Uncertainties and certainties in GMO analytics using qpcr

Transcription

1 Uncertainties and certainties in GMO analytics using qpcr PD Dr. Philipp Hübner Kantonales Labor Basel-Stadt Abteilung Lebensmittel Postfach CH-412 Basel

2 using qpcr historical review Short crash course in validation Accuracy Limits of detection and quantitation Error calculation Identity Tag for Quantitation Conclusion

3 Limit of detection of 35S-PCR GMO labeling depended on PCR testing with reference method targeting 35S promotor P35S factor 2 (=2%) 2 ng 1 ng.5 ng.2 ng (5) (1) (2) (3) (8) (7) (4) 95% Upper Mean 95% Lower.1 ng (1) (6) (9)

4 Implementation of Quantitative Competitive PCR Co-amplification of 5 ng target DNA with a defined amount of standard DNA (equivalent to 1 % GMO) external standards samples controls L #1 #2 #3 #4 #5 S R N #1: Lecithine 1% #2: Flour <.1% #3: Protein 2% #4: Grist > 5% #5: Grist 1-2% S: Standard DNA R: Roundup Ready soybean DNA N: PCR negative control (H 2 O)

5 QC-PCR for the Determination of the GMO content Example: Sample #4 grist Neg. % 1 % 2 % 5 % 1 % 2 % 5 % standard target titration with internal competitor DNA relative intensities migration (cm) migration (cm) migration (cm) migration (cm) migration (cm) 24 log (standard/target) y = x, r 2 =.998 for y =, x = x = = 1.5 Sample #4 contains 1% GMO log [standard]

6 Results of inter-laboratory study participants, mainly Swiss labs 25 2 sample #6 flour LMC137 QC-PCR GM (LMC): <.1% 35S-PCR (RAP;LMC): neg RRS-PCR (RAP): neg RRS-PCR (LMC): 2pos/3neg sample #7 flour LMC17 QC-PCR GM (LMC): 6% 35S-PCR (RAP): pos RRS-PCR (RAP;LMC): pos GM RRS P35S GM RRS P35S GMO content I x <.5% II* IV.5% < x < 2% x > 2% Categories depended on available reference material -> semiquantitative results

7 using qpcr historical review Short crash course in validation Accuracy Limits of detection and quantitation Error calculation Identity Tag for Quantitation Conclusion

8 Which validation parameter are important? we would like to measure accurately: trueness and precision we need a robust method which gives comparable results in other laboratories

9 Accuracy: trueness and precision high trueness high precision high trueness low precision low trueness high precision low trueness low precision

10 what is needed for validation? for determination of relative trueness: certified reference material* samples material which can be traced back to CRMs interlaboratory studies or proficiency testing schemes (e.g. FAPAS) * commercial availability is restricted to view events

11 what is needed for validation? (2) for determination of precision: replicates of sample extraction and rt-pcr are informative for the precision of the whole procedure whereas replicates of extracted DNA are only informative for the rt-pcr part of the procedure reproducibility can be assessed by replicate measurements in different labs

12 Accuracy: Trueness and precision Table 5 Quantitative estimates with DNA extracted from mixed and processed foodstuffs. Each foodstuff was analysed four times (A D) Food sample RRS A B C D Mean SD RSD r Bias % (%) Biscuit Biscuit Acidified soybeans Infant formula RSD r varied from 7% to 34% Bias varied from 1% to 29% Berdal and Holst-Jensen (21)

13 Accuracy (trueness and precision) should be determined at the legal limit (i.e..9 % GMO) RSDs in the range of 25% to 35% have been achieved in ring trials FAPAS indicates a target standard deviation of.2 log 1. This means that the lower 95% confidence intervall is at 4% of the mean, the upper limit at 25% of the mean (zscore:=2) Example mean:= 1%; 95% confidence intervall: 4% - 25%.2 log 1 is equivalent to.66 log 2. This means C t - values up to.33 are acceptable by FAPAS

14 Limit of Detection (LOD) Limit of Quantification (LOQ) Definitions (IUPAC) The limit of quantification of an analytical procedure is the lowest amount or concentration of analyte in a sample which can be quantitatively determined with an acceptable level of precision and accuracy the limit of detection is the smallest amount or concentration of analyte in the test sample, that can be reliably distinguished with stated significance, from the background or blank level.

