International Comparison of Enumeration-Based Quantification of

Subscriber access provided by University of Otago Library

Article

An International Comparison of Enumeration-based Quantification of DNA Copy-concentration Using Flow Cytometric Counting and Digital Polymerase Chain Reaction Heebong Yoo, Sang-Ryoul Park, Lianhua Dong, Jing Wang, Zhiwei Sui, Jernej Pavši#, Mojca Milavec, Muslum Akgoz, Erkan Mozio#lu, Philippe Corbisier, Janka Matrai, Bruno Cosme, Janaina Japiassu de Vasconcelos Cavalcante, Roberto Becht Flatshart, Daniel Gerard Burke, Michael Forbes-Smith, Jacob L. H. McLaughlin, Kerry R. Emslie, Alexandra S. Whale, Jim Francis Huggett, Helen Parkes, Margaret Kline, Jo Lynne Harenza, and Peter M Vallone Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b03076 • Publication Date (Web): 10 Nov 2016 Downloaded from http://pubs.acs.org on November 12, 2016

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 27

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

1

An International Comparison of Enumeration-based Quantification of DNA Copy-

2

concentration Using Flow Cytometric Counting and Digital Polymerase Chain Reaction

3

Hee-Bong Yoo,1,2 Sang-Ryoul Park,1,2,∗ Lianhua Dong,3 Jing Wang,3 Zhiwei Sui,3 Jernej

4

Pavšič,4 Mojca Milavec,4 Muslum Akgoz,5 Erkan Mozioğlu,5 Philippe Corbisier,6 Matrai

5

Janka,6 Bruno Cosme,7 Janaina J. de V. Cavalcante,7 Roberto Becht Flatshart,7 Daniel Burke,8

6

Michael Forbes-Smith,8 Jacob McLaughlin,8 Kerry Emslie,8,∗ Alexandra S. Whale,9 Jim F.

7

Huggett,9 Helen Parkes,9 Margaret C Kline,10 Jo Lynne Harenza,10 Peter M. Vallone10

8 9

*Correspondences

10

(S-R Park) [email protected]: correspondence on overall study

11

(K. Emslie) [email protected]: correspondence on digital PCR techniques

12 13

1. Korea Research Institute of Standards and Science, Daejeon, Republic of Korea

14

2. University of Science and Technology, Daejeon, Republic of Korea

15

3. National Institute of Metrology, Beijing, P.R. China

16

4. National Institute of Biology, Ljubljana, Slovenia

17

5. TUBITAK UME National Metrology Institute, Kocaeli, Turkey

18

6. Institute for Reference Materials and Measurements, Joint Research Centre, European

19

Commission, Geel, Belgium

20

7. National Institute of Metrology, Quality and Technology, Xerém, Brazil

21

8. National Measurement Institute Australia, Lindfield, Australia

22

9. LGC, Teddington, UK

23

10. National Institute of Standards and Technology, Gaithersburg, USA

24 25

ACS Paragon Plus Environment


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

1

ABSTRACT

2

Enumeration-based determination of DNA copy-concentration was assessed through an

3

international comparison among national metrology institutes (NMIs) and designated

4

institutes (DIs). Enumeration-based quantification does not require a calibration standard

5

thereby providing a route to ‘absolute quantification,’ which offers the potential for reliable

6

value assignments of DNA reference materials, and International System of Units (SI)

7

traceability to copy number 1 through accurate counting. In this study, 2 enumeration-based

8

methods, flow cytometric (FCM) counting and the digital polymerase chain reaction (dPCR),

9

were compared to quantify a solution of the pBR322 plasmid at a concentration of several

10

thousand copies per microliter. In addition, 2 orthogonal chemical-analysis methods based on

11

nucleotide quantification, isotope-dilution mass spectrometry (IDMS) and capillary

12

electrophoresis (CE) were applied to quantify a more concentrated solution of the plasmid.

13

Although 9 dPCR results from 8 laboratories showed some dispersion (relative standard

14

deviation [RSD] = 11.8%), their means were closely aligned with those of the FCM-based

15

counting method and the orthogonal chemical-analysis methods, corrected for gravimetric

16

dilution factors. Using the means of dPCR results, the RSD of all 4 methods was 1.8%, which

17

strongly supported the validity of the recent enumeration approaches. Despite a good overall

18

agreement, the individual dPCR results were not sufficiently covered by the reported

19

measurement uncertainties. This finding suggests that some laboratories may not have

20

considered all factors contributing to the measurement uncertainty of dPCR, and further

21

investigation of this possibility is warranted.

22 23

Keywords: Enumeration-based quantification, DNA copy-concentration, flow cytometric

24

counting, digital PCR, chemical analytical quantification of DNA, national metrology

25

institute, international comparison

26 27


Page 2 of 27

Page 3 of 27

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


1

INTRODUCTION

2

The increasing application of quantitative molecular methods to measurements across several

3

sectors requires a robust mechanism for their quality control and quality assurance. This is

4

especially true where adequate analytical methods are required to provide accurate

5

measurements on which legislation can be based or clinical decisions can be made. However,

6

a lack of higher-order reference methods and materials is a major hindrance for deriving

7

traceability and measurement comparability, which impacts accreditation and regulatory

8

compliance.

9

In this study, we aimed to address this need by evaluating the performance of 2 different

10

enumeration-based, DNA-quantification methods as candidate reference methods that can be

11

applied without the requirement for calibration standards. Such methods would provide the

12

basis for a substantial increase in the range of molecular measurements that can rely on SI

13

traceability (to 1) through counting.

