Interlaboratory Reproducibility of Droplet Digital Polymerase Chain

Oct 2, 2017 - Acoustic liquid dispensing of DNA into wells of a 96-well plate was undertaken using an automated Access workstation (Labcyte Inc.) comp...
1 downloads 20 Views 881KB Size
Subscriber access provided by LONDON METROPOLITAN UNIV

Article

Interlaboratory reproducibility of droplet digital polymerase chain reaction using a new DNA reference material format Leonardo B. Pinheiro, Helen O’Brien, Julian Druce, Hongdo Do, Pippa Kay, Marissa Daniels, Jingjing You, Daniel Gerard Burke, Kate Griffiths, and Kerry R. Emslie Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b05032 • Publication Date (Web): 02 Oct 2017 Downloaded from http://pubs.acs.org on October 3, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Interlaboratory reproducibility of droplet digital polymerase chain reaction using a new DNA reference material format §

Leonardo B. Pinheiro,*† Helen O’Brien, ‡ Julian Druce, Hongdo Do, ¶ Pippa Kay, # Marissa Daniels, ┴, Jingjing You, ^ Daniel Burke, † Kate Griffiths,† and Kerry R. Emslie,†



National Measurement Institute, Lindfield, New South Wales, Australia



Research and Development, Australian Red Cross Blood Service, Kelvin Grove,

Queensland, Australia §



Victorian Infectious Diseases Reference Laboratory, Melbourne, Victoria, Australia Olivia Newton-John Cancer Research Institute - Translation Genomics and Epigenomics

Laboratory, Heidelberg, Victoria, Australia #

Agri-Bio Molecular Genetics Biosciences Research Division, Bundoora, Victoria, Australia

┴┴

The Prince Charles Hospital University of Queensland Thoracic Research Centre,

Chermside, Queensland, Australia ^ Save Sight Institute, Sydney Eye Hospital, Sydney Medical School, University of Sydney, New South Wales, Australia *Corresponding author. E-mail: [email protected]. Tel +61 2 84673735. 1 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

ABSTRACT Use of droplet digital PCR technology (ddPCR) is expanding rapidly in the diversity of applications and number of users around the world. Access to relatively simple and affordable commercial ddPCR technology has attracted wide interest in use of this technology as a molecular diagnostic tool. For ddPCR to effectively transition to a molecular diagnostic setting requires processes for method validation and verification, and demonstration of reproducible instrument performance. In this study, we describe the development and characterisation of a DNA reference material (NMIA NA008 High GC reference material) comprising a challenging methylated GC-rich DNA template under a novel 96-well microplate format. A scalable process using high precision acoustic dispensing technology was validated to produce the DNA reference material with a certified reference value expressed in amount of DNA molecules per well. An interlaboratory study, conducted using blinded NA008 High GC reference material to assess reproducibility among seven independent laboratories, demonstrated less than 4.5% reproducibility relative standard deviation. With the exclusion of one laboratory, laboratories had appropriate technical competency, fully functional instrumentation and suitable reagents to perform accurate ddPCR based DNA quantification measurements at the time of the study. The study results confirmed that NA008 High GC reference material is fit for the purpose of being used for quality control of ddPCR systems, consumables, instrumentation and workflow.

INTRODUCTION Over the past few years, use of digital polymerase chain reaction (PCR) technology has rapidly expanded and diversified into a multitude of areas within life sciences. The expansion 2 ACS Paragon Plus Environment

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

in application of digital PCR technology is evident from literature reports in various areas of research including pathogen detection,1-4 monitoring of food and water safety,5-7 and microbial ecology.8 Because digital PCR can resolve small differences in the amount of nucleic acids, the largest uptake of this technology is in clinical research such as biomarker quantification, 9-11 detection of genetic variants in cancer patients,12,13 detection of copy number alterations in stem cells,14 monitoring transplant recipients,15 quantification of massively parallel sequence libraries16-18 and genome editing.19,20 Recognition that digital PCR can provide unprecedented levels of precision, accuracy and resolution for quantification of nucleic acids, together with development and availability of affordable instrumentation, have been the main reasons for the rapid expansion of digital PCR. Transitioning digital PCR technology from a research laboratory to a clinical setting requires processes for method validation and verification, and demonstration of method reproducibility and instrument performance. Many DNA reference materials have been prepared for validation and calibration of specific real-time quantitative PCR (qPCR) assays. Because digital PCR is a primary measurement method that does not require a calibrator and is in principle a counting technique, digital PCR has been used for property value assignment of some DNA reference materials with traceability to the International System of Units.21-25 Reference materials have also been prepared to assess biases in specific steps of a qPCR process such as biases that may arise in measurement of methylated DNA after bisulphite conversion.26 To date, such reference materials often comprise a plasmid DNA solution with an assigned copy number concentration or copy number ratio. Even if a validated PCR protocol is followed, instrument related factors can introduce error or bias to results. Accurate thermal cycling conditions during PCR amplification are critical for reproducible results. Inaccurate thermal cycling temperatures or poor uniformity of 3 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

