Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling...
Transcript of Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling...
![Page 1: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/1.jpg)
faculty of science and engineering analytical biochemistry
Biomarker Best Practices: confounding factors, preanalytics and method validation
Rainer [email protected]
EU-COST - Basel, March 28, 2019
![Page 2: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/2.jpg)
Causes of irreproducibility
Freedman et al., PLoS Biol 2015, 13, e1002165.
![Page 3: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/3.jpg)
Research antibodies• Standards to validate antibodies used in basic or clinical research, and in
bio-pharma• Standards to ensure reproducibility of antibodies over time between
manufacturers and lots• Best practices, such as correct use of negative and positive controls in
antibody-based tests
Bradbury and Plückthun, Nature 2015, 518, 27-29.
![Page 4: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/4.jpg)
Biological specimen
Pre-analytical variables affect outcomes unrelated to biology:Surgical procedures, storage, freeze-thaw cycles, analysis, etc. Gene expression, protein levels, post-translational protein modification, etc.Ideally, we would:Carefully follow a standard operating procedure (SOP)Link all specimens to the SOP used to collect them
Malm et al., J Proteomics 2013, 95, 38-45.
![Page 5: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/5.jpg)
Example CSF: preanalytical factors
Human: 3-5 mL
![Page 6: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/6.jpg)
Delayed storage after samplingCSF was left at room temperature directly after sampling for 0, 30, 120 min
Freeze/thaw cyclesSnap freeze in liq. N2 and thaw on ice 1-10 times
Autosampler stabilityLeave in autosampler for extended time periods
Sample handling & stability
![Page 7: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/7.jpg)
Delayed storage after sampling
Rosenling et al., J. Proteome Res., 8: 5511-5522, 2009.
![Page 8: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/8.jpg)
Freeze-thaw cycles
Transthyretin
Rosenling et al., J. Proteome Res., 8: 5511-5522, 2009.
![Page 9: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/9.jpg)
Storage in autosampler at 4oC
Rosenling et al., J. Proteome Res., 8: 5511-5522, 2009.
![Page 10: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/10.jpg)
Conditions of blood coagulationTh
rom
bosp
ondi
n-1
Klont et al., Methods Mol Biol 2019, 1959, 1-22.
![Page 11: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/11.jpg)
FACTOR LEVELSTube Type BD368430 BD367784Heamolysis Level Low HighClotting Time 2 hours 6 hoursFreeze-Thaw cycles 1 3Trypsin Concentration 1:100 1:20Deactivate Trypsin No YesSample Stability 0 days 30 days
Sample preparation taken accountfactors and levels
rt
Inte
sity
LC-MS analysis
rtm/z Sample I
Final Common Peak List Format
peak id Sample II Sample n
p p g
-10 -8 -6 -4 -2 0 2 4 6 80
2
4
6
8
10
12
14Peak selection with -log10(p-value): 1.301 and log2(fold ratio): -1 and 1
log2(fold change)
-log 10
(p-v
alue
)
blood collection tubehaemolysisclotting timefreeze-thaw cycletrypsin digestionstopping trypsinsample stability
p f g
avgX X X X Eα β αβ= + + + + +X
How to assess the relevance of preanalytical factors?
Mitra et al., Analytical Chemistry 2016, 88, 4229-4238; Klont et al., Methods Mol Biol 2019, 1959, 1-22.
![Page 12: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/12.jpg)
Effect of data (pre)processing
Rosenling et al., J. Proteome Res., 11, 2048-2060 (2012)
![Page 13: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/13.jpg)
From raw data to results
Suits et al., Anal. Chem., 80: 3095-3104 (2008)Christin et al., Anal. Chem., 80: 7012-7021 (2008)Suits et al., Anal. Chem., 83: 7786-7794 (2011)Hoekman et al., Mol. Cell. Proteomics, 11 (6) (2012)Christin et al., Mol. Cell. Proteomics, 12, 263-276 (2013)
![Page 14: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/14.jpg)
Same data – different results
Hoekman et al., Mol. Cell. Proteomics, 2012: 11 (6), 10.1074/mcp.M111.015974
![Page 15: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/15.jpg)
A scoring system to assess the quality of data (pre)processing
Hoekman et al., Mol. Cell. Proteomics, 2012: 11 (6), 10.1074/mcp.M111.015974
![Page 16: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/16.jpg)
72.0 72.5 73.0 73.5 74.0 74.5 75.0 time [min]0
1
2
3
4
x106Intens.
Quantification
b) c)MediumHigh Low
e)
a)
c)
b) d)
ManualMzmine
Peak splitting: one possible source of errors
Klont et al., Methods Mol Biol 2019, 1959, 1-22.
![Page 17: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/17.jpg)
Christin et al., Mol. Cell. Proteomics, 2013: 12 (1), 263-276
Same processed data – different results
![Page 18: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/18.jpg)
Method validation: Quantification of IGF-1 by MALDI-MS
Klont et al., Analytical Chemistry 2017, 89, 6188-6195.
![Page 19: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/19.jpg)
› 15N-IGF1 as internal standard reliable compensation for sources of variability
› Rat plasma as calibration matrix adequate mimicking of human plasma
› SDS-based IGF1/IGFBP-complex dissociation proper liberation of IGF1 from its binding proteins
› Workflow automation matching MALDI-TOF’s high throughput capabilities
› Identification of quality indicators improving MALDI-TOF performance & data quality
Method development
![Page 20: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/20.jpg)
Relevance of the internal standard
Klont et al., Analytical Chemistry 2017, 89, 6188-6195.
