TECHNICAL NOTE pubs.acs.org/ac
Highly Sensitive Pyrosequencing System with Polymer-Supported Enzymes for High-Throughput DNA Analysis Masataka Shirai,†,* Mari Goto,† Shigeya Suzuki,‡ Kenji Kono,§ Tomoharu Kajiyama,† and Hideki Kambara† †
Central Research Laboratory, Hitachi, Ltd., 1-280, Higachi-koigakubo Kokubunji-shi, Tokyo, Japan Noda Development, Group Plannig & Administration Department, Kikkoman Biochemifa Company, 376-2, Kamihanawa, Noda-shi, Chiba, Japan § Product Development & Marketing Division, Hitachi Maxell, Ltd., 1, Koizumi, Oyamazaki, Otokuni, Kyoto, Japan ‡
ABSTRACT: A highly sensitive massively parallel pyrosequencing system employing a gel matrix to immobilize enzymes at high density in microreaction chambers is demonstrated. Reducing the size of microreaction chambers in a DNA analyzer is important to achieve a high throughput utilizing a commercially available detection device or camera. A high-performance system can be attained by detecting signals from one reaction chamber with one photopixel of around several micrometers by utilizing a 1:1 image magnification. However, the use of small beads immobilizing DNA has a disadvantage in detecting luminescence because only small amounts of DNA can be immobilized on the bead surfaces for sequencing. As luminescence intensity could be enhanced by increasing the luciferase density in the chambers, we overcame this difficulty by using a gel matrix to immobilize luciferase at a high concentration in the microreaction chambers. Luminescence 1 order of magnitude higher could be observed with the new method compared to the conventional method. Consequently, the chamber size and bead size immobilizing DNA could be reduced to as small as 6.5 and 4 μm, respectively. This can be successfully applied to achieving small, inexpensive, pyrosequencing systems with high throughput.
M
icro-DNA arrays are generally used for high-throughput gene expression profiling.1,2 Next-generation DNA sequencers have recently been used for more accurate analyses of gene expression by counting mRNA.35 Despite the recent rapid improvements in the throughput of DNA sequencers,6 there are still huge challenges in cost and throughput for the next-generation DNA sequencers to become standard tools for the analyses of gene expression. One of the most promising techniques is pyrosequencing from the aspects of long base sequencing, improving the overall throughput, and increasing the speed of analyses.79 Enzymatic reactions are used in pyrosequencing to produce luminescence from pyrophosphates (PPi) released as byproduct of DNA extension reactions. All cDNA should be sequenced independently in gene expression analyses by digitally counting mRNA with highthroughput DNA sequencers. Consequently, increasing sequencing throughput and lowering sequencing cost are key issues. To achieve a high throughput inexpensive pyrosequencer, an inexpensive detection device should be used with a high density array of numerous microreaction chambers containing DNA immobilized beads. Since the most sensitive system of optical detection can be attained with a 1 to 1 image magnification system according to the Lagrange invariant law,10 the size of the microreaction chambers should be reduced down to the size of photopixels in a commercially available and inexpensive image sensor. Unfortunately, the present sensitivity in pyrosequencing is not sufficient to decrease the diameter of DNA r 2011 American Chemical Society
immobilized beads down to the size of photopixels as small as a few micrometers. The use of small bead immobilizing DNA templates in small microreaction chambers may not yield sufficient photosignals for detection. Therefore, innovative improvements to sensitivity in pyrosequencing are necessary to achieve a system with an inexpensive detection device coupled with an array of small microreaction chambers. We overcame this difficulty by using a gel matrix to immobilize enzymes, especially luciferase, instead of packing beads to immobilize them in the microreaction chambers. The enzymes were homogeneously immobilized in a gel matrix at high density. The total number of enzymes could be drastically increased more than those with the conventional methods because the enzymes could be held three dimensionally instead of two dimensionally on the bead surfaces as in the conventional methods. We evaluated the detection sensitivity obtained with 22 and 4 μm beads to confirm the high performance of the newly developed gel-matrix method. Generally, 28 μm beads are used in a conventional pyrosequencer.9 Since the diameter of the 4 μm beads is 1/7th of the bead diameter used in conventional systems, the possible amount of immobilized cDNA on a bead surface is 1 order of magnitude lower than that in conventional systems. Therefore, the luminescence detected with the new system Received: May 29, 2011 Accepted: August 21, 2011 Published: August 21, 2011 7560
dx.doi.org/10.1021/ac201357t | Anal. Chem. 2011, 83, 7560–7565
Analytical Chemistry
TECHNICAL NOTE
Figure 1. Schematic of pyrosequencing system. A reaction plate was setup vertically to prevent the accumulation of air bubbles. Two pinch valves were used to switch the aspiration of reagents and putting accumulated reagents in the syringe.
should be 1 order of magnitude lower than that in conventional systems.