14

Calibration standards for assuring the performance of a measurement method need to be

15

characterized by a suitable measurement method. In the case of trace amounts of DNA, a

16

calibration standard may be prepared by consecutive gravimetric dilutions of a concentrated

17

DNA stock that can be accurately determined by conventional chemical analysis.1-5 However,

18

the reliability of consecutive gravimetric dilutions of DNA solutions has yet to be confirmed,

19

and the manipulation of DNA materials may lead to their degradation through, for example,

20

shear damage.6 If possible, the value for a DNA reference material should be assigned using

21

methods that are capable of high-accuracy measurement across a DNA concentration range

22

that encompasses the concentration of the reference material. In the case of low or trace

23

concentrations of DNA, enumeration-based quantification may be a suitable approach for

24

absolute DNA quantification. If all target DNA molecules or sequences in a known volume

25

are successfully counted one by one (so called ‘exhaustive counting’), a calibrator commonly

26

needed to determine the relationship between the signal intensity and the sample

27

concentration is not required to determine the DNA concentration. Most simply, a single

28

target DNA molecule could be detected for enumeration using high-sensitivity, laser-induced

29

fluorescence detection provided that the target DNA is large enough to produce detectable

30

fluorescence with appropriate staining dyes. The sample solution should be sufficiently dilute

31

to result in singly separated DNA particles. This approach lead to the measurement principle



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 27

1

of flow cytometric counting (FCM counting) of target DNA particles, which was used in this

2

study.7,8 Alternatively, if target DNA sequences are distributed across a large number of

3

uniformly sized PCR chambers (partitions), they can be amplified by PCR and then

4

enumerated based on the number of PCR-positive partitions and the total number of partitions,

5

provided that a portion of the partitions contain no target DNA sequences. This detection

6

strategy is known as digital PCR (dPCR).9-13 Techniques of dPCR

7

studied not only as a superior alternative to real time PCR but also as an absolute

8

quantification method.14-19 Several NMIs recognized the potential advantage of these 2

9

enumeration strategies, and have developed or investigated intensely as candidate reference

have been actively

10

analytical procedures for absolute quantification of low-level DNA concentrations.7,8, 20-29

11

Although the FCM counting method follows a simple, straightforward measurement principle,

12

there could be measurement biases due to, for example, some DNA particles bypassing the

13

detection zone or multimeric appearances in the detection zone. These possibilities have

14

previously been thoroughly investigated and reduced. For example, multimeric appearances

15

in the detection zone were detected with accordingly larger signal intensities or enlarged

16

exposure times and comprised the largest component of the measurement uncertainty.8 Unlike

17

the FCM counting method, dPCR does not distinguish between partitions containing a single

18

copy of target DNA and partitions containing multiple copies of target DNA. The total

19

number of copies of target DNA across the partitions in dPCR is estimated by Poisson

20

statistics. This estimate has some uncertainty dependent upon the total number of partitions

21

and the proportion of partitions containing at least 1 copy of target DNA.24,25 As for all

22

analytical procedures, dPCR requires careful optimization. For accurate quantification by

23

dPCR, consistent amplification from a single copy of target DNA is required. For this, primer

24

sets, PCR reagents including polymerase and master mix, and thermal cycling conditions

25

need to be evaluated. Unintended melting of double-stranded DNA leads to overestimation

26

because the single strands are distributed independently across partitions, resulting in a higher

27

proportion of partitions containing amplifiable target sequences.27 Several actions required

28

for successful dPCR have been reported.20-29

29

For the acceptance of both FCM-based counting and dPCR as reference analytical procedures,

30

it is important to determine the level of comparability that is achievable among NMIs.

31

Because FCM counting and dPCR are fundamentally different measurement techniques,

32

comparability between these 2 enumeration methods would provide greater confidence in


Page 5 of 27

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


1

their performance. For this reason, an international comparison of enumeration-based

2

quantification of low level DNA was organized in this study.

3

In this comparison, a simple form of DNA was used as the test sample. Plasmid DNA (for

4

example, the pBR322 plasmid of 4.36 kb) can be obtained in a relatively pure form and is

5

detectable by both FCM counting and dPCR. Plasmid DNA can be measured in a circular

6

form (mostly in a supercoiled form) or after linearization with a restriction enzyme. To

7

investigate the applicability of these measurement techniques to both the structurally

8

hindered supercoiled form and the linear form, both forms of a plasmid DNA were tested.

9

Two chemical-analysis methods were also applied, namely high pressure liquid

10

chromatography (HPLC)-isotope-dilution mass spectrometry (IDMS)3,24 and capillary

11

electrophoresis with detection by ultraviolet absorbance (CE-UV)5, which quantify

12

nucleotides (or deoxynucleoside monophosphates [dNMPs]) that are enzymatically released

13

from the target DNA. These methods require substantially higher DNA concentrations than

14

do dPCR or FCM counting. Concentrated DNA samples were prepared, with the

15

gravimetrically determined concentrations. Samples were delivered to testing facilities across

16

several continents. The short-term stability of the test samples the under conditions expected

17

during transport was carefully evaluated after confirming their homogeneity. Participants in

18

the study were either NMIs or DIs of the Bioanalysis Working Group (BAWG) under the

19

Consultative Committee on the Amount of Substance (CCQM) for the Meter Convention.

