Real time and Quantitative (RTAQ) PCR. so I have an outlier and I want to see if it really is changed

Full text

(1)

Real time and Quantitative (RTAQ) PCR

or….. for this audience…

“ so I have an outlier and I want to see if it really is changed”

Nigel Walker, Ph.D.

Laboratory of Computational Biology and Risk Analysis, Environmental Toxicology Program, NIEHS

(2)

I’ve got 30 minutes- what am I going to say?

z

What I am going to tell you

Ñ Overview of the technology

Ñ Overview of Software based quantitation Ñ Assay Designs

Ñ Types of Quantitative analysis

z

What I am not going to tell you

Ñ How to design primers

Ñ How to use different types of PCR machines Ñ All there is to know- this is just a brief intro

(3)

Conventional RT-PCR

z

Reverse transcription (RT) of cDNA from RNA

Ñ Oligo d(T)

Ñ Random Hexamer Ñ Gene specific primer

z

PCR amplification of a defined DNA sequence from cDNA

Ñ Traditionally a 3-phase multi-cycle reaction Ñ Denaturation, annealing, primer extension

z

Electrophoretic separation of PCR products (amplicons)

Ñ PCR for specific number of cycles Ñ Run products on an agarose gel

(4)

Real-time PCR is kinetic

z

Detection of “amplification-associated fluorescence” at each cycle during PCR

z

No gel-based analysis at the end of the PCR reaction

z

Computer based analysis of the cycle-fluorescence time course

Increasing fluorescence

Linear plot

PCR cycle

(5)

Specialized Instrumentation is needed

z

96-well format

Ñ ABI SDS 7700 Ñ ABI 7000

Ñ ABI 5700

Ñ BioRad Icycler

Ñ Stratagene Mx4000

z

Capillary tube format-

Ñ Roche Light cycler

z

384-well format HT systems

Ñ ABI SDS 7900

(6)

Software-based analysis

z

Data acquisition

Ñ Fluorescence in each well at all cycles.

Ñ Software-based curve fit of fluorescence vs cycle #

z

Threshold

Ñ Fluorescence level that is significantly greater than the baseline.

Ñ Automatically determined/User controlled

z

C

T

(Cycle threshold)

Ñ Cycle at which fluorescence for a given sample reaches the threshold.

Ñ CT correlates, inversely, with the starting concentration of the target.

Ñ Varies with threshold- not transferable across different plates

(7)

Software-based analysis

Baseline

C T (cycle threshold)

Increasing fluorescence

Threshold (10sd)

Log plot

(8)

Example analysis of CYP1A1

•SYBR Green detection

•10-fold dilution series

No RNA controls C

T

values

Threshold

(9)

Linear range for CYP1A1 by RTAQ-PCR

1pg-100ng Total RNA

95% E

z CT correlates, inversely, with the starting concentration of the target.

(10)

Amplification-associated fluorescence

z

Fluorescent dye

Ñ Detects accumulation of DNA (SYBR green)

z

FRET (Fluorescent Resonance Energy Transfer) based.

Ñ Detect accumulation of a fluorescent molecule (Taqman)

Ñ Detect accumulation of specific DNA product-(Molecular Beacons, LC)

(11)

Methods of fluorescence detection

Light Cycler Molecular

Beacons Taqman

SYBR Green

R

Taq

Q

R Q

“Direct-NS” “Indirect” “Direct-S”

D A

“Direct-S”

(12)

Dye-based FRET-based

z Upside Ñ Quick

Ñ No primer/probe optimization

z Downside

Ñ Non-specific

z Application

Ñ Many genes few samples

Ñ “That sounds like just the ticket for checking my microarray data”

z Upside

Ñ Specific

z Downside

Ñ Primer/probe optimization Ñ More costly

z Application

Ñ Many samples few genes

(13)

Primer sets

z There are no large databases of primer sets

Ñ We attempted to get a database going early on with little success.

z Commercial pre-developed assays are available

z Dye-based

Ñ Existing primers could be adapted but may not be the best

z Probe-based

Ñ Unlikely that existing PCR primer pairs will be suitable

Ñ Primers and probes designed to match specific reaction conditions.

z Primer design

Ñ Primer Express™ software facilitates design (for ABI system) (SCL copies) Ñ No optimization of primer annealing temperature

Ñ Multiple primer sets can use same default cycling conditions on SDS7700

(14)

Quantitation

z

Absolute quantitation (eg. copies/ug RNA)

Ñ Interpolate CT vs standard curve of known copies of nucleic acid Ñ Total RNA, in vitro transcribed RNA, DNA

z

Unit-less quantitation (arbitrary values/ug RNA)