’ EXPERIMENTAL METHOD AND MATERIALS Principle Underlying Pyrosequencing. Pyrosequencing is based on the detection of PPi during stepwise synthesis of DNA by polymerase. Four enzymes were employed in the technique including DNA polymerase for incorporating nucleotides into DNA strands, ATP sulfurylase for catalyzing PPi to ATP, luciferase for generating light (signals) from ATP and luciferin, and apyrase for degrading ATP and excess nucleotides. Each reaction for incorporating nucleotides was carried out by injecting four nucleotide species in turn into a flow cell device having numerous microreaction chambers. Although the beads in the reaction chambers basically had different DNA templates on their surfaces, each bead had the same DNA template. Each DNA sequence could be determined by observing a bioluminescence pattern (pyrogram) corresponding to the DNA sequence on the bead. Experimental Apparatus. A schematic view of the pyrosequencing system we developed is shown in Figure 1. It consists of three parts (1) a flow cell device including a disposable reaction plate with microreaction chambers and DNA immobilized beads, (2) a flow system consisting of a rotary bulb, a syringe pump, and PEEK tubes to supply reagents for nucleotide incorporation reactions and for washing, and (3) an imaging system with an electron-multiplying charge-coupled device (EMCCD) to acquire the bioluminescent images in synchronization with the reagent supply. We used two types of disposable reaction plates with microreaction chambers of different sizes. The first one had about 100 000 of 21 pL microreaction chambers (31 μm deep and 35 μm in diameter) made of polyolefin (Hitachi Maxell Ltd. and Maxell Finetech Ltd.) coated with a 100 nm thick titanium film (Maxell Finetech Ltd.). The metal coating reduced the optical cross-talks between neighboring microreaction chambers and increased bioluminescence by reflecting the emitted luminescence in the opposite direction. The titanium coating was formed by arc ion plating at room temperature to prevent the plate from thermal deterioration. We also formed 1020 nm thick gold film on another plate by using an ion coater (SANYU Electron QUCIK COATER SC701). These plates with large microreaction chambers were mainly used for comparing the gel matrix and the conventional methods in sensitivity. Each
microreaction chamber contained a DNA immobilized bead of 22 μm in diameter. The beads were made of zirconia and coated with avidin (Hitachi Maxell Ltd.). They were heavy enough (6.0 g/cm3) to sink down in the chambers by gravity without a centrifuge. The effect of Ti or Au coating on the microreaction chambers was evaluated by observing luminescence produced by single base extension reactions. The second plate was made of polycarbonate and had 1 000 000 of 0.2 pL microreaction chambers (6.5 μm in depth and diameter) placed at a 13 μm pitch (Fujidenolo Co., Ltd.). Its surface was coated with 1020 nm thick Au. Each microreaction chamber had one small magnetic bead 4 μm in diameter coated with streptavidin (Micromod Partikeltechnologie GmbH) to immobilize DNA templates. The maximum number of DNA strands immobilized on a bead was as much as the number of streptavidins on the bead, which was about 2.0 105 for the 4 μm bead. Six reagent reservoirs for four buffers including dNTPs, washing buffer (WB) including Apyrase, and two types of cleaning buffer (CB1, 2) for washing out the buffer including apyrase were connected to the flow system that could deliver the reagents in turn into the microreaction chambers. The reagent flows could be switched from one to another with a rotary valve (C35Z-31820D VICI Valco Instruments Co. Inc.). The flow speed was controlled by operating a syringe pump with a stepping motor. All the reagents except for CB1 were stored at 4 °C during the experiments to prevent the decomposition of dNTPs, apyrase, and pyrophosphatase (ppase). We created an air gap of 100 μL between reagent injections to avoid contamination from the reagents used in the previous reaction step. The reagent injections flowed through the reaction plate in turn to enter the microreaction chambers by diffusion through the gel matrixes to carry out the nucleotide incorporation reactions followed by the cascade reactions for bioluminescence. The order of reagent flows was (1) the buffer including the first nucleotide species, (2) the washing buffer including apyrase to remove the buffer including dNTP and degrade the residual dNTP, (3) the cleaning buffer to remove apyrase used in the second process, and then (4) the buffer including the second nucleotide species and so on. The four types of dNTPs were injected in turn, which were followed by two additional injections of washing buffer and cleaning buffer. The washing buffer was composed of a 0.6 mL washing solution with 0.1 U/mL apyrase and the cleaning buffer composed of 1.8 mL cleaning solution 7561
dx.doi.org/10.1021/ac201357t |Anal. Chem. 2011, 83, 7560–7565
Analytical Chemistry
TECHNICAL NOTE
Figure 2. (a and b) Schematics of the intersectional view of microreaction chambers. Beads immobilizing enzymes were used as enzyme supporter in part a, and photoreactive polymer was used to immobilize enzymes in part b. (c) SEM image of the intersection of the reaction plate. The red rectangle corresponds to areas in parts a and b. (d) Reaction diagram for gelling and immobilization of enzymes.