20 21

EXPERIMENTAL

22

Comparison scheme, sample preparation, and delivery

23

All samples were prepared at the NMI of Korea (KRISS) using gravimetric dilutions,

24

according to the scheme shown in Figure 1. The mixed pool of pBR322 plasmid DNA

25

(#SD0041; Fermentas, Burlington, Canada) was divided into 2 parts. One part was diluted

26

approximately 100-fold in deionized water to generate a high level, intact sample of circular

27

plasmid DNA (Level I, type-C). The other part was linearized using the EcoRI restriction

28

enzyme (#ER0271; Fermentas), which resulted in approximately a 10-fold dilution of the

29

original material. The linearized plasmid was then further diluted 10-fold, resulting in

30

approximately a 100-fold dilution (Level I, type-L) from the original material.



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

1

Both Level I, type-C and type-L were serially diluted by approximately 125,000-fold in a

2

gravimetric manner in 1× Tris-EDTA buffer to generate the Level II samples. Two hundred 1-

3

mL aliquots of Level II samples were prepared with both type-C and type-L samples, which

4

were immediately stored at −70ºC. Sample homogeneity was tested by FCM counting of 10

5

randomly chosen aliquots from both the Level II, type-C and type-L samples. Sample

6

stability was tested at 4 time intervals (1 day, 2 days, 5 days, and 10 days) and at 4 different

7

storage temperatures (−20ºC, 4ºC, 25ºC, and 45ºC). For each time and temperature, 2 aliquots

8

were randomly chosen for each type of Level-II aliquot and quantified at least 3 times by

9

FCM counting. Following the homogeneity and stability studies, each participating laboratory

10

was sent 5 vials of both Level II type-C and type-L samples packaged with cool-packs.

11 12 13

Figure 1. Schematic representation of the CCQM P-154 pilot study: absolute quantification of DNA

14 15

Participants and analytical methods. The participating laboratories and their analytical

16

methods are listed in Table 1. The full names of the participating institutes are as follows:

17

NMIA, National Metrology Institute of Australia; INMETRO, National Institute of Metrology,

18

Quality and Technology; NIM, National Institute of Metrology; IRMM, Institute of Reference

19

Materials and Measurements; KRISS, Korea Institute of Standards and Science; NIB,

20

National Institute of Biology; TUBITAK UME, National Metrology Institute of Turkey; LGC,


Page 6 of 27

Page 7 of 27

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


1

Laboratory of Government Chemist; NIST, National Institute of Standards and Technology.

2

Refer to Table 1 for the economy of each institute. Each of the applied analytical methods is

3

described in detail below. The details of dPCR protocol used by each participant are

4

described in the references shown in Table 1.

5 6 7

Table 1. Study participants and applied methods Institute* NMIA NMIA

Method d-PCR HPLC-IDMS (for Level I)

INMETRO

d-PCR

Slovenia Turkey United Kingdom

NIM IRMM KRISS KRISS NIB TŰBITAK-UME LGC

United States

NIST

Economy Australia Brazil China European Union Korea, Republic of

8

References

d-PCR d-PCR FCM counting CE-UV (for Level I) d-PCR d-PCR d-PCR

Instrumentation QX-100 (Bio-Rad) 6510 Q-TOF LCMS (Agilent) QX-100 (Bio-Rad) Biomark HD (Fluidigm) Biomark HD (Fluidigm) Home-made flow cytometry system G1600AX (Agilent) QX100 (Bio-Rad) Quant Studio3D (Thermo Fisher Sci.) Biomark HD (Fluidigm)

d-PCR

Biomark HD (Fluidigm) & QX100 (Bio-Rad)

31,32

36 24 38 33,34 28,37 8 5 30 35 21,39

*

Refer to the text for the detailed name of each participating institute.

9 10 11

dPCR

12

Three dPCR platforms were used in this comparative study (Table 1). Three participants used

13

only the BioMark platform, (Fluidigm; chip-based system), 1 participant used the

14

QuantStudio 3D platform (Thermo Fisher Scientific; chip-based system), and 3 participants

15

used only the QX100 platform (Bio-Rad; emulsion-based system). In addition, 1 participant

16

used both the BioMark and QX100 systems. One participant measured the Level 1, type-L

17

material using dPCR after the necessary dilution. Well-documented procedures for dPCR20,21

18

were suggested as standard protocols. Nevertheless, participants were allowed to choose their

19

own instrumentation and experimental protocols. An example of how each laboratory utilized

20

dPCR to quantify a DNA target is provided in Table 1 in the form of reference publications.

21

In addition, the details of dPCR conditions of each participant are provided in Table S-1A and

22

B. As the aim of this study was to assess the general comparability among the enumeration



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

1

methods, it was beyond the scope of this study to investigate whether specific experimental

2

conditions affected the quantification results.

3

FCM counting

4

Level-II samples were diluted 5-fold in a Tris-based running buffer containing 10% (v/v)

5

DMSO and 0.1× SYBRTM-Gold dye. Part of this diluted solution was introduced by pressure

6

into an approximately 1-m long, 50 µm × 50 µm square capillary tube (Polymicro

7

Technologies, Phoenix, AZ) for flow cytometry. The DNA particles were counted until the

8

mixture was completely driven out from the capillary, with fresh running buffer following the

9

sample solution. For this procedure, 0.07 MPa was applied for 2 min, with 20 kV applied to

10

the end of the capillary where the direction of the electric field was opposite to the direction

11

of the flow. The details of the experimental conditions are available elsewhere.8

12 13

Nucleotide analysis by CE-UV

14

Level-I samples were concentrated by approximately 10-fold by vacuum drying, and the

15

change in the sample mass following drying was gravimetrically determined. The

16

concentrated DNA was enzymatically digested to dNMPs using DNase I (PN 78311; USB,

17

Cleveland, OH) and phosphodiesterase I (PN 20240; USB) (Figure 1). The dNMPs were