temperature across heating block units can result in no amplification or low amplification efficiency in wells where optimal temperatures are not reached. PCR assays targeting guanine and cytosine (GC)-rich DNA sequences are especially prone to inefficient amplification resulting from suboptimal denaturation and annealing temperatures.27 This was highlighted in a recent study involving a qPCR based measurement targeting a methylated, GC-rich DNA sequence in which diagnostic results were incorrectly interpreted due to instrument derived denaturation temperature differences between samples within 96-well plates.28 In this study, we describe the development, characterisation and interlaboratory validation of a DNA reference material under a 96-well format designed for digital PCR instrument validation, monitoring of instrument performance, and training in digital PCR instrumentation and procedures. The reference material is produced using a ‘scalable’ high precision robotic acoustic droplet ejection (ADE) (Labcyte®) technology.29 ADE is an ultrasound-based liquid ejection technology with several advantages over traditional pipetting techniques. ADE does not require tips or pins29 for dispensing and is capable of delivering precise nanoliter volumes of solution, thus, utilising much higher DNA concentrations than required for dispensing the same amount of DNA using traditional microliter volume techniques. Together, these features reduce the risk of DNA loss due to adsorption to the surface of the pipette tip or storage tube. Unlike many other DNA reference materials, the reference value is expressed in DNA copy number amount per well and assigned by droplet digital PCR (ddPCR). DNA reference materials in this format, in addition to verifying instrument performance, could also be used to support analytical validation of specific PCR methods. EXPERIMENTAL SECTION DNA materials

4 ACS Paragon Plus Environment

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

The procedure used for production of DNA materials is described in Supporting Information S-1). Briefly, DNA materials were produced by end point PCR amplification of human genomic DNA (Human placental DNA, Sigma-Aldrich product number D 4642) followed by deproteination, ethanol precipitation, high pressure liquid chromatography (HPLC),fractionation and ultrafiltration. CDKN2A_550 is a 550 base pair (bp) amplicon corresponding to part of the human cyclin-dependent kinase inhibitor 2A (CDKN2A) gene promoter region. CDKN2A_183 was prepared by digestion of CDKN2A_550 using MspI restriction enzyme (New England BioLabs) followed by the same purification procedure used for materials produced by end-point PCR amplification. TLX3_304 is a 304 bp amplicon corresponding to part of the human T-cell leukemia homeobox 3 gene promoter region. Methylated CDKN2A_550 (mCDKN2A_550) and methylated TLX3_304 (mTLX3_304) were prepared by in vitro methylation using M.SssI CpG Methyl transferase (New England BioLabs). NMIA NA008 Copy Number Verification Plate – High GC DNA reference material (hereafter referred to as NA008 High GC reference material) consists of a defined number of copies of mTLX3_304 acoustically dispensed into each well of a 96-well PCR microplate. All dilutions of DNA were performed using 1 X TE0.1 buffer (10 mM Tris, 0.1 mM EDTA, pH 8.0). Instrumentation Digital PCR QX100™ and QX200™ Droplet Digital PCR™ systems (Bio-Rad Laboratories Pty Ltd, Australia) including DG8™ cartridges for droplet generators (Bio-Rad Laboratories Pty Ltd, Australia) were used by all laboratories participating in the interlaboratory study for ddPCR measurements, except for one laboratory (Laboratory 4) which used an automated droplet generator (AutoDG™) and DG32™ cartridges for AutoDG™ (Bio-Rad Laboratories Pty Ltd,

