scholar]. Random error is associated with flawed experimental design parameters, which are primarily manifested in biological variability and subsampling error. Systemic error arises from improper use or calibration of equipment and computational software, leading to technical and calculation errors. Each can contribute significantly to the total error derived from an experiment, producing nonstatistically significant and/or artefactual data that is unrepresentative of the tested parameters. By isolating error sources and understanding how to minimize their impact through careful experimental design and good technique, solid, reproducible, and statistically significant results can be achieved that will stand the test of time in the literature. The most challenging and least considered aspect of many experiments is the appropriate selection of a randomized set of individual samples (i.e., biological replicates) per biological group (i.e., treatment/experimental conditions) while minimizing their inherent variability [13Suresh K.P. An overview of randomization techniques: an unbiased assessment of outcome in clinical research.J. Hum. Reprod. Sci. 2011; 4: 8-11Crossref PubMed Scopus (627) Google Scholar]. Both the transcriptome and proteome are highly sensitive to the inherent biological differences between samples. In fact, gross differences in transcription have been demonstrated between individual cells plated from the same passage in a single petri dish [14Moignard V. Gottgens B. Transcriptional mechanisms of cell fate decisions revealed by single cell expression profiling.Bioessays. 2014; 36: 419-426Crossref PubMed Scopus (21) Google Scholar, 15Battich N. et al.Control of transcript variability in single mammalian cells.Cell. 2015; 163: 1596-1610Abstract Full Text Full Text PDF PubMed Scopus (204) Google Scholar]. Therefore, careful thought and planning must go into the sourcing, randomized selection, and number of samples per biological group to ensure statistically significant and reproducible results that give the precision and accuracy for publication (Figure 1A). Apart from the inappropriate selection of a randomized set of samples that represent the population, nonreproducible error can also be introduced from: (i) lot-to-lot differences in growth medium or animal feed [16Lanoix D. Vaillancourt C. Cell culture media formulation and supplementation affect villous trophoblast HCG release.Placenta. 2010; 31: 558-559Crossref PubMed Scopus (7) Google Scholar]; (ii) variation in temperature, time, incubation conditions, and circadian regulation between samples (Figure 1B) [17Kang H.J. et al.Spatio-temporal transcriptome of the human brain.Nature. 2011; 478: 483-489Crossref PubMed Scopus (1286) Google Scholar, 18Gao S. et al.Tracing the temporal-spatial transcriptome landscapes of the human fetal digestive tract using single-cell RNA-sequencing.Nat. Cell Biol. 2018; 20: 721-734Crossref PubMed Scopus (77) Google Scholar, 19Panda S. et al.Coordinated transcription of key pathways in the mouse by the circadian clock.Cell. 2002; 109: 307-320Abstract Full Text Full Text PDF PubMed Scopus (1869) Google Scholar]; (iii) tissue sections or cell subpopulations extracted from a specimen (Figure 1C) [17Kang H.J. et al.Spatio-temporal transcriptome of the human brain.Nature. 2011; 478: 483-489Crossref PubMed Scopus (1286) Google Scholar, 18Gao S. et al.Tracing the temporal-spatial transcriptome landscapes of the human fetal digestive tract using single-cell RNA-sequencing.Nat. Cell Biol. 2018; 20: 721-734Crossref PubMed Scopus (77) Google Scholar]; (iv) particularly for female animal models, the synchronization of hormonal cycles [20Byers S.L. et al.Mouse estrous cycle identification tool and images.PLoS One. 2012; 7e35538Crossref PubMed Scopus (626) Google Scholar, 21Mitko K. et al.Dynamic changes in messenger RNA profiles of bovine endometrium during the oestrous cycle.Reproduction. 2008; 135: 225-240Crossref PubMed Scopus (99) Google Scholar]; and (v) for environment and plant material, time of day, amount of light, temperature, soil microenvironment, and in situ sampling and storage techniques [22Woo H.R. et al.Plant senescence: how plants know when and how to die.J. Exp. Bot. 2018; 69: 715-718Crossref PubMed Scopus (53) Google Scholar, 23Filipecki M. Malepszy S. Unintended consequences of plant transformation: a molecular insight.J. Appl. Genet. 