18

separated and quantified by CE-UV using a G1600AX instrument (Agilent), with iothalamate

19

serving as the internal standard as described previously.5 Calibration of each nucleotide was

20

performed with nucleotide calibration-standard solutions that were determined by inductively

21

coupled plasma-optical emission spectroscopy (ICP-OES) in conjunction with purity

22

determinations by high-performance liquid chromatography-ICP-mass spectrometry (HPLC-

23

ICP-MS). The quantity of DNA was estimated by dividing the quantity of each dNMP by its

24

number in the DNA sequence, which resulted in 4 predictions for dAMP, dTMP, dGMP, and

25

dCMP. The average of these 4 values was assigned as the quantity of the 10-fold concentrated

26

Level-I DNA. Considering the gravimetric concentration factor and the dilution factor, the

27

quantity of Level-II DNA was also estimated.

28 29

Nucleotide analysis by HPLC-IDMS


Page 8 of 27

Page 9 of 27

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


1

Level-I, Type-L study materials were analyzed as they were received. A sample blend was

2

prepared by adding a mixture of stable isotopically labeled dNMPs to each of 5 100-mg

3

subsamples of the study material. A calibration blend was prepared by adding the same stable

4

isotopically labeled dNMP mixture to a mixture of reference dNMPs. Both the sample and

5

calibration blends were enzymatically hydrolyzed using a mixture of DNase I and

6

phosphodiesterase I. After hydrolysis, each blend was ultrafiltered through Pall Nanosep 3K

7

Ω membranes and the filtrate was analyzed by LC-MS on an Agilent 6510 QTOF system.

8

Details of the measurement system, reagents, and calculations were published previously.24

9 10 11

RESULTS AND DISCUSSION

12

Testing of comparison samples

13

The results of homogeneity testing using FCM counting for the study materials are shown in

14

Figure 2. For this test, 10 vials were randomly chosen from a batch of 200 vials. The relative

15

standard deviations (RSDs) of measurements for each vial ranged from 0.2 to 2.5% (n = 3).

16

The results of F-testing (df1 = 9, df2 = 20) were 0.35 and 0.89 for type-C and type-L samples,

17

respectively. Those values were substantially smaller than Fcritical value of 2.39 at α = 0.05.

18

Therefore, it was concluded that test samples were prepared homogeneously, within the

19

measurement uncertainty (~5%). It was necessary to avoid the potential problem of high

20

viscosity of the concentrated DNA stock solution while preparing homogenous samples.

21

Preparing Level-I samples after a ~100-fold dilution of the stock solution to ~5 ng/µL (Figure

22

1) led to satisfactory homogenization as expected, which was necessary to assure the

23

comparability of measurement results obtained by different methods. Agreement in the

24

measurement results for different vials obtained during measurement-repeatability testing by

25

both CE-UV and LC-IDMS demonstrated the homogeneity of the Level-I samples (data not

26

shown).

27



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

1 2 3 4 5

Figure 2. The results of homogeneity testing by FCM counting of Level-II samples: (A) typeC samples; (B) type-L samples. Each sample vial was measured 3 times, and 10 vials were measured simultaneously in exactly the same manner. Error bars indicate the CVs of 3 repeated measurements.

6 7

The short-term stability of the test samples during international delivery was examined by

8

measuring samples left at room temperature or at an elevated temperature (45ºC) for several

9

days. Measurement was performed by FCM counting, the results of which are presented in

10

Figure S-1. No significant reduction in the concentration of pBR322 was observed for up to

11

10 days. Instead, a few samples showed a slightly elevated level (not exceeding 3%, indicated

12

with red arrows) after 5 or 10 days, which was not directly linked to the storage temperature.

13

This was more likely due to the evaporation of water through loose caps. After this

14

observation, newly prepared samples were firmly sealed with great care, and no observable

15

changes beyond measurement uncertainty were observed after storage at 45ºC for 10 days

16

(data not shown). The pBR322 plasmid was stable in 1× TE buffer against chemical or

17

physical degradation for up to 10 days, indicating that accidental exposures to ambient

18

temperatures during sample delivery would not cause a problem. No considerable exposure to

19

a degrading environment during sample delivery was reported. All participating laboratories

20

(except for 1) received the test samples within 4 days. The laboratory in question received the

21

samples 7 days after shipment, but were stored in a refrigerator for some of the transit time.

22


Page 10 of 27

Page 11 of 27

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


1

Results obtained with the linearized pBR322 (type-L sample)

2

The reported results for the linearized pBR322 plasmid (type-L samples) are compared in

3

Table 2. One result by dPCR (result report 1) was more than an order of magnitude lower

4

than the other results. This result was reported by a laboratory relatively new to performing

5

dPCR. Subsequently, it was found that the calculation of the results did not include the

6

consideration of the dilution used to prepare the final test solutions. The dilution factor-

7

corrected result is shown as result report 5 in Table 2 and was relatively well aligned with the

8

other results. During the subsequent data processing, the initial result of the clearly identified

9

mistake (Result report 1) was excluded. For the enumeration methods, dilution of original

10

samples down to the proper concentration range (the linear dynamic range) for the selected

11

analytical method was frequently necessary. Considerations of such dilution factors when

12

calculating the results should not be neglected.

13

Table 2. Result reports for the linearized plasmid DNA sample. Unless noted, the samples were Level II, type-L. Results of Level-I samples were normalized to the concentrations of Level II, based on the dilution factor from Level I to Level II (refer to Table 3).