(1.2 mL of CB1 (room temperature) and 0.6 ml of CB2 (4 °C)) to remove apyrase completely. After the microreaction chambers had been cleaned with these two solutions, 0.55 mL of solution containing one of the nucleotides (10 mM of dTTP, 10 mM of dGTP (deaza-dGTPdGTP = 3:1), 10 mM of dCTP, or 20 mM of dATPαS (GE Healthcare)) with 0.4 mM of luciferin, 2 μM of adenosine 50 -phosphosulfate (APS), and 10 mM of magnesium acetate (Sigma-Aldrich Corporation) were injected into the reaction chambers. Deaza-dGTP was used for improving the nucleotide incorporation of successive “G” by a relaxing compression of the structure of synthesized DNA. The reaction plate was maintained at 30 °C throughout the experiments. The EMCCD camera was cooled at 57 °C, and the exposure time was 20 s. A luminescence image of all the microreaction chambers was focused on a back-illuminated EMCCD camera with a lens system consisting of two tandemly aligned identical lenses (NA = 0.42, f = 50 mm, Nikon Corporation) to achieve the unity image magnification. The EMCCD camera (Andor Technology PLC) had a 13.3 mm square sensor chip with 1024 1024 pixels. The pitch of the photopixels was 13 μm which coincided with that of the microreaction chambers of 6.5 μm. A reaction plate having the microreaction chambers was placed so that a luminescence image from one microreaction chamber should be focused and positioned in one photopixel of the EMCCD. For the moment, we did not have a fine image positioning system and we could not make sure whether the luminescence image was focused in one photopixel or not. The defocusing of the luminescence image may result in unstable detection of signals during the experiments. This instability may not occur for the use of large reaction chambers because their images were detected with a plural of photopixels which enables one to focus the luminescence image on the detector perfectly. The image data were acquired and processed with Labview (National Instruments Corporation) based programs. The EMCCD chip size was the same as the microreaction chamber area in the reaction plate. Preparation of DNA Samples. The model target was biotinylated oligo-nucleotide (50 -biotin- TGTGCGCCGGTCTCTCCCAGGACAGGCACAAACACGCACCTCAAAGCTGTTC-30 ) immobilized on the beads. They were hybridized with sequence
primers (50 -GAACAGCTTTGAGGTGCGTGTT-30 ). Two types of beads, 4 μm in diameter (25 μg) and 22 μm in diameter (2.25 mg), were used. They were incubated in a 50 μL of preparation buffer with 0.1 U/mL (0.5 U/mL) apyrase and 0.2 U/μL exoKlenow fragment (Promega Corporation) at room temperature (RT) for 3 min to reduce background signals. The total quantity of 4 μm beads in a reaction plate was estimated to be 6.8 105; these beads were dispersed in the reaction plate followed by centrifuging so as to settle the individual bead efficiently down into each microreaction chamber. The number of zirconia beads (22 μm) was estimated to be 7.3 104, and they were settled down in microreaction chambers by gravity. Immobilization of Enzymes in Gel Matrix. The crosssectional views of microreaction chambers for the conventional method and the gel matrix method are shown in Figure 2a,b. The beads immobilizing enzymes were produced by mixing 50 μL of undiluted suspended bead solution (Dynabeads M280 Streptavidin Invitrogen), 6.6 μL of biotinylated luciferase (BLU-Y, Kikkoman Corp., 2.3 107 RLU), and 6.6 μL of biotinylated ATP sulfurylase (homemade, 0.022 U/μL) followed by incubating them for 30 min at RT. These beads were packed in the microreaction chambers by centrifuge (at 2000g for 5 min) after settling 22 μm beads immobilizing DNA in the chambers. It