2006; 47: 277-286Crossref PubMed Scopus (94) Google Scholar]. Each of these experimental design factors can significantly increase biological error and variability, which can ultimately lead to false conclusions. In most molecular biology experiments, a portion of the sample (a subsample) is tested. The error associated with quantification of the number of molecules from a portion of the sample and extrapolating the result to the total is termed subsampling error (Figure 2) [24Gerlach R.W. Nocerino J.M. Guidance for Obtaining Representative Laboratory Analytical Subsamples from Particulate Laboratory Samples. US Environmental Protection Agency, 2003Google Scholar]. For example, if a 1 μl subsample of cDNA is taken from a total volume of 25 μl with a concentration of six target molecules per μl, the expected number of copies in the subsample would be six. However, the result would likely range between four and eight copies. Assuming no technical error in pipetting, the deviation in the result from the predicted six copies is the subsampling error, which is much more pronounced in sample concentrations below about two copies per microliter (Figure 2). The coefficient of variation (CV) measures the variation around the mean, and is a useful tool to assess the degree of data inconsistency (Figure 2C). The standard deviation (SD) and CV associated with subsampling error can be calculated from the expected number of copies in a given sample as follows: M = expected number of target molecules under binomial approximationSD=MCV=MM From the previous example, the average number of copies for several 1 μl subsamples will be six but the SD of this measurement is (√6 = 2.45) and the %CV is (2.45/6*100 = 40.8%). Subsampling error contributes to more than 10% of the variance when the subsample is below 100 copies, and above 30% when the subsample is below ten copies (Figure 2C,D) [24Gerlach R.W. Nocerino J.M. Guidance for Obtaining Representative Laboratory Analytical Subsamples from Particulate Laboratory Samples. US Environmental Protection Agency, 2003Google Scholar]. This necessitates running more technical replicates (i.e., subsamples) for samples of very low concentration. One of the most common and problematic sources of technical variability in molecular biology stems from pipetting and the use of inappropriate pipettes and sample volumes. Many laboratories inadvertently use poorly calibrated pipettes coupled with tips and technique that may be inaccurate for small volumes [25Vaccaro W. Minimizing liquid delivery risk: operators as sources of error.Am. Lab. 2007; 39: 16-17Google Scholar, 26Carle A.B. Five good reasons: the argument for pipetting technique training.Med. Lab. Obs. 2013; 45: 58PubMed Google Scholar]. This can produce variability ranging between 5% and 37% CV [27Albert K.J. Bradshaw J.T. Importance of integrating a volume verification method for liquid handlers: applications in learning performance behavior.J. Assoc. Lab. Autom. 2007; 12: 172-180Crossref Scopus (19) Google Scholar]. It becomes a major problem when samples are serially diluted because the pipetting errors between each dilution are propagated, resulting in gross variability between data points [28Higgins K.M. et al.The effect of serial dilution error on calibration inference in immunoassay.Biometrics. 1998; 54: 19-32Crossref PubMed Scopus (46) Google Scholar, 29Walling L.A. An inline QC method for determining serial dilution performance of DMSO-based systems.J. Assoc. Lab. Autom. 2011; 16: 235-240Crossref Scopus (4) Google Scholar]. For qPCR, a good methodology to minimize pipetting error is to apply a 60:40 ratio of master mix to cDNA/gDNA sample (Figure 3A), which permits: (i) the use of the same pipette and tips, (ii) efficient mixing of sample in the reaction mix, and (iii) pipetting volumes that correspond to the midrange of the pipette for high accuracy. Nucleic acid contaminants represent another major source of error in qPCR. This necessitates the inclusion of no template controls (NTCs) on each plate to assess the level of cDNA or gDNA contamination that may stem from the individual reaction components (Figure 3A). No reverse transcription controls (NRTs) are also important to assess the levels of genomic DNA contaminants from the initial RNA extracts of a given project [30Hashemipetroudi S.H. et al.Assessment of DNA contamination in RNA samples based on ribosomal DNA.J. Visualized Exp. 2018; 131e55451Google Scholar]. Given the sensitivity of the transcriptome to even minor changes in the surrounding environment, care and rigor must be applied to ensure uniformity in: (i) sample handling throughout the experiment; (ii) activity/potency between lots of the same compound or reagent(s) used for treatment [16Lanoix D. Vaillancourt C. Cell culture media formulation and supplementation affect villous trophoblast HCG release.Placenta. 