14 15

16

Result report

Method

Instrumentation

1 2 3 4 5 6 7 8 9 10 11 12 13

d-PCR d-PCR d-PCR d-PCR d-PCR ID/MS (Level I) d-PCR CE-UV (Level I) direct counting d-PCR d-PCR d-PCR (Lelevl I) d-PCR

Bio-Rad QX-100 Bio-Rad QX-100 Thermo Fisher QuantStudio 3D Fluidigm Biomark HD Bio-Rad QX-100 Agilent 6510 QTOF LC-MS Fluidigm Biomark HD Agielnt G1600AX Home-made Fluidigm Biomark HD Bio-Rad QX100 Bio-Rad QX100 Fluidigm Biomark HD

Value (cp/mg) 329 6640 6730 6990 7304 7550 7610 7680 7880 8060 8460 8840 9090

Uncertaint Coverage Factor y (cp/mg) 4.8 2.0 370 2.0 1050 2.0 560 2.6 37 2.0 220 2.0 240 2.2 500 2.0 490 2.0 550 2.0 250 2.0 450 2.0 920 2.0

17 18

While assessing the comparability of the applied methods, the mean of 9 dPCR results, 7747

19

copies/mg, was taken as the representative value for dPCR. The standard uncertainty of this

20

mean value was calculated to be 304 copies/mg (the standard deviation of 912 copies/mg

21

divided by the square root of nine, which was the number of independent results). Then, the

22

expanded uncertainty of the mean of the dPCR data set was calculated to be 687 copies/mg



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

1

based on the coverage factor of 2.26 (95% confidence level, degree of freedom = 8). This

2

representative result for dPCR is compared with the results of other methods in Figure 3. The

3

results of all 4 methods were in good agreement (RSD = 1.8%) within their measurement

4

uncertainties. Between 2 enumeration approaches performed without the use of calibration

5

standards, the relative deviation from their average value was only 0.9% (7747 vs. 7880

6

copies/mg). The close agreement of results obtained using methods with totally different

7

measurement principles strongly supported the validity of those methods. Therefore, the

8

validity of the new approaches of enumeration-based quantification of DNA copy-

9

concentration was confirmed through this comparison study, which was unique in terms of

10

employing multiple methods performed at multiple sites.

11 12 13 14 15 16 17

Figure 3. Comparison of the results obtained with type-L samples using different quantification methods. The results obtained for Level-I samples were normalized to the predictions of Level-II sample, based on the gravimetric dilution factors from Level I to Level II. The representative dPCR results are presented as the mean of all dPCR results, with an expanded uncertainty for the 95% confidence interval. Error bars shown for other methods are also their expanded uncertainties for the 95% confidence interval.

18


Page 12 of 27

Page 13 of 27

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


1

The observed agreement confirms the absence of a significant hidden bias in FCM counting

2

using unique homemade instrumentation. The possible underestimation of FCM-counting due

3

to some DNA particles straying away from the detection zone was confirmed to be

4

insignificant in this comparison study. Although FCM counting has a limitation in terms of

5

the lack of specificity for target DNA sequences, the method can be used as a credible

6

analytical method for quantifying pure DNA reference materials. FCM counting operates

7

under a straightforward measurement principle and enables real-time monitoring and

8

recording of raw counting signals (Figurer S-2), which facilitates the ready detection of errors

9

during measurement.

10 11

dPCR, the major enumeration approach adopted by several NMIs and DIs, was validated for

12

the first time through a comparison among multiple laboratories, in parallel with other

13

measurement methods. Although the dispersion of dPCR results was not insignificant (RSD =

14

11.8%), the distribution was rather symmetric to the mean value that was in close agreement

15

with the results of the other methods used. Therefore, dPCR in general is not likely to carry

16

substantial measurement bias. The dispersion of dPCR results could have been in part due to

17

the different experimental conditions optimized at each laboratory. As shown in many

18

reports,20-29 dPCR requires sophisticated optimization of experimental conditions to avoid

19

measurement biases. Each participant in this comparative study freely determined the

20

optimum measurement conditions, including the PCR priming sites, reagents, and

21

instrumentation, which was necessary to assess the comparability of dPCR techniques as

22

practiced at individual laboratories. Variations in dPCR protocols used among the participants

23

are presented in Table S-1. Briefly, 3 laboratories analyzed the samples using a single PCR

24

assay, 3 laboratories used 2 PCR assays, while 2 laboratories utilized either 4 or 8 separate

25

PCR assays. The PCR assays targeted sequences between nucleotides positions 748 and 2105

26

bp, or between 2900 and 4276 bp on the pBR322 plasmid (Accession Number J01749). At

27

this stage, the major causes of variation from the mean value are uncertain, and further

28

investigation on those causes is beyond the scope of this study. However, we are rapidly

29

gaining knowledge on dPCR performance characteristics,20-41 and comparability of dPCR

30

results among multiple laboratories may be significantly improved in the near future. At

31

present, the results of is study suggest an RSD of ~12% as the range of possible scattering of

32

dPCR results from multiple laboratories, if experimental conditions are not restricted. As



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

1

shown in Figure 4, some results were barely within the range of the representative dPCR

2

result, but with large measurement uncertainties, whereas some laboratories reported

3

substantially smaller uncertainties. Therefore, further investigations on the proper assessment

4

of the measurement uncertainty of dPCR are necessary.

5

6 7 8 9

Figure 4. Results of dPCR quantification for type-L samples aligned with their mean and its expanded uncertainty for the 95% confidence interval. The error bar of each result also indicates the reported expanded uncertainty for the 95% confidence interval.