5 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 25

Australia). Most Droplet Digital PCR™ systems included a C1000 Touch™ Thermal Cycler (Bio-Rad Laboratories Pty Ltd, Australia) for thermal cycling of droplets in ddPCR assays, except for three laboratories participating in the interlaboratory study; Laboratory 2 used an Arktik Thermal Cycler Type 5020 (Thermo Scientific) and Laboratories 5 and 6 used a T100 Thermal Cycler (Bio-Rad Laboratories Pty Ltd, Australia). For experimental work to identify DNA materials sensitive to thermal cycler temperature uniformity, Mastercycler ep gradient S (Eppendorf), Nexus (Eppendorf) and C1000 (Bio-Rad Laboratories Pty Ltd, Australia) thermal cyclers were employed. The first two of these individual thermal cyclers were chosen as they were known to have inaccurate thermal profile from prior calibration verifications using temperature probes (data not shown). ddPCR data acquisition was performed using QuantaSoft™ software (Bio-Rad Laboratories Pty Ltd, Australia). The procedure used for digital PCR quantification of DNA materials using QX100 Droplet Digital PCR system is given in Supporting Information S-2. Primers and probes sequences used for digital PCR assays of DNA materials are given in Table S-1. Liquid handling Acoustic liquid dispensing of DNA into wells of a 96-well plate was undertaken using an automated Access™ workstation (Labcyte Inc.) comprising an Echo® 550 Liquid Handler (Labcyte Inc.), a 4-axis Precise PF-400 Robot plate handler and a PlateLoc Thermal Microplate Sealer (Agilent Technology, Australia) inside a custom built HEPA filtered air enclosure (LabSystems Pty Ltd, Australia). A Tempo™ automation control software (Labcyte Inc.) was used for protocol workflow definition and running the Access™ workstation. For ddPCR quantification of DNA in wells of a 96-well plate during development work and batch production of the DNA reference material, ddPCR mix was dispensed into wells using a CAS-1200N Liquid handler (Corbett). Calibrated pipettes and

6 ACS Paragon Plus Environment

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

calibrated analytical balance XP205 (Mettler-Toledo) with 0.01 mg resolution were used for gravimetric dilutions of DNA and preparation of digital PCR mixes. High pressure liquid chromatography (HPLC) A prominence SPD-M20A HPLC (Shimadzu) fitted with a Gen-Pak FAX column (Waters) was used to fractionate amplicon restriction fragments by ion-exchange HPLC. Quantitative HPLC measurements of DNA materials were performed using an Ultimate 300 rapid separation liquid chromatography (RSLC) nano system (Dionex) with a PepSwift monolithic column (Dionex). Details of analytical methods used for both ion-exchange HPLC fractionation and quantitative HPLC measurements of DNA materials have been previously described.30 Capillary electrophoresis Amplicon size (bp) was verified on a 2100 Bioanalyzer (Agilent Technologies, Australia) using a DNA 1000 kit as per manufacturer’s instructions. Optical microscopy An optical microscope (Leica DM6000M) with digital CCD camera (Leica DFC490) and 1 µ-Slide VI flat uncoated microscopy chambers (IBID Germany) was used to image droplets generated from the ddPCR™ system droplet generators (Bio-Rad) to determine droplet volume as previously described.31 Acoustic dispensing of DNA combined with ddPCR quantification The procedure of combining high precision acoustic dispensing of a DNA solution with ddPCR quantification of number of dispensed DNA molecules comprised the following steps: (A) quantification and confirmation of identity of high concentration stock of DNA

7 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

material, gravimetric dilution of DNA to a Source Stock at targeted dispensing concentration and confirmatory digital PCR quantification of Source Stock, (B) transfer of Source Stock to individual wells of an Echo qualified 384-well Source Plate (Labcyte Inc.), robotic-driven acoustic dispensing of Source Stock into 96-well x 0.2 ml PCR microplates (BIOPlastics B, Landgraaf Netherlands) and robotic-driven foil sealing of dispensed plates under HEPA filter air environment, (C) ddPCR quantification of the number of DNA molecules present in individual wells of dispensed PCR microplates (Figure 1).