2010; 31: 558-559Crossref PubMed Scopus (7) Google Scholar]; (iii) tissue or cell harvesting and homogenization techniques; and (iv) sample storage and freeze/thaw conditions [31Kim S.J. et al.Effects of storage, RNA extraction, genechip type, and donor sex on gene expression profiling of human whole blood.Clin. Chem. 2007; 53: 1038-1045Crossref PubMed Scopus (46) Google Scholar]. RNA/DNA extraction procedures and storage can also play a significant role in data quality, and good kits are available to produce excellent RNA, DNA, or protein extracts to ensure consistency in the quantity, quality, and purity of extracted nucleic acids [31Kim S.J. et al.Effects of storage, RNA extraction, genechip type, and donor sex on gene expression profiling of human whole blood.Clin. Chem. 2007; 53: 1038-1045Crossref PubMed Scopus (46) Google Scholar, 32Tan S.C. Yiap B.C. DNA, RNA, and protein extraction: the past and the present.J. Biomed. Biotechnol. 2009; 2009574398Crossref PubMed Scopus (416) Google Scholar, 33Cankar K. et al.Critical points of DNA quantification by real-time PCR – effects of DNA extraction method and sample matrix on quantification of genetically modified organisms.BMC Biotechnol. 2006; 6: 37-52Crossref PubMed Scopus (178) Google Scholar]. However, the appropriate choice of extraction kit is critical to minimize background contaminants that can alter the Cq values. This is highly dependent on the sample and no single kit serves all applications [32Tan S.C. Yiap B.C. DNA, RNA, and protein extraction: the past and the present.J. Biomed. Biotechnol. 2009; 2009574398Crossref PubMed Scopus (416) Google Scholar]. For RT-qPCR specifically, significant variability can arise from the reverse transcription (RT) of mRNA to cDNA where the results for the same sample can vary by two to threefold depending on the amount and quality of mRNA used in the RT reaction [34Bustin S. et al.Variability of the reverse transcription step: practical implications.Clin. Chem. 2014; 61: 201-212Google Scholar]. Furthermore, the amount of target mRNA with respect to background can have a major impact on the downstream results, which can also be RT kit dependent [35Levesque-Sergerie J.-P. et al.Detection limits of several commercial reverse transcriptase enzymes: impact on the low- and high-abundance transcript levels assessed by quantitative RT-PCR.BMC Mol. Biol. 2007; 8: 93-111Crossref PubMed Scopus (60) Google Scholar]. Finally, the components of the RT reaction can be inhibitory to the downstream qPCR reaction [35Levesque-Sergerie J.-P. et al.Detection limits of several commercial reverse transcriptase enzymes: impact on the low- and high-abundance transcript levels assessed by quantitative RT-PCR.BMC Mol. Biol. 2007; 8: 93-111Crossref PubMed Scopus (60) Google Scholar]. Hence, care must be taken in adequately diluting the RNA and the resulting cDNA, to ensure efficient RT and qPCR reactions that reflect the true proportion of the mRNA target(s) in the original samples. When the target DNA is very dilute (less than ten copies per microliter), the primers may not land on all the template molecules in the first cycle. Thus, different fractions of the original starting template are amplified from the first and second cycles before complete amplification of all template has been initiated (Figure 3B). This can lead to large variability between the technical replicates of a given sample that can easily range between 10% to 200% CV, particularly when template is below 100 copies per reaction [i.e., the quantitative cycle (Cq) values are greater than 29 cycles]; (Figure 2D) [6Bustin S.A. et al.The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments.Clin. Chem. 2009; 55: 611-622Crossref PubMed Scopus (10347) Google Scholar, 36Peccoud J. Jacob C. Theoretical uncertainty of measurements using quantitative polymerase chain reaction.Biophys. J. 1996; 71: 101-108Abstract Full Text PDF PubMed Scopus (148) Google Scholar, 37Forootan A. et al.Methods to determine limit of detection and limit of quantification in quantitative real-time PCR (qPCR).Biomol. Detect. Quantif. 