From DNA Sample to Data Point Characterizing Trait Testing Methods

1
From DNA Sample to Data Point Characterizing
Trait Testing Methods
David Hondred Senior Research Scientist Pioneer
Hi-Bred International
2
Measuring the Needle in the Haystack
Ancillary Presence in Seed
Ancillary Presence in Grain
3
Testing for Adventitious PresenceCompilation of
many areas
PCR Testing
Grinding/ Homogenization
Interpretation
Sub-Sample
Sample
4
Testing for Adventitious PresenceOnly as good as
the weakest component
Failure in One Area Compromises the Entire
Analysis
Success Depends on Recognizing Weak Points and
Compensating for Them
5
Accuracy
How relevant are the test results to the Real or
True Value?
Pertains to likelihood that test results bracket
Real Value.
6
Precision
How closely do the test results cluster?
Pertains to the degree of agreement among results

7
Best Test Accurate and Precise
8
Factors Affecting Accuracy and Precision

• Precision
• Variability in Raw Data

• Accuracy
• Systematic Influence

• Common Sources
• Seed Sampling
• Flour Sub-Sampling
• DNA Sub-Sampling

• Common Sources
• Source of Standard Material
• Endogenous Gene Reference
• Specific or Generic Assay

• Impact on Interpretation

• Impact on Interpretation

• Identified as Variation in Replicated Data

• Difficult to Identify

• Single data points less informative

• DNA to Seed Conversion

9
Validated Robust AssayA Given

• Multiplexed or Singleplexed
• Optimized Primer Concentrations
• Optimized Probe Concentration
• Validated Amount of DNA per reaction
• Well Characterized Dynamic Range of Assay

10
Sources of Data VariationImpact on Precision
11
Variation between replicate PCR
Outlier?
Spike level was 0.057 2 Target Kernels per 3500
Kernels
12
Variation between like Sub-Samples
Spike level was 0.029 1 Target Kernel per 3500
Kernels
13
Variation between Replicate Pools
Spike level was 0.057 2 Target Kernels per 3500
Kernels
14
Sources of Variation
Sampling

• Random Sampling
• Representative Sample
• Homogeneous Sample
• Sample Size Corresponds to Detection Limit

15
Sources of Variation
GRINDING

• Particle Size
• Effects sub-sampling
• Homogenous mixture
• Cleanliness
• Equipment purge and decontamination
• Airborne dust

16
Sources of Variation Flour Sub-Sampling
Large Sub-Sample
Small Sub-Sample
1000 mg
100 mg

• More particles
• Better representation

• Fewer particles
• More variability

• Specialized DNA extraction
• Slower process

• Automated DNA extraction
• Higher through-put

17
Sources of Variation Flour Sub-Sampling
100 mg
1000 mg

• Coarse Grinds Aggravates the Problem
• Fewer particles per unit mass
• Fewer particles per unit volume
• Greater variability
• Small sample may result in negative result.

18
Sources of Variation DNA Sub-Sampling
23 chance any one test gives results in this
area
23 chance any one test gives results in this
area
Large Number of DNA Targets Less Assay Variation
19
Sources of Variation DNA Sub-Sampling
Small Number of DNA Targets More Assay Variation
20
Sources of Variation Cross Contamination
Negative
Positive
Biologically Relevant Level Minimum amount
corresponding to one seed
21
Sources of Variation Cross Contamination
Negative samples with No GMO in 3500 seeds
(0.000)
1 GMO in 3500 seed would equal 0.0286
22
Systematic Factors that Impact Accuracy and
Influence Data Interpretation
23
Variation Factor Seed-Based
Pericarp (ALL maternal genetic material)
3N
Starchy and Vitreous Endosperm (21
maternalpaternal genetic material) gt80 of the
Kernel 50 of DNA
2N
Embryo (11 maternalpaternal genetic
material) lt20 of the Kernel 50 of DNA
24
Percent DNA to Percent Seed Hemizygous Male
Seed as Standard
17 over estimate
40 over estimate
25
Percent DNA to Percent Seed Homozygous Seed as
Standard
24 over estimate
13 under estimate
26
Variation Factor Impact of Copy Number
Hypothetical Test Result
35S Promoter Test Performed on 2000 seed pool
Standard Curve produced with Mon810
Results indicate level of 35S Promoter DNA in
unknown sample is at 0.1
From the Literature
Mon810 contains 1 copy of 35S promoter per
haploid genome
Bt11 contains 2 copies of 35S promoter per
haploid genome
Question How many GMO seeds are present
in the 1000 seed pool?
Results would be compatible with either

• 2 seeds of Mon810

• 1 seed of Bt11

27
Percent DNA to Percent Seed Multicopy Event
Equal number of GMO Seeds
3x over- estimation
28
Variation Factor Impact of Inbred on
Endogenous Gene Reference
Example of unacceptable endogenous gene reference
test
Evidence for Inbred Attenuation (SNP Effect)
29
Testing for Adventitious PresenceHigh
Quality/Reliable Data Attainable
-6
(0.50)
-7
(0.25)
-8
(0.125)
delta Ct
-9
(0.10)
-10
(0.05)
-11
(0.029)
-12
-3.6
-3.4
-3.2
-3.0
-2.8
-2.6
-2.4
-2.2
log ( AP)
30
Testing for Adventitious PresenceHigh
Quality/Reliable Data Attainable
1 Target seed in 3500 seeds (0.029)
Each pool sub-sampled 5x with PCR
performed in triplicate
31
Testing for Adventitious PresenceHigh
Quality/Reliable Data Attainable
Percent Target DNA
Percent Target Seed
But it takes work!