Figure 1. Schematic workflow of combining acoustic dispensing with ddPCR quantification used in the production of NA008 High GC reference material. (A) High concentration DNA material is quantified by HPLC and expected sequence identity and size verified by Sanger DNA sequencing and capillary electrophoresis, respectively. High concentration DNA material is gravimetrically diluted to generate Source Stock at suitable concentration for acoustic dispensing. Source Stock concentration of DNA material is verified by ddPCR. (B) 8 ACS Paragon Plus Environment

Page 9 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Samples from Source Stock of DNA material are transferred to Labcyte Echo qualified Source Plate wells. Single 2.5 nL aliquots of Source Stock of DNA material are acoustically dispensed into individual wells of 96-well PCR microplates and foil sealed using a roboticdriven automated workstation under HEPA filter air enclosure. (C) Preparations of ddPCR mix containing primers and probe are either manually pipetted or robotically dispensed into each well of the microplate. The DNA contents in each well are redissolved, mixed and collected to the bottom of wells by centrifugation and quantified following standard BioRad QX100/200 ddPCR workflow. The number of positive droplets and number of accepted droplets obtained from BioRad QuantaSoft software output is used to calculate the number of DNA molecules in each well using an NMI-designed Microsoft Excel proforma.

(end of Figure 1 legend)

The acoustic dispensing procedure involved transfer of defined nanoliter volumes of Source Stock into individual wells of each 96-well PCR microplate. The acoustic dispensing liquid handling system ejects single or multiple 2.5 nL droplets (≤ 8% coefficient of variation (CV) accuracy; ≤ 5% CV precision) (manufacturer’s specification) from a Source Plate into the wells of an inverted plate (Destination Plate) that is suspended above the Source Plate. The nanoliter volume of DNA solution evaporates immediately after acoustic dispensing, resulting in PCR microplates containing dry DNA molecules at the bottom of individual wells. The sealed plates are stored at -20 °C until used. To assay DNA in a Destination Plate, a defined volume between 20 and 25 µL ddPCR mix containing appropriate primers and probe was either manually pipetted or robotically dispensed (CAS-1200N) into each well of the microplate. The DNA contents in each well were then redissolved by incubating the Destination Plate at 62 °C for 2 min, followed by mixing using a plate vortex (MixMate PCR 96 Eppendorf) and centrifugation (Centrifuge 9 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

5430 Eppendorf) to collect the solution to the bottom of the wells. The QX100/200 ddPCR workflow was then followed for quantification of the number of DNA molecules present in the wells. A simplified DNA redissolving procedure, which excluded the heat incubation step, was used for the interlaboratory study. The simplified method consisted of pipetting 25 µL of ddPCR mix containing appropriate primers and probes into each well of a microplate, mixing by pipetting the reaction solution up and down five times, brief centrifugation of the plate to collect the solution at the bottom of each well followed by the standard QX100/200 ddPCRworkflow. RESULTS AND DISCUSSION Identification of DNA materials sensitive to thermal cycler temperature uniformity The first evidence of an effect of well position on results from ddPCR quantification was noticed when analysing mCDKN2A_550 using an assay (Assay F) targeting a 70 bp sequence with 71% GC content and 9 CpG methylation sites. (Table S-1 and Figure S-1). Results from ddPCR quantification of mCDKN2A_550 using assay F showed poor repeatability among replicate assays placed in wells positioned at the edge of the thermal cycler heating block. To further investigate the thermal cycler well position effect in the assays, a bulk ddPCR mix containing primers, probe and mCDKN2A_550 sufficient for 96 wells was prepared, manually dispensed across an entire 96-well plate (3 replicate plates) and analysed using ddPCR. Thermal cycling of each plate was undertaken on a different thermal cycler. Depending on the thermal cycler unit used, the measured copy number concentration showed up to 56% relative standard deviation (RSD) across the 96 wells of the plate, and up to four fold difference between adjacent wells at the edge of the heating block. The thermal cycler unit with the lowest well to well variation (thermal cycler C) resulted in