2017; 12: 1-6Crossref PubMed Scopus (277) Google Scholar]. This variability can only be minimized by interrogating a larger proportion of the sample from more technical replicates and using the average Cq for downstream calculations. One of the advantages of qPCR is the ability to perform multiplexed detection and quantification of several targets within the same sample using probe-based chemistry. This is particularly useful when working with very precious samples containing low quantities of nucleic acids. However, qPCR is entirely dependent on reaction efficiency where a Cq value is only representative of the template amount when the reaction efficiency is near 100% (Figure 4A–C). This necessitates a rigorous validation of individual targets in a representative sample (Figure 4B,C) to ensure their reaction efficiencies and annealing temperatures are similar prior to multiplexing, coupled with limited crosstalk between primers and probes when combined in one reaction [38Elnifro E.M. et al.Multiplex PCR: optimization and application in diagnostic virology.Clin. Microbiol. Rev. 2000; 13: 559-570Crossref PubMed Scopus (616) Google Scholar]. Multiplexed preamplification offers the advantage of increasing the starting quantity of a selected panel of targets to permit their detection and quantification from individual SYBR Green containing qPCR reactions without the constraints of probe-based multiplexing [39Andersson D. et al.Properties of targeted preamplification in DNA and cDNA quantification.Expert Rev. Mol. Diagn. 2015; 15: 1085-1100Crossref PubMed Scopus (32) Google Scholar, 40Okino S.T. et al.Evaluation of bias associated with high-multiplex, target-specific pre-amplification.Biomol. Detect. Quantif. 2016; 6: 13-21Crossref PubMed Scopus (22) Google Scholar]. Furthermore, qPCR offers flexibility in running small projects with hand pipetting, and large-scale projects requiring high-throughput sample processing using robotics liquid handlers and automated plate loading and reading. The data analysis and associated calculations for relative quantification involve multiple steps (Figure 5): (i) extract the Cq values for each sample after inter-run calibration and calculate the mean Cq from the associated technical replicates; (ii) determine the relative quantity by raising one + the PCR efficiency (E) [determined from the standard curve (Figure 4C)] to the ΔCq [i.e., (1 + E)ΔCq]; (calculated by subtracting the mean Cq of all the samples within the control group from the mean Cq of the technical replicates from each sample); (iii) enumerate the normalized relative expression (equivalent to ΔΔCq after log transformation) per sample by dividing the relative quantity of a given target/sample (step 2) by the geometric mean of the relative quantities of two or more reference targets; (iv) compute the average normalized expression of the samples in each biological group; and (v) perform statistical analysis based on the log transformed normalized expression per sample (Figure 5: column 9) [41Kitchen R.R. et al.Statistical aspects of quantitative real-time PCR experiment design.Methods. 2010; 50: 231-236Crossref PubMed Scopus (44) Google Scholar].These calculations are complex because qPCR data are relative by nature between samples or groups of samples, with normalization to reference genes, and calibration between plates. Additionally, qPCR measurements are made on the log scale (Cq value) with statistical analysis performed in Cq space (i.e., ΔΔCq values or using log-transformed relative normalized expression), while expression levels are reported after linear transformation of the ΔΔCq results. The methodology by which normalized expression calculations are performed, coupled with the amalgamation of data from multiple experiments that typically span multiple plates, can be daunting for any researcher. Many laboratories create Microsoft Excel spreadsheets with premade formulas to automate the analysis process. This requires copy/paste manipulations from the raw data files, necessitating the careful triage of results to fit the confines of the formulas and data flow. Multi-plate experiments can introduce another challenge in the manual fashion by which the Cq values from each experiment are entered and potentially reconfigured and rearranged in Excel to accommodate different plate configurations between individual experiments. In the hands of even the most experienced and methodical scientist, the physical management of large data sets in Excel can result in calculation errors that may go unnoticed, resulting in erroneous summarized results and conclusions [42Panko R.R. What we know about spreadsheet errors.J. Organ. End User Comput. 