was necessary to fill the microreaction chambers with a highdensity enzyme solution to make a gel matrix. It was done by mixing 20 μL of an enzyme solution including 2000 GLU Luciferase (Luc-H, Kikkoman Corp.), 40 U/mL of ATP sulfurylate (M0394L, New England Biolabs, Inc.), and 50 mU/mL of apyrase (Sigma-Aldrich Corp.) with 10 μL of AWP (6 wt %, Toyo Gosei Co., Ltd.) followed by dispensing it into the microreaction chambers. The excess solution was removed from the microreaction chambers by spinning them out at 6500 rpm. As the gel solution contained photopolymers, enzymes could be immobilized in the gel matrix by making cross-links in the polymer by irradiating with UV light (302 nm, 2 mW/cm2, and UVP) for 1 min. The reaction plate was filled with a cleaning solution and then incubated to make the gel swell in the microreaction chambers. The enzymes could be immobilized uniformly in the gel matrix as outlined in Figure 2b. Figure 2d is a reaction diagram in which active nitrene is generated from azide 7562
dx.doi.org/10.1021/ac201357t |Anal. Chem. 2011, 83, 7560–7565
Analytical Chemistry
TECHNICAL NOTE
Figure 3. (a) Comparison of bioluminescence from microreaction chambers with and without metal coatings. (b) Bioluminescence image acquired associated with single base extension.
by UV irradiation to react with other nitrenes on polyvinyl chains or amine on the enzymes. All the substrates for the reactions to incorporate nucleotides could migrate into the microreaction chambers through the gel matrix within 1 min. The apyrase in the gel matrix is used to degrade the excess dNTPs which should be removed from the chambers at a level of less than 1% of the original concentration. Enhancement of Luminescence by Metal Coating on the Chamber Surfaces. Coating of the microreaction chamber surfaces with reflecting materials was effective to increase luminescence collected with an EMCCD. Actually bioluminescence was enhanced by four times as shown in Figure 3. The factors related to the signal enhancement might be the multiple reflections from and the refractions by the coatings and beads in the microreaction chambers. A typical example of a bioluminescence image associated with single base extensions is shown in Figure 3b. Measurement of Diffusion Constants. As dNTPs are supplied into the microreaction chambers by diffusion through gel matrixes, it is important to keep the diffusion time comparable to that in a free solution for completing nucleotide incorporation reactions and removing the residual reagents. Therefore, the diffusion constants of Cy5 labeled dCTP (GE Healthcare Life Science) in a buffer solution (1xC buffer) as well as in the gel matrix obtained from AWP with 060 s irradiation of UV light were evaluated by fluorescent correlation spectroscopy (FCS) with a laser scanning confocal microscope (Carl Zeiss LSM510). The solutions of 200 μL with 109 M Cy5-dCTP and 2% AWP in plastic chambers were irradiated with 302 nm UV (2 mW/cm2) for 0, 10, 20, 30, and 60 s to make the gels in different conditions to obtain the diffusion constants of Cy5-dCTP in solution as well as in the gels.
’ RESULTS AND DISCUSSION Improvements in Detection Sensitivity with Gel Matrix. As the signals obtained with the small 4 μm beads by the conventional method were too small for the comparative study with the gel matrix method, we used large zirconia beads (22 μm in diameter) as DNA supports to evaluate the luminescence enhanced by the gel matrix. After hybridization of the sequencing primers to the DNA templates on zirconia beads, the zirconia
Figure 4. Comparison of signal intensity and SNR with the conventional method.