15 Pipetting of target molecules mean StDev RSD (%) µ 1 µ Codex alimentarius: LOQ is at RSD = 25% LOD is at RSD = 33%

16 Theoretical uncertainty of measurements using quantitative PCR Jean Peccoud and Christine Jacob σ =.19 σ =.95

17 realtime PCR: estimation of LOQ and LOD α=.95 pipetting PCR efficiency CRM mean x>µ RSD RSD RSD RSD µ % m=1.9 overall LOQ LOD Codex alimentarius: LOQ is at RSD = 25% LOD is at RSD = 33%

18 Influence of plant genome size on LOQ common name scientific name genome size a genome copies 1% LOQ b [in Mia bp] [per 2 ng] corn Zea mays % rice Oryza sativa % soybean Glycine max % wheat Triticum aestivum % a published genome sizes (per 2C) were taken from Arumuganathan et al. b the theoretical limit of quantitation is expressed as the fraction (in %) of 36 copies divided by the number of copies of the corresponding plant species within 2 ng DNA

19 The LOQ depends on the amount of the ingredient (matrix match) 1 absolute LOQ relative LOQ 4% 1.4% %.4% 36 genome copies soya ingredient (in %) LOQ (in % GMO) copies

20 Limit of Detection (LOD) Limit of Quantification (LOQ) the sensitivity of the whole procedure can only be determined by analysing CRMs with low GMO contents in the range of.1% to.1%; at present CRMs with.1% GMO can be tested, all published methods reach this sensitivity the sensitivity of the rt-pcr part of the procedure can be assessed by measuring replicates of diluted GMO-DNA test samples. Attention: this sensitivity does not reflect the sensitivity of the whole procedure! LOQ: 3 to 5 gene copies LOD: 1 to 2 gene copies

21 using qpcr historical review Short crash course in validation Accuracy Limits of detection and quantitation Error calculation General Quantitation Marker Conclusion

22 amount DNA fluorescence realtime PCR in theory N = N ( + m) n 1 n mn ( ) 1 mn ( ) = m(n) 1 Plateauphase exponential phase Threshold C t time number of cycles C t

23 error calculus of realtime PCR Which parameter has the greatest influence on the measured Ct-value (n)? PCR-efficiency m ("Biochemistry") pipetting error N error of fluorescence measurement N n law of error propagation by Gauss: f ( x) f ( x) = x x

24 (1) PCR-efficiency m ( biochemistry") n = ln( N / N ) n ln( 1 + m) n f ( m) = m m n = n 1 ( 1+ m) ln( 1+ m) m n ( C t ) is proportional to m and to the number of cycles n and inverse proportional to the mean PCR efficiency (1+m)

25 (2) pipetting error N n = ln( N / N ) n ln( 1 + m) n f ( N ) = N N n = 1 ln( 1+ m) N N n ( C t ) is proportional to the relative pipetting error N /N

26 (3) error of fluorescence determination N n n = ln( N / N ) n ln( 1 + m) n f ( N n ) = N n N n n = 1 ln( 1+ m) N N n n n ( C t ) is proportional to the relative error of the fluorescence measurement N n /N n

27 what are the contributions of the different terms? N N N N n n relative pipetting error: 2.5% relative error of fluorescence measurement: 1 to 1% (depends on used apparatus) n exp(specification) :.16 to.387 n 1 n 2 n 3.1 to to.14 1 PCR efficiency 2 pipetting 3 Fluorometer n 1 n 3 > n 2 The precision of the measurement is mainly determined by the factors biochemistry and fluorescence measurement

28 Influence of Taq DNA-Polymerases on Ct D (n:=8) D (Enhancer) D (Wiederholung) D using 1% Soja (ca. 4 copies) enzyme D yielded lowest C t -values C (n:=8).39 C (Kit, Wiedrholung) C (Kit) C ±.1.93 to 1.7 RSD = 7% B A (1u/PCR; BSA) A (2u/PCR).19 RSD =.1/3 =.3% A (1u/PCR) Ct specification

29 using qpcr historical review Short crash course in validation Accuracy Limits of detection and quantitation Error calculation Identity Tag for Quantitation Conclusion

30 Unique identifiers for authorized GMOs Commission Regulation (EC) No 65/24 Applicants shall, in accordance with the formats set out in the Annex, develop the unique identifier for each GMO concerned,... Standardised code containing 9 alphanumerical signs 3 components A. Identification of applicant (2-3 signs) B. Definition of event (5-6 signs) C. Control of code (1 sign) Example: Roundup Ready Soja MON A B C

31 Detection of Saccharomyces cerevisiae carrying ID-tags transformation of yeast realtime PCR Isolation of DNA cultivation of GM-yeast baking bread

32 using qpcr historical review Short crash course in validation Accuracy Limits of detection and quantitation Error calculation Identity Tag for Quantitation Conclusion

33 ingredient (%) composed food with several ingredients protein Limitation: flour protein flour GMO soya flour soya flour GMO soya protein soya protein wheat flour GMO analysis per ingredient product A product B product C possible possible not possible

34 Limits of GMO analytics Availability of reference materials (quantitative and qualitative) sample matched LOQ (practical LOQ) Products with several ingredients derived from corn or soya Purified ( DNA-free ) ingredients or products

35 GMO-ID-Tag fast and reliable screening, identification and quantitation of GMOs no cross reactivity Analysis without reference material gets possible amplifiable yeast-dna could be extracted from bread

36 GMO-Analyses at KL BS number of samples 1 % 8% 6% 4% 2% % n = n = n = n = n = n = 91 po s itive (>1%) po s itive (.1%<x<1%) po s itive (<.1%) negative n = 54