Importance ofExperimental Design

in QPCR

Review of good practice in QPCR for meaningful and robust results

Experimental Design

March, 2009Page 1

Dr. Steffen Müller Sr. Field Application Scientist

� Introduction – Source of Variance

� Sampling and Sample Preparation

Importance of Experimental Designin QPCR

Experimental Design

March, 2009Page 2

� Quality of Template

� Inhibition

Experimental Design

Sample preparationand purification

Experimental DesignExperimental DesignUnderstanding experimental varianceUnderstanding experimental variance

Experimental Design

March, 2009Page 3

Experimental Design

cDNA

Total RNA

AAAA

AAAAAAAA

Reverse TranscriptionReal time QPCRPost-run Analysis

Sources of variance in QPCR experiments:

� Biology of experimental systemUse of biological replicates

� Technical varianceUse of technical replicates

Experimental DesignExperimental DesignUnderstanding experimental varianceUnderstanding experimental variance

Experimental Design

March, 2009Page 4

� Pipetting errorAvoid pipetting small volumes and use of calibrated pipettes

� Varying template quality and/or quantityUse of sample QC and proper normalization

� Run-to-run variabiltyUse of inter-run calibrating samples

Number of biological replicates should be determined in a pilot experiment.Usually 3 technical replicates are sufficient.

Sampling andDNA/RNA extraction

Nucleic acid quantificationand QC

Experimental DesignExperimental DesignUnderstanding experimental varianceUnderstanding experimental variance

� Sample complexity� Efficiency of extraction procedures� Quality of template preparation (Inhibitors, RNA

or DNA quality)

� There is no gold standard method� Different methods will generate different results� Any type of QPCR quantification assumes

similar template qualities

Experimental Design

March, 2009Page 5

Reverse Transcription

QPCR

similar template qualities

� Known to be the major source of variability� RT efficiency is dependent on initial RNA

quantity as well as individual level of acertain mRNA

� Type of enzyme and priming method willyield different results

� Assay robustness affects results� Variance of chosen normalizer� Plasticware and reagents

Quantity?

Efficiency of sample andstandard amplification

Linear dynamic rangeof assay

Experimental DesignExperimental DesignApplication specific varianceApplication specific variance

Differences in sample andstandard amplificationefficiency lead to over- orunderestimation of samplequantity

Quantification should beperformed in the rangewhere there is a lineardependency of Ct and

Experimental Design

March, 2009Page 6

Fold Change?

Quality of samplesVariance of chosennormalizer

quantitydependency of Ct andlog(quantity)

Low quality can have anegative effect on resultsIf not taken care of

Reference genes should have similarexpression levels as the GOI due to RTproperties

Experimental DesignExperimental DesignApplication specific varianceApplication specific variance

What are the limitations of comparative quantification? Calibrator

Unknown 1Unknown 2

� Needs to be positive for GOI and reference� Consider exogenous Calibrators (eg. QPCR reference RNA)

� Should have a Ct value in the range of the unkowns(max 10-12 Cts difference to unknowns � ~1000-5000 fold change)

Calibrator

Experimental Design

March, 2009Page 7

Unknown 1Unknown 2 Calibrator

Ct

The further away an unknown is from thecalibrator the less accurate the FC will be

Sample in the center of the data populationbest possible Calibrator (reduced variance)

Experimental DesignExperimental DesignApplication specific varianceApplication specific variance

Melt curve analysis is influenced by many variables :

� Salt concentration of the reaction mix(Spink and Chaires, Biochemistry 38, 1999; Zipper et al., NAR 32, 2004)

� Other compounds present in the reaction mix(Spink and Chaires, Biochemistry 38, 1999)

� Concentration and type of double strand binding dye

Experimental Design

March, 2009Page 8

� Concentration and type of double strand binding dye(Gudnason et al., NAR 35, 2007; Rasmussen et al., BMC Bioinformatics 8, 2007)

� Ramp rate of the thermal system(Gundry et al., Clin. Chem. 49, 2003)

� Data sampling rate and algorithms used in data analysis

� Sequence of the DNA(Li et al., BioTechniques 35, 2003; Giglio et al., NAR 31, 2003; Zipper et al., NAR 32, 2004;Rasmussen et al., BMC Bioinformatics 8, 2007)

Sample PreparationSample PreparationInfluence of Sampling and Sample PreparationInfluence of Sampling and Sample Preparation

Sample preparation influences QPCR results

Quality of template Amount of InhibitorsAmount of co-purified

salts

� Quantification assumes � Inhibitors can lead to � Affects primer and probe

Experimental Design

March, 2009Page 9

� Quantification assumescomparable quality

� Low quality can leadto failure of detection

� Inhibitors can lead todelayed or failure ofdetection

� Affects primer and probebinding affinity

Sample preparation affects QPCR assay performanceresulting in lower assay sensitivity if not optimize d!