beads were settled in microreaction chambers together with a cleaning solution. Then, 2.8 μm magnetic beads immobilizing enzymes (luciferase and ATP sulfurylase) were spun down into the microreaction chambers in the conventional method. The bioluminescence intensities obtained with these two methods are shown in Figure 4. The results were averaged over five measurements, each of which was obtained with a 50 50 pixel area in the acquired images associated with nucleotide extension reactions. The signal to noise ratios (SNRs), which are defined as signals divided by the standard deviations of background signals (signals without extensions), have been plotted with the round markers in the figure as well. As shown in the figure, the new gel-matrix method yielded about 40 times more luminescence than that obtained with the conventional method. This was due to the high density of immobilized luciferase which catalyzes the rate-limiting reaction in pyrosequencing.11 Pyrosequencing. The signal enhancement by the gel matrix enables the use of very small microreaction chambers together with small beads immobilizing DNA. Actually, this could reduce the bead size to 4 μm and the microreaction chamber size to 6.5 μm. The numbers of DNA molecules immobilized on a bead surface were about 510 105 for a 4 μm bead and 2 107 9 for a 28 μm bead, respectively. Although the use of small beads required highly sensitive detection technology, we accomplished this with the refraction coating on the microreaction chamber 7563
dx.doi.org/10.1021/ac201357t |Anal. Chem. 2011, 83, 7560–7565
Analytical Chemistry
TECHNICAL NOTE
Figure 5. Example of a pyrogram obtained with a bead which was modified with a signal decaying factor.
Figure 6. Relative diffusion constants of Cy5-labeled dCTP in gel matrix.
surfaces and the use of a gel matrix to support the enzymes. A pyrogram obtained with 4 μm beads that immobilized DNA is shown in Figure 5. It was normalized so as that the initial signal was 1. We estimated the amount of DNA on a 4 μm bead to be 510 105 molecules by measuring the bioluminescence emitted from DNA immobilized beads during single nucleotide extension reactions. The signals were measured with a small pyrosequencer.8 The estimated DNA density immobilized on a bead’s surface was 12 copies/100 nm2, which coincided with results from other experiments. The conventional method required more than 2 107 molecules per bead for sequencing,9 which was 2040 times greater than those used in our system. The estimated values here were consistent with the improvement of SNRs by more than 40 times. Pyrograms obtained with the gel matrix and the conventional methods for 28 μm beads immobilizing template DNA were similar although a few peaks due to negative and positive phase shifts which were caused by partially imperfect nucleotide incorporation reactions and by residual nucleotides in the next step of extensions were observed for the gel-matrix method. It might be overcome by improving the gel making process to give a more uniform pore size of the gel. A program obtained with small beads of 4 μm in a gel matrix is shown in Figure 5. Although the signal intensities were strong enough for sequencing, the intensities were instable as indicated in the figure. The reason
for that is not clarified yet. One possibility may be the defocus of the luminescence image on the detector and a small vibration during the experiments which may result in the change of luminescence intensities on the detector. We believe that the development of a positioning system to settle the reaction plate at the proper position and to reduce the possible vibration of the system may improve the stability of the signal and increase the readable DNA length. Diffusion Coefficient in Gel Matrix. A time-correlation fluorescence method was used for the evaluation. The diffusion coefficients of dNTPs in solutions without and without photoreactive polymer were measured with a laser excited confocal microscope. With the laser irradiation time for polymerization changed, the changes in the diffusion coefficients due to gel polymerization were obtained. The relative diffusion coefficient of Cy5-labeled dCTP monomer depended on the irradiation time as shown in Figure 6. The relative diffusion coefficients in gel matrix relative to that in a buffer solution were between 0.5 and 0.8 where the threshold time of UV irradiation for making the gel was about 30 s. As a result, we found that the gel matrix had no effect on the supply of dNTPs for pyrosequencing. We also estimated the absolute values of the diffusion coefficient (D) for Cy5-dCTP in the buffer solution. Since the number of fluorescent molecules (N) was estimated to be 10.5 ( 0.15 for 10 nM of Cy5 in the confocal area, the confocal volume (V = N/(NAc) was estimated to be 1.7 ( 0.025 fL with Avogadro number NA and Cy5 concentration c. The confocal volume V is expressed as π1.5ωr2 ωz (where ωr and ωz are the lateral and axial focus radii), and the structure parameter S = ωz/ωr was estimated to be 11.6 ( 2.6. As the diffusion time (τ = ωr2/(4D)) through the focal volume was estimated to be 69.0 ( 2.3, then D in the buffer solution was estimated to be 3.2 ( 0.33 1010 m2/s, so that in the gel, D = 1.62.6 1010 m2/s. This means that it takes less than 0.3 s for dNTPs to diffuse into or out of a microreaction chamber that is 6.5 μm in depth when the diffusion time can be derived from the equation of 0.5 = Erf(6.5 μm/(4Dt)) (Erf means error function).