10 ACS Paragon Plus Environment

Page 11 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

3.0% RSD in the measured concentration of mCDKN2A_550 across the entire 96 wells (Figure S-2). To verify if the thermal cycler well position effect could be detected using other high GC content DNA, a second series of assays was performed using mTLX3_304. First, a preparation of mTLX3_304 was diluted and then quantified by ddPCR to approximately 1 x 106 copies per µL using TLX3 assay 1 that targets a 112 bp sequence of TLX3_304 with 66% GC content and 13 CpG methylation sites (Supporting Information Figure S-3). The standard procedure for acoustic dispensing followed by digital PCR quantification was used to dispense 10 nL (four 2.5 nL droplets) of mTLX3_304 preparation into each well of a 96well PCR microplate. Digital PCR quantification of mTLX3_304 plates using TLX3 assay 1 (Table S-1) showed a similar thermal cycler well position pattern as detected for mCDKN2A_550. Thermal cycler A, which previously produced poorest uniformity in measured values across wells containing mCDKN2A_550, showed 71% RSD in measured DNA concentration across the plate with consistently lower values around the edge of the plate whereas there was a 3.6% RSD across the plate when using thermal cycler C (Figure S4). These results prompted us to consider the possibility of preparing reference plates predispensed with a defined number of copies of methylated DNA material to be used for verifying thermal cycler uniformity and reproducibility of ddPCR systems. Validation of acoustic dispensing for production of DNA reference materials Validation of acoustic dispensing of DNA was performed using accurately quantified preparations of CDKN2A_183. A HPLC fraction corresponding to CDKN2A_183 was collected and quantified to an estimated concentration of 1.05 x 1010 copies per µL by quantitative HPLC. CDKN2A_183 was gravimetrically diluted to produce three Source 11 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

Stocks: levels 1, 2 and 3 containing 400, 40 and 4 DNA molecules per nL, respectively. The standard procedure for acoustic dispensing followed by digital PCR quantification was used to dispense either 2.5, 5 or 10 nL of CDKN2A_183 preparation into each well of a 96-well PCR microplate in replicates of seven. This resulted, theoretically, in sets of seven wells containing either 4 000, 2 000, 1 000, 400, 200, 100, 40, 20 or 10 copies in each well. Three replicate plates were dispensed. The dispensing format was designed to assess the acoustic dispensing precision of both single and multiple ejections of the 2.5 nL volume (2.5, 5 and 10 nL) for delivery of different numbers of DNA molecules. Quantification of CDKN2A_183 in the wells was performed by ddPCR using CDKN2A assay F as the 183 bp fragment spans the assay F target region within CDKN2A_550. For data analysis, wells with less than 10 000 accepted droplets were excluded. This comprised a complete column plus two additional wells on one of the plates. Data analysis of the number of CDKN2A_183 molecules in each well across the three plates indicated excellent linearity across the range from 10 to 4 000 molecules (r2 = 0.99995) between the nominal number expected assuming an acoustic ejection volume of 2.5 nL (manufacturer’s specification) and the number measured by ddPCR (Supporting information Figure S-5). The relative expanded uncertainty of the number of CDKN2A_183 molecules in each well ranged from 7.0% for 2 500 dispensed molecules to 48% for 10 DNA molecules (Figure 2). The increase in uncertainty observed when dispensing very low numbers of DNA molecules is predominantly due to stochastic effects31,32 rather than increased variability in the acoustic dispensing process. The results indicate that either single or multiple 2.5 nL aliquots can be acoustically dispensed with high precision allowing for a flexible dispensing design format using 96-well PCR microplates. Overall, the results demonstrated that the process of acoustic dispensing of DNA molecules in combination with quantification of dispensed molecules using ddPCR is fit for the purpose of producing DNA reference materials. 12 ACS Paragon Plus Environment

Page 13 of 25

10000 CDKN2A_183 copies/well — measured by ddPCR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1000

100

10

1 (A)

4

2

1

4

2

1

4

2

1

(B)

1

1

1

2

2

2

3

3

3

Figure 2. ddPCR analysis of acoustically dispensed CDKN2A_183. Each data point corresponds to the average of either 6 or 7 replicates for each of 3 plates except for the data point with a nominal 10 copies per well which was based on 2 plates. Error bars represent expanded uncertainty calculated by multiplying the combined standard uncertainty by a coverage factor of between 2.10 and 2.18 depending on the number of wells analysed. This provides a level of confidence of 95% in the expanded uncertainty. Row (A) corresponds to the number of acoustically dispensed droplet ejections and row (B) corresponds to the Source Stock dilutions used: level 1, 2 and 3 containing 400, 40 and 4 DNA molecules per nL, respectively.