Why Quality MattersWhy Quality Matters

Sample/templatequality

RNA quality control (Quality of template):

� RNA degrades naturally due to enzymatic or autocatalyticmechanisms: � Any 5’ or 3’ biased design might fail

or produce misleading results� Wrong priming strategy in the RT step

can produce misleading results

� Knowing RNA quality allows to accommodate theamplicon design and set expectations avoiding wronginterpretation of results

Experimental Design

March, 2009Page 10

interpretation of results

� All quantifications rely on comparable template quality tobe meaningful

Why Quality MattersWhy Quality Matters

Sample/templatequality

QPCR assay validation/optimization (Quality of resul ts):

Experimental Design

March, 2009Page 11

Quality of assay

Robust and meaningfulresults

� The resolution of SYBR Green meltcurves is limited

� Tm depends on dye/template ratio

� Verifying the size of PCR products is a recommendedvalidation procedure: � Resolution of slab gels limited!

Experimental workflowExperimental workflow

RNA extraction

Nucleic acid quantificationand QC

Extraction from 5x106

HEK cells usingAbsolutely RNA ® mini

RT from 1 µg of total

Quantification of 1 µlsample on Nanodrop

QC on Bioanalyzer :RNA 6000 nano kit

RNA degradation @ 70°C

Experimental Design

March, 2009Page 12

Reverse Transcription

QPCR Assay validationQPCR 5‘ and 3‘ assays

RT from 1 µg of totalRNA usingAffinityScript ™

Analysis of QPCR products onBioanalyzer: DNA 1000 kit

Mx3005P®

Brilliant ® II SYBR® Green

RIN 8.90 min

A RIN 6.530 min

B

RIN 4.645 min

C RIN 2.375 min

D

Effect of RNA quality ongene expression results:

RNA was extracted from HEK293cells and thermally degraded

All RNAs were tested on theAgilent Bioanalyzer

Quality and Impact onQuality and Impact onGene Expression ResultsGene Expression Results

Experimental Design

March, 2009Page 13

intron 2-3: 23.6 kb

GAPDH

HPRT1

YWHAZ-14

-12

-10

-8

-6

-4

-2

0

RIN 8.9 RIN 6.5 RIN 4.6 RIN 2.3

GAPDH 5' assay

GAPDH 3' assay

GAPDH

-10

-9

-8

-7

-6

-5

-4

-3

-2

-1

0

RIN 8.9 RIN 6.5 RIN 4.6 RIN 2.3

HPRT1 5' assay

HPRT1 3' assay

HPRT1

YWHAZ

-6

-5

-4

-3

-2

-1

0

RIN 8.9 RIN 6.5 RIN 4.6 RIN 2.3

YWHAZ 5' assay

YWHAZ 3' assay

Agilent Bioanalyzer

Results:Assay design:

Why DNA Quality mattersWhy DNA Quality matters

Assay

Small DNA fragmentscompete with primers:Unspecific amplification,

Reduced population of DNAwith full length of amplicon:Underestimation of quantity

DNA degradation in preserved biological tissue, forensic samples orsamples commonly used in pathogen detection can negatively impactassay performance and produce misleading results

Experimental Design

March, 2009Page 14

Assayperformance

andsuccess

Unspecific amplification,Underestimation of quantity

Alterations of bases:Reduced affinity of primersand probes

Competition by abortiveamplicons:Loss of sensitivity or inhibition

Flu

ores

cenc

e

Threshold

Sample Quality Sample Quality -- InhibitionInhibition

� Various compounds can act asinhibitors: eg. phenols, polysaccharides,cell debris, EDTA, lipids, high amountsof rRNA + tRNA or gDNA, componentsof RT reactions,

� Inhibition affects the kinetics of thePCR reaction resulting in later Ct valuesand reduced amplification efficiency

Experimental Design

March, 2009Page 15

Cycle#

CtnormCtinhibHEX labeled probe against ALIEN®

FAM labeled probe against β2m� Inhibited samples either need to be excluded

from quantification or if possible rerun at higherdilutions

� Internal positive controls (eg. ALIEN® Inhibitor Alert)are a good tool to monitor the varying degreesof inhibition in each sample

DNA/RNA extraction

Nucleic acid quantificationand QC

Optimization of sample extraction for yield, quality and low amounts of inhibitors

Choosing the right

RealReal--time PCRtime PCRAssay OptimizationAssay Optimization

Experimental Design

March, 2009Page 16

Reverse Transcription

QPCR assay validationand optimization

QPCR

Choosing the rightenzyme and primingmethod to increaseyield

QPCR uses smallamplicons. Shortenthermoprofile timings for increased speed

Optimize primer and probe concentrationfor increased specificity and efficiency

Summary� It is important to understand sources of experimental variance

� If variability exists try to minimize this by adjusting your experimental design

� All types of quantification assume similar template quality

� Assay validation and optimization are not optional but crucial to obtainrobust and meaningful results

Experimental Design

March, 2009Page 17

� Make sure the analysis parameters are meaningful and make sense inrelation to the data obtained in the experiment

� With the drive to detect ever smaller quantities or fold changes it is highlyimportant to address experimental variability at all stages

� Finally we have to accept that in the worst case the overall variabilitymay prevent us from achieving a certain sensitivity of our results

Thanks for your attention!Thanks for your attention!Thanks for your attention!

Importance of Experimental Designin QPCR

Experimental Design

March, 2009Page 18