’ CONCLUSIONS We have demonstrated a novel pyrosequencing method by utilizing enzymes supported in a gel matrix to increase bioluminescence intensity. The use of a gel matrix seems very promising to improve the detection sensitivity by tens of times, and this enables the pyrosequencing to use beads as small as 4 μm immobilizing template DNA in microreaction chambers as small as 6.5 μm. Although we should improve the positioning system of the reaction plate to improve the stability of signals in 7564
dx.doi.org/10.1021/ac201357t |Anal. Chem. 2011, 83, 7560–7565
Analytical Chemistry
TECHNICAL NOTE
pyrograms, this has encouraged us to use an inexpensive commercially available CCD detector coupled with a 1:1 magnification lens system to achieve an inexpensive massively parallel pyrosequencer.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT We wish to express our thanks to Professor Y. Ito at the RIKEN Nano Medical Engineering Laboratory for the productive discussions we had with him on photoreactive polymers. This research was supported by the Strategic International Cooperative Program of the Japan Science and Technology Agency. ’ REFERENCES (1) Fodor, S. P.; Rave, R. P.; Huang, X. C.; Pease, A. C.; Holmes, C. P.; Adams, C. L. Nature 1993, 364, 555. (2) Kurimoto, K.; Yabuta, Y.; Ohinata, Y.; Ono, Y.; Uno, K. D.; Yamada, R. G.; Ueda, H. R.; Saitou, M Nucleic Acids Res. 2006, 34, e42. (3) Kambara, H. Proceedings of The 1st International Workshop on Approaches to Single Cell Analysis, 2006. (4) Tang, F.; Barbaciou, C.; Wang, Y.; Nordman, E; Lee, C.; Xu, N.; Wang, X; Bodeau, J.; Tuch, B. B.; Siddiqui, A.; Lao, K.; Surani, A. Nat. Methods 2009, 6, 377. (5) Hashimoto, S.; Qu, W.; Ahsan, B.; Ogoshi, K.; Sasaki, A.; Nakatani, Y.; Lee, Y.; Ogawa, M.; Suzuki, Y.; Sugano, S.; Lee, C.; Nutter, R.; Morishita, S.; Matsushima, K. PloS One 2009, 4, e4108. (6) Wang, Z.; Gerstein, M.; Snyder, M. Nat. Rev. Genet. 2009, 10, 57. (7) Ronaghi, M.; Uhlen, M.; Nyren, P. Science 1998, 281, 363. (8) Zhou, G.; Kajiyama, T.; Gotou, M.; Kishimoto, A.; Suzuki, S.; Kambara, H. Anal. Chem. 2006, 78, 4482–4489. (9) Margulies, M.; Egholm, M.; Altman, W.; Attiya, S.; Bader, J. S.; Bemben, L. A.; Berka, J.; Braverman, M. S.; Chen, Y.-J.; Chen, Z.; Dewell, S. B.; Du, L.; Fierro, J. M.; Gomes, X. V.; Godwin, B.C.; He, W.; Helgesen, S.; Ho, C. H.; Irzyk, G. P.; Jando, S. C.; Alenquer, M. L. I.; Jarvie, T. P.; Jirage, K. B.; Kim, J.-B.; Knight, J. R.; Lanza, J. R.; Leamon, J. H.; Lefkowitz, S. M.; Lei, M.; Li, J.; Lohman, K. L.; Lu, H.; Makhijani, V. B.; McDade, K. E.; McKenna, M. P.; Myers, E. W.; Nickerson, E.; Nobile, J. R.; Plant, R.; Puc, B. P.; Ronan, M. T.; Roth, G. T.; Sarkis, G. J.; Simons, J. F.; Simpson, J. W.; Srinivasan, M.; Tartaro, K. R.; Tomasz, A.; Vogt, K. A.; Volkmer, G. A.; Wang, S. H.; Wang, Y.; Weiner, M. P.; Yu, P.; Begley, R. F.; Rothberg, J. M. Nature 2005, 437, 376. (10) Greivenkamp, J. E. Field Guide to Geometrical Optics ; SPIE Field Guides, SPIE: Bellingham, WA, 2004; Vol. FG01, p 28. (11) Agah, A.; Aghajan, M.; Mashayekhi, F.; Amini, S.; Davis, R.; Plummer, J.; Ronaghi., M.; Griffin, P. Nucleic Acids Res. 2004, 32 (21), e166.
7565
dx.doi.org/10.1021/ac201357t |Anal. Chem. 2011, 83, 7560–7565