![Page 21: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/21.jpg)
Summary of validation data
QC-low QC-medium QC-highCV Bias CV Bias CV Bias
Accuracy & precision (3 runs in 6-fold) run 1 5% 1% 5% 8% 13% 4%run 2 6% -4% 4% -6% 15% -10%run 3 10% 2% 4% -2% 15% 6%
Bench-top stability (24h, in 3-fold) 14% -9% - - 1% 9%Freeze-thaw stability -20 °C (3 cycles, in 3-fold) 6% -13% - - 4% 12%
MALDI-sample stability (7 days, in 6-fold) day 0 5% 1% 5% 8% 13% 4%day 7 3% 10% 4% 10% 4% 12%
20 ng/mL calibrant QC-low QC-mediumCV Bias CV Bias CV Bias
IGFBP3 challenge test (in 5-fold) 5% -7% 4% -6% 1% -2%+ 25 ng/mL + 100 ng/mL + 500 ng/mL
CV Bias CV Bias CV BiasSpike recovery (6 different plasma samples) 9% 4% 7% 11% 12% 4%
Klont et al., Analytical Chemistry 2017, 89, 6188-6195.
![Page 22: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/22.jpg)
Quality indicator: oxidation
Klont et al., Analytical Chemistry 2017, 89, 6188-6195.
![Page 23: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/23.jpg)
Quality indicator: 4-spot variability
Klont et al., Analytical Chemistry 2017, 89, 6188-6195.
![Page 24: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/24.jpg)
Confounding factors: sRAGE and COPD
Progressive loss of lung function with a large impact on the quality of life
20% of the smokers develop COPD, more than 200 million people have COPD
Insufficient insight in the molecular mechanisms of COPD
Limited therapeutic options
![Page 25: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/25.jpg)
Expression of the AGER Gene
Fagerberg et al., Molecular & Cellular Proteomics 2014, 13, 397-406.
![Page 26: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/26.jpg)
Receptor for Advanced Glycation Endproducts (RAGE)
sRAGE
![Page 27: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/27.jpg)
RAGE effects
RAGE
S100B
NFƙB-mediated pro-inflammatory responses
Airway inflammation
Reduced cell growth andcell migration, less collagen
production
Alveolar tissue damage
![Page 28: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/28.jpg)
sRAGE methodology
28Klont et al., Talanta 2018, 182, 414-421.
![Page 29: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/29.jpg)
Immunoaffinity LC-MS vs. ELISA
Klont et al., Talanta 2018, 182, 414-421.
![Page 30: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/30.jpg)
Alternative Affinity Agents: Affimers
Tiede et al., Elife 2017, 6, e24903.
![Page 31: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/31.jpg)
Affimer-Based Enrichment vs. Antibody-Based Enrichment (1)
Klont et al., J. Proteome Res., 8, 2018, 2892-2899
![Page 32: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/32.jpg)
Affimer-based Enrichment vs. Antibody-based Enrichment (2)
Klont et al., J. Proteome Res., 8, 2018, 2892-2899
![Page 33: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/33.jpg)
sRAGE and acute smoking
Serum samples COPD patients GOLD I-IV
Healthy controls young (20-40 y) or old (40-80 y) that are either smokers or non-smokers.
2 days 1 hour 2 hours
Stop smoking Serum Smoke 3 cigarettes Serum
Lo Tam Loi, et al., BMJ Open 2013, 3, pii: e002178, 002110.001136/bmjopen-002012-002178.
Young
Non-susceptible
Susceptible
Old
Non-susceptible
COPD
![Page 34: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/34.jpg)
Pouwels et al., Am. J. Respir. Dis. Crit. Care Med., 2018, 198, 1456-1458
![Page 35: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/35.jpg)
Ongay et al., Journal of the COPD Foundation 2016, 3, 560-569.
![Page 36: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/36.jpg)
• Biomarker discovery and validation studies suffer from a lack in well-
characterized biochemical reagents and a lack in consistent guidelines (such as
are available for regulated, quantitative bioanalysis) as well as from a lack of
compliance with such procedures/guidelines when available.
• Preanalytical factors may affect sample quality and ultimately the reliability of
results.
• Confounding factors need to be taken into account when it comes to designing
biomarker discovery and validation studies.
• Data (pre)processing as well as the ensuing statistical analysis may affect the
final results considerably even when the (pre)processed data were identical.
Summary & Conclusions
![Page 37: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/37.jpg)
Acknowledgements
Biomarker Development CenterAlain van Gool (Radboud UMC)
Theo Luider (EMC)Nick ten Hacken (UMCG)
IGF-1 & sRAGEFrank Klont
Daan Pouwels (UMCG)Marrit Hadderingh
Stefan Bakker (UMCG)
Preanalytics & ConfoundersNatalia GovorukhinaTherese Rosenling
Sara Ongay
Data (Pre)processing) & StatisticsChristin ChristinBerend HoekmanFrank Suits (IBM)Peter Horvatovich
University of Groningen
![Page 38: Biomarker Best Practices: confounding factors, preanalytics and … · 2019. 6. 24. · sampling for 0, 30, 120 min. Freeze/thaw cycles. Snap freeze in liq. N. 2 and thaw on ice 1-10](https://reader036.fdocuments.in/reader036/viewer/2022063020/5fe3066c2d19eb78a876cc5e/html5/thumbnails/38.jpg)
The Analytical Biochemistry Research Group