(end of Figure 2 legend) Production and characterisation of acoustically dispensed DNA reference material mTLX3_304 was selected for batch production of DNA reference material because, in addition to its generic potential as quality control for digital PCR, mTLX3_304 shows high sensitivity to the accuracy and uniformity of thermal cycler temperature. For preparation of the acoustic dispensing Source stock, a preparation of mTLX3_304 was gravimetrically diluted to approximately 1.2 x 107 copies per µL, with concentration verified by ddPCR using 13 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

TLX3 assay 1. This dilution was chosen with the objective of acoustically dispensing single 2.5 nL aliquots for delivery of approximately 30 000 copies of mTLX3_304 into each well of a 96-well PCR microplate. The target of approximately 30 000 DNA molecules per well was selected to minimise the contribution to uncertainty from the Poisson model assuming the ddPCR would comprise approximately 20 000 droplets31 and, thus, optimise conditions for identifying instrument-related factors that could affect ddPCR performance. The dispensing workflow of a single 2.5 nL aliquot per well maximises the number of plates that can be dispensed from a single well of Source Stock in the source plate. Acoustic dispensing was programmed to consecutively dispense a single aliquot in each well of each column commencing at A01, moving down column 1 to H01 and then returning to A02 etc untill the last aliquot was dispensed into the final well, H12. An automated workflow was programmed for the Access workstation using the Tempo™ software for dispensing a batch of 300 microplates in six sequential sets of 50 plates. Each set of 50 plates was dispensed from a single Source well. The automated dispensing process involved barcoding and foil heat sealing each plate after acoustically dispensing DNA into each of the 96 wells. After dispensing, the batch of plates of NA008 High GC reference material was transferred to storage at -20 °C.

Homogeneity assessment The first and one of the most important steps for determining success in the production of a batch of reference materials is homogeneity assessment. Assessment of the inter-plate homogeneity of the acoustically dispensed batch of NA008 High GC reference material was conducted in accordance with ISO Guide 35.33 The measurand was defined as the average number of mTLX3_304 DNA molecules per well across the eight wells of a column. Homogeneity data was obtained from ddPCR quantification of the number of mTLX3_304 molecules present in each of 96-wells using seven plates randomly selected from the batch. 14 ACS Paragon Plus Environment

Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Initial assessment of the data revealed a slightly higher number of molecules (approximately 10% higher) detected for the first few wells of the first column of each dispensed plate. The trend possibly resulted from a deviation in the volume of the acoustically dispensed aliquot caused by effects from fluid surface tension on the meniscus in the source well at the start of the dispensing cycle for each individual plate (Labcyte, personal communication). In order to assess homogeneity while taking into account the dispensing trend, data analysis was performed by dividing each plate into four quadrants according to the dispensing sequence of columns. Plate quadrants 1, 2, 3 and 4 comprised columns 1 to 3, 4 to 6, 7 to 9, and 10 to 12, respectively. ANOVA of data from each of the four quadrants was used to separate the variation due to precision (within-plate relative standard deviation, Sw,rel) and homogeneity (between-plate relative standard deviation, Sb,rel) and these factors were used for the relative standard uncertainties for precision and homogeneity, respectively. The relative standard uncertainties for both homogeneity (ub,rel) and precision (uw,rel) for all four quadrants were below 3% (Table 1).

Stability assessment Stability of the produced batch of material was assessed using a classical stability study design following ISO guide 3533 recommendations. Short-term stability was conducted at 20 °C and 40 °C for up to 4 weeks to simulate the most extreme conditions that may arise under transport. Long-term stability at -20 °C has been under continuous monitoring for 76 weeks at the time of the study. Both short- and long-term stability data were obtained from ddPCR quantification of the number of DNA molecules per well from three randomly selected plate units at each time point tested. As for assessment of homogeneity, data was collected and analysed across the four quadrant regions comprising columns 1 to 3, 4 to 6, 7 to 9, and 10 to 12 . Results were assessed for trends in the measured number of DNA

15 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25

molecules using a t-test and the regression slope over time. There was no significant linear trend over time for any of the quadrants when plates were stored at either 20 °C or 40 °C for up to 4 weeks (p-values ranging from 0.128 to 0.779). Estimates for the relative standard uncertainty for short-term stability at 40 °C for 4 weeks, an unlikely worst possible transport condition, were below 1.2% per week. There was no significant linear trend over time for any of the quadrants when plates were stored at -20 °C for 76 weeks (p-values ranging from 0.257 to 0.852). ANOVA evaluation of the long-term stability data revealed a significant time point effect (p-values