10 11

The results of orthogonal chemical analysis (HPLC-IDMS and CE-UV), normalized using the

12

gravimetric dilution factors from Level I to Level II, deviated from the average of all methods

13

only by −2.1% and −0.4%, respectively (Table 2). The originally reported data and their

14

conversions to the Level-II concentrations are shown in Table 3, which also includes a dPCR

15

result. Even with a dilution of ~1/125,000, significant deviations were not evident, which

16

strongly suggested that gravimetric dilution of a plasmid solution can be performed reliably.

17

Note that dilution was performed in only 2 steps to simplify the procedure (Figure. 1). Unlike


Page 14 of 27

Page 15 of 27

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


1

the enumeration methods, the chemical-analysis methods were calibrated with calibration

2

standards, which were nucleotide standard solutions quantified by ICP-OES to measure the

3

quantity of phosphorus in the nucleotides. The credibility of the nucleotide-measurement

4

methods was also demonstrated by the good agreement with enumeration-based methods,

5

along with the validity of gravimetric dilution.

6 7 8 9

10

Table 3. Report results for a Level-I, type–L sample and the predicted Level-II concentrations based on the dilution factor from Level I to Level II Result report 6 8

HPLC-IDMS CE-UV

12

dPCR

Method

Report data (copies/mg) (9.54 ± 0.28) x 108 (9.70 ± 0.63) x 108 (1.116 ±0.057) x109

Dilution factor* 7.918 x 10-6

Prediction of Level II (copies/mg) 7550 ± 220 7680 ± 500 8840 ± 450

* The uncertainty of dilution factor was ignored as the weighing uncertainties were relatively negligible.

11 12

Results for the circular pBR322 plasmid (type-C sample)

13

The results of the analysis of intact, circular pBR322 plasmid DNA (type-C samples) are

14

shown in Table 4. Three out of 5 dPCR results were substantially lower than the results of

15

FCM counting and CE-UV, which were not susceptible to the target-DNA structure during

16

quantification (internal communication with the KRISS laboratory). One dPCR result (result

17

report 4) was obtained after on-site linearization of the type-C sample, which dramatically

18

raised the result to ~80% of that of FCM counting.

19

The possibility of underestimating circular DNA concentrations by dPCR was revealed by the

20

fluorescence responses observed during real-time PCR. As shown in Figure S-3, the

21

fluorescence response curves for circular DNAs were markedly less robust compared to those

22

observed with linearized DNA, which implies that a substantially higher failure rate may

23

occur during PCR amplification of circular DNA. A substantial portion of circular DNA is in

24

the form of compactly packed supercoils (Figure S-4), which could structurally hinder PCR

25

primers from accessing the target sequences. Interestingly, 1 participant reported a dPCR

26

result (result report 6) that closely agreed with the FCM-counting result. The particular dPCR

27

conditions used for this exceptional result featured a specific PCR master mix combined with



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 27

1

a fluorescent probe.33 This finding raises the possibility of accurate quantification of circular

2

DNAs by dPCR without linearization.

3 4 5 6 7

8

Table 4. Results for the circular pBR322 plasmid (type-C sample). The reported values are listed together with the expanded uncertainty and coverage factor to provide the 95% confidence level. The result for each Level-I sample was normalized to that predicted for Level II, using the gravimetric dilution factor from Level I to Level II. Result report

Method

Instrumentation

Value (cp/mg)

1 2 3 4 5 6

d-PCR d-PCR d-PCR d-PCR CE-UV d-PCR

Thermo Fisher QuantStudio 3D Bio-Rad QX-100 Fluidigm Biomark HD Fluidigm Biomark HD Agielnt G1600AX Fluidigm Biomark HD

2690 2710 4140 6980 8200 8680

Expanded uncertainty 250 180 330 530 700 620

7

direct counting

Home-made

8700

500

Coverage factor 2.0 2.0 2.6 2.0 2.0 2.0 2.0

9 10 11

Other noteworthy findings

12

Estimation of the sample volume or partition volume. During FCM counting, the sample

13

volume is typically about 1.5 µL. Accurate determination of this small volume was confirmed

14

only after demonstrating satisfactory comparability among several different measurement

15

methods.8 For dPCR-based methods, an accurate estimation of the partition volume is also

16

critically important. While calculating the DNA concentration, the average number of

17

counted entities per partition is divided by the partition volume. An inaccurate determination

18

of the partition volume directly leads to an error in the final concentration result. This was

19

exemplified by a recent report in which a measurement bias was only removed after

20

correcting the partition volume to the self-measured value.28 The partition volume in dPCR is

21

often less than 1 nL, and accurate determination of this volume is challenging. In this

22

comparison study, some participants measured the partition volume together with its

23

associated uncertainty, while other participants used the manufacturer-defined partition

24

volume, which has no stated uncertainty. The information on the partition volume provided

25

by the manufacturers was not sufficient to affirm the accuracy and evaluate the related

26

uncertainty. One participant demonstrated a dramatic difference in the final result: −18% or


Page 17 of 27

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


1

+27% of the reference value, depending upon whether the partition volume was as provided

2

by the manufacturer or self-measured, respectively. It is likely that variation in determining

3

the partition volume contributed to the observed variation in the reported results. Therefore,

4

the use of accurately determined (self or given) sample or partition volumes should be

5

considered to ensure accuracy and comparability of enumeration-based quantification

6

methods.