Ñ Interpolate CT vs dilution curve of a “quantitator standard RNA”

(15)

Relative quantitation

z

∆C

T

between “control” and “treated” RNAs on a single plate

Ñ Fold-difference

Ñ Cannot compare Ct between samples on different plates

z

∆C

T

between “calibrator” RNA sample and unknown RNA

Ñ Same calibrator RNA can be on multiple plates

z

∆∆C

T

between “control” and “treated”

Ñ Fold change-normalized to a separate reference gene/sample

(16)

Relative fold change

z

C

T

inversely correlated with starting copies

z

Each cycle there is a “doubling” of amplicons (assuming 100%

efficiency)

z

Difference in 1 cycle therefore a 2=fold difference in copies Fold change = 2

∆ CT

∆Ct = 3.31

Fold difference in starting copy number = 2

3.31

= 9.9

(17)

Correlation of real-time and microarray

R2 = 0.6888

0.1 1 10 100

0.1 1 10 100

Microarray fold

Data from Martinez et al (submitted), SYBR, 2∆Ct

(18)

Efficiency adjustment

For a ∆C

t

=1, Fold change = efficiency

z

2

∆ CT

assumes a 100% efficient amplification

z

For single gene efficiency adjustment use

Ñ Fold change = e ∆Ct

Ñ Where efficiency(e%) = 10 (-1/slope)

z

For a ∆Ct=1, Fold change = efficiency

(19)

Calculation of Efficiency

z

Based on a linear plot of C

T

vs. log copies:

z

Efficiency(e%) = 10

(-1/slope)

z

100% efficiency (2 copies each cycle) slope of –3.3219.

Slope = -3.462

e = 10

(-1/3.462)

= 1.95 1.95 copies per cycle

∆Ct = 3.3

Fold= (1.95)

3.31

= 9.1 fold

(20)

Efficiency adjusted Normalization

z

Fold-change can be “normalized” relative to a “reference gene”

z

Reference can be a separate sample on the plate

z

Beware of the interpretation of a normalized fold change

Ñ Assumption that the reference gene is “unaffected” by treatment

(21)

Can I really detect <2X changes?

z

How can you be sure a difference is real?

Ñ T-test on triplicates CTs

z

Sensitivity of discrimination is dependent upon..

Ñ Efficiency

Ñ Assay variability

Ñ Number of Replicates

(22)

Variability impact on ∆Ct

z

Taqman or SYBR

z

Standard deviations (n = 9)

Ñ 0.2-0.5 cycles

Ñ CV, 1-2% on CT values

z

Power calculation for ∆Ct =1, 90%, p<0.05 (T-test)

Ñ sd=0.25, n = 3 Ñ sd =0.33, n = 4 Ñ sd= 0.5. n = 7

(23)

Issues of assay design

z

RNA specific sets -ie Primers spanning intron location

Ñ If you know the gene and have the time go for it.

Ñ Not all genes in database and annotated esp. rat

z

Do you need RNA specific sets?

Ñ RNA expression 103-108 copies/100ng total RNA

Ñ 100 ng RNA approx = 100 single gene copies (assuming 1% DNA contam)

z

Reverse transcription

Ñ Gene specific primer is best especially if using a synthetic RNA standard Ñ Oligo d(T)-may not be good for 5’ end targets

Ñ Random hexamers - poor for synthetic RNA standard

(24)

Some Take-home advice

z

You’re not in Kansas anymore, so do your homework first

Ñ Learn the concepts before you do the assay

z

Specialized machines

Ñ Make contact with someone who will let you use their machine

z

The devil is in the details.

Ñ You can get the same CT from very different curves of different quality

z

You still need gels

Ñ While quantitation is gel-free, a picture tells a thousand words

z

Replicate

Ñ You can detect a 2X change with duplicates but is it for real?

(25)

Acknowledgements/Information sources

z Chris Miller- now the ABI Field Application Specialist!

z NIEHS Real-time PCR webpage Ñ http://dir.niehs.nih.gov/pcr

z Applied Biosystems

Ñ http://www.appliedbiosystems.com/apps

z BioRad ICycler

Ñ http://www.bio-rad.com/iCycler/

z Stratagene

Ñ http://www.stratagene.com/q_pcr/index.htm

z Light Cycler

Ñ http://biochem.roche.com/lightcycler/lc_sys/lc_sys.htm

(26)

Additional Reference

Michael W. Pfaffl. A new mathematical model for relative quantification in real- time RT-PCR. Nucleic Acids Research 29(9): 2005-2007, 2001.

Figure

Updating...

References

Updating...

Related subjects :