1998; 10: 15-21Crossref Scopus (245) Google Scholar]. The common sources of error in a given experiment cannot be entirely avoided, but by clearly understanding where and how they can arise, measures can be taken for their minimization. Given that scientists continue to push the limits of sensitivity for detection of biomolecules in smaller samples, more care must be taken to reduce the sources of error, improve precision, and achieve reproducible, statistically significant data particularly with low target abundance. In order to achieve reproducible data that truly reflect the experimental parameters, a sequence of key steps must be strictly followed to circumvent and minimize both random and systemic sources of error (Table 1), one of the most important of which is primer validation to assess annealing temperature, melt curve, and amplicon size (Figure 4A–B) [43Opel K.L. et al.A study of PCR inhibition mechanisms using real time PCR.J. Forensic Sci. 2010; 55: 25-33Crossref PubMed Scopus (357) Google Scholar]. Subsequent standard curve analysis is critical, because only when reaction efficiency is close to 100% are the Cq values representative of the target concentration in each sample (Figure 4C) [44Svec D. et al.How good is a PCR efficiency estimate: recommendations for precise and robust qPCR efficiency assessments.Biomol. Detect. Quantif. 2015; 3: 9-16Crossref PubMed Scopus (280) Google Scholar]. Furthermore, accurate normalization of qPCR data from the geometric mean stability of multiple reference genes is important because an unstable target can produce artefactual data that do not represent the true expression differences between samples (Figure 4D) [45Robledo D. et al.Analysis of qPCR reference gene stability determination methods and a practical approach for efficiency calculation on a turbot (Scophthalmus maximus) gonad dataset.BMC Genomics. 2014; 15: 648-663Crossref PubMed Scopus (104) Google Scholar, 46Vandesompele J. et al.Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes.Genome Biol. 2002; 3 (research0034)Crossref PubMed Google Scholar]. For projects that are not conducive to the selection of endogenous reference genes (i.e., developmental studies), an exogenous internal positive control may be required [47Kavlick M.F. Development of a universal internal positive control.Biotechniques. 2018; 65: 275-280Crossref PubMed Scopus (5) Google Scholar]. Finally, to avoid the ramifications of interplate variability, it is best to pipette all samples and associated replicates for a given target on a single plate (Figure 3A).Table 1Stepwise Approach to Performing a Valid qPCR ExperimentaAdapted and updated under a Creative Commons license from [8].StepDescriptionSub stepsInstruction1Experimental designProcedureList the best targets, samples, and treatments based on previously acquired, vetted, and validated data or literature (Figure 1A). align="center"Biological groupsDefine the appropriate biological groups (i.e., treatments, knockout, time points, etc.) and various combinations thereof (Figure 1).ReplicatesBiological (samples): number of animals or number of cell culture plates per biological group, which is determined with the aid of a biostatistician and/or a statistical power analysis (highly dependent on the complexity of the organism). Technical: number of wells pipetted per cDNA sample from each biological replicate (typically two or three); (Figure 1).Experimental conditionsCarefully note all controllable factors, such as: lot consistency in cell culture media, FBS, BSA, animal feed and drugs or compounds, sex, and phenotype. Take pictures throughout the experiment and carefully note any unusual changes in specific samples or specimens that may become outliers.2Tissue and cell harvestingSample extraction (complex organisms)Sacrifice animals, extract and dissect tiss