7

For this study, participating laboratories were required to report the results in terms of

8

copies/mg, rather than the more common unit of copies/µL. Glycerol is a common ingredient

9

in PCR master mixes and has a density of 1.261 g/cm3, making the density of the PCR master

10

mix larger than the density of the DNA solution. Depending on whether laboratories mixed

11

the DNA sample and dPCR mix gravimetrically or volumetrically, it was necessary for the

12

laboratory to account for the difference in density between these 2 solutions when reporting

13

the results in copies/mg. During FCM counting, the density of the final sample solution was

14

1.02 due to the presence of 10% of DMSO, which has a density of 1.10 g/cm3 at room

15

temperature.

16 17

Platform dependence of dPCR results. The 2 main instrument platforms used in this

18

comparison, chip-based dPCR from Fluidigm and droplet dPCR from Bio-Rad, did not show

19

a significant difference in their averaged results, which were 7940 copies/mg vs. 7980

20

copies/mg, respectively. Therefore, the choice of instrument between these 2 platforms could

21

be a minor factor affecting the comparability of dPCR results. However, a large variation in

22

intra-laboratory measurement results was observed using the newer platform from Thermo

23

Fisher Scientific, which may have been due to the large uncertainty in determining the

24

partition volume.

25 26

CONCLUSIONS

27

In this study, the comparability of enumeration-based quantification of plasmid DNA by

28

FCM counting and dPCR was evaluated through an international comparison among national

29

metrology institutes and equivalent organizations. Tight agreement (0.9% deviation from the

30

average value) between the result of FCM counting and the mean of the dPCR results was



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

1

demonstrated for linearized plasmid DNA, which is impressive considering that the

2

measurements were made without the help of calibration standards (absolute quantification).

3

Furthermore, the results of the enumeration methods also closely agreed with the results of 2

4

orthogonal chemical analysis methods (HPLC-IDMS and CE). The RSD of all 4 methods was

5

1.8%. Therefore, all 4 measurement approaches were confidently validated through this

6

comparison, which was unique in terms of the performance of multiple measurements at

7

multiple sites using multiple approaches. Regarding individual dPCR results, however, the

8

dispersion of the results was not insignificant (RSD = 11.8%). The scattering was not fully

9

accounted for by the reported measurement uncertainties, which suggests the need for further

10

improvement and investigation. As expected, dPCR underestimated the concentration of the

11

circular plasmid, although a possible resolution for this problem was suggested. The

12

importance of accurately determining sample volumes and partition volumes in enumeration-

13

based quantification was also addressed in this study.

14

The results of this study demonstrated a rational basis for the direct counting of nucleic acid

15

(plasmid) copy number, validated by orthogonal approaches, strongly supported the validity

16

of these enumeration approaches for absolute DNA quantification, and allow for claims of SI

17

traceability to 1 through counting.

18 19

Acknowledgements

20

The cost for the preparation, characterization, and delivery of the test samples of this

21

international comparison was fully covered by the KRISS internal research fund,

22

Establishment of measurement standards for quality of life.

23 24

Conflict of Interest Disclosure

25

The authors declare no competing financial interest’.

26 27 28


Page 18 of 27

Page 19 of 27

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


1

References

2

(1) Yang, I.; Han, M.-S.; Yim, Y.-H.; Hwang, E.; Park, S.-R. Anal. Biochem. 2004, 335, 150-

3

161.

4 5

(2) Holden, M. J.; Rabb, S. A.; Tewari, Y. B.; Winchester, M. R. Anal. Chem. 2007, 79,

6

1536-1541.

7 8

(3) O'Connor, G.; Dawson, C.; Woolford, A.; Webb, K. S.; Catterick, T. Anal. Chem. 2002, 74,

9

3670-3676.

10 11

(4) Donald, C. E.; Stokes, P.; O'Connor, G.; Woolford, A. J. J. Chromatogr. B Anal. Technol.

12

Biomed. Life Sci. 2005, 817, 173-182.

13 14

(5) Hong, N.-S.; Shi, L.-H.; Jeong, J.-S.; Yang, I.; Kim, S.-K.; Park, S.-R. Anal. Bioanal.

15

Chem. 2011, 400, 2131-2140.

16 17

(6) Yoo, H.-B.; Lim, H.-M.; Yang, I.; Kim, S.-K.; Park, S.-R. J. Biophys. Chem. 2011, 2,

18

102-111.

19 20

(7) Lim, H.-M.; Yoo, H.-B.; Hong, N.-S.; Yang, I.; Han, M.-S.; Park, S.-R. Metrologia 2009,

21

46, 375-387.

22 23

(8) Yoo, H.-B.; Oh, D.; Song, J.-Y.; Kawaharasaki, M.; Hwang, J.; Yang, I.; Park, S.-R.

24

Metrologia 2014, 51, 491-502.

25 26

(9) Sykes, P. J.; Neoh, S. H.; Brisco, M. J.; Hughes, E.; Condon, J.; Morley, A. A.

27

BioTechniques 1992, 13, 444–449.

28 29

(10) Pohl, G.; Shih, I. M. Expert Rev Mol. Diagn. 2004, 4, 41–47.

30 31

(11) Pekin, D.; Skhiri, Y.; Baret, J. C.; Le-Corre, D.; Mazutis, L.; Ben-Salem, C.; Millot, F.;

32

El Harrak, A.; Hutchison, J. B.; Larson, J. W.; Link, D. R.; Laurent-Puig, P.; Griffiths, A. D.;



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

1

Taly, V. Lab Chip 2011, 11, 2156–2166.

2 3

(12) Hindson, B. J.; et al. Anal. Chem. 2011, 83, 8604–8610.

4 5

(13) Baker, M. Nature Methods 2012, 9, 541–544.

6 7

(14) Kim, T. G.; Jeong S.-Y.; Cho, K.-S. Biotechnology Reports 2014, 4, 1-4.

8 9

(15) Drandi, D.; et al. J. Molecular Diagnostics 2015, 17, 652-660.

10 11 12

(16) Coudray-Meunier, C.; Fraisse, A.; Martin-Latil, S.; Guillier, L.; Fach, P.; Pérelle, S. Int. J. Food Microbiol. 2015, 201, 17-26.

13 14

(17) Köppel, R.; Bucher, T. Eur. Food Res. Technol. 2015, 241, 427-439.

15 16

(18) Köppel, R.; Bucher, T.; Waiblinger, H.-U. Eur. Food Res. Technol. 2015, 241, 521-527.

17 18 19

(19) Taylor, S. C.; Carbonneau, J.; Shelton, D. N.; Bovin, G. J. Virol. Methods 2015, 224, 5866.

20 21

(20) Bhat, S.; Curach, N.; Mostyn, T.; Bains, G. S.; Griffiths, K. R.; Emslie, K. R. Anal.

22

Chem. 2010, 82, 7185–7192.

23 24

(21) Sanders, R.; Huggett, J. F.; Bushell, C. A.; Cowen, S.; Scott, D. J.; Foy, C. A. Anal.

25

Chem. 2011, 83, 6474–6484.

26 27

(22) Whale, A. S.; Huggett, J. F.; Cowen, S.; Speirs, V.; Shaw, J.; Ellison, S.; Foy, C. A.;

28

Scott, D. J. Nucleic Acids Res. 2012, 40, e82.

29 30

(23) Sanders, R.; Mason, D. J.; Foy, C. A.; Huggett, J. F. PLoS One 2013, 8, e75296.

31


Page 20 of 27

Page 21 of 27

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


1

(24) Burke, D. G.; Dong, L.; Bhat, S.; Forbes-Smith, M.; Fu, S.; Pinheiro, L.; Wang, J.;

2

Emslie, K. R. Anal. Chem. 2013, 85, 1657–1664.

3 4

(25) Bhat, S.; Herrmann, J.; Armishaw, P.; Corbisier, P.; Emslie, K. R. Anal. Bioanal. Chem.

5

2009, 394, 457-467.

6 7

(26) Devonshire, A. S.; Honeyborne, I.; Gutteridge, A.; Whale, A. S.; Nixon, G.; Wilson, P.;

8

Jones, G.; McHugh, T. D.; Foy, C. A.; Huggett, J. F. Anal. Chem. 2015, 87, 3706-3713.

9 10

(27) Bhat, S.; McLaughlin, J. L.; Emslie, K. R. Analyst 2011, 136, 724-732.

11 12

(28) Corbisier, P.; Pinheiro, L.; Mazoua, S.; Kortekaas, A. M.; Chung, P. Y.; Gerganova, T.;

13

Roebben, G.; Emons, H.; Emslie, K. Anal. Bioanal. Chem. 2015, 407, 1831-1840.

14 15 16

(29) Féllix-Urquidez, D.; Pérez-UrquizaM.; Valdez Torres, J. B.; León-Félix, J.; GarciaEstrada, R.; Acatzi-Silva, A. Anal. Chem. 2016, 88, 812-819.

17 18

(30) Pavšič, J.; Žel, J.; Milavec, M. Bioanal. Chem. 2015, 408, 107-121.

19 20

(31) Kline, M. C.; Romsos, E. L.; Duewer, D. L. Anal. Chem. 2016, 88, 2132-2139.

21 22

(32) Duewer, D. L.; Kline, M. C.; Romsos, E. L. Anal. Bioanal. Chem. 2015, 407, 9061-9069.

23 24

(33) Dong, L.; Yoo, H.-B.; Wang, J.; and Park, S.-R. Scientific Reports 2016, 6, 24230.

25 26

(34) Dong, L.; Meng, Y.; Sui, Z.; Wang, J.; Wu, L.; Fu, B. Scientific Reports 2015, 5, 13174.

27 28

(35) Nivedita, Majumdar.; Thomas, Wessel.; and Jeffrey, Marks. PLoS One 2015, 10,

29

e0118833

30 31

(36) Pinheiro, L. B.; et al. Anal. Chem. 2012, 84, 1003-1011.

32 33

(37) White, H.; et al. Leukemia. 2015, 29, 369-376.



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

1 2

(38) Flatschart, R. B.; Almeida, D. O.; Heinemann, M. B.; Medeiros, M. N.; Granjeiro, J. M.;

3

Folgueras-Flatschart, A. V. J. Physics. Conference Series 2015, 575, 12038-12041.

4 5

(39) Whale, A. S.; Cowen, S.; Foy, C. A.; Huggett, J. F. PLoS One 2013, 8, e58177.

6 7

(40) White III, R. A.; Blainey, P. C.; Fan, C. H.; Quake, S. R. BMC Genomics 2009, 10, 116.

8 9 10

(41) Hindson, C. M.; Chevillet, J. R.; Gallichotte, E. N.; Ruf, I. K.; Hindson, B. J.; Vessella, R. L.; Tewari, M. Nature Methods 2013, 10, 1003-1005.

11 12


Page 22 of 27

Page 23 of 27

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

1


For TOC only

2 3 4 5 6



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

132x78mm (96 x 96 DPI)


Page 24 of 27

Page 25 of 27

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


132x53mm (96 x 96 DPI)



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

63x48mm (96 x 96 DPI)


Page 26 of 27

Page 27 of 27

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


119x94mm (96 x 96 DPI)


International Comparison of Enumeration-Based Quantification of

Recommend Documents