Microfabricated bioprocessor for integratednanoliter-scale Sanger DNA sequencingRobert G. Blazej*, Palani Kumaresan†, and Richard A. Mathies*‡§

*University of California, San Francisco�University of California, Berkeley, Joint Bioengineering Graduate Group, †Department of Mechanical Engineering,and ‡Department of Chemistry, University of California, Berkeley, CA 94720

Communicated by Ignacio Tinoco, Jr., University of California, Berkeley, CA, March 27, 2006 (received for review March 1, 2006)

An efficient, nanoliter-scale microfabricated bioprocessor integrat-ing all three Sanger sequencing steps, thermal cycling, samplepurification, and capillary electrophoresis, has been developed andevaluated. Hybrid glass–polydimethylsiloxane (PDMS) wafer-scaleconstruction is used to combine 250-nl reactors, affinity-capturepurification chambers, high-performance capillary electrophoresischannels, and pneumatic valves and pumps onto a single micro-fabricated device. Lab-on-a-chip-level integration enables com-plete Sanger sequencing from only 1 fmol of DNA template. Up to556 continuous bases were sequenced with 99% accuracy, dem-onstrating read lengths required for de novo sequencing of humanand other complex genomes. The performance of this miniaturizedDNA sequencer provides a benchmark for predicting the ultimatecost and efficiency limits of Sanger sequencing.

capillary electrophoresis � genetic analysis � microfluidic

The Sanger method of DNA sequencing is remarkable in itsability to generate long and accurate sequence reads. Alter-

native low-cost methodologies aimed at displacing Sanger se-quencing currently produce shorter and less accurate sequencereads, a drawback that could limit their usefulness in de novosequencing of complex genomes and in the detection of struc-tural variations. Comparative analysis of mammalian genomes isfocusing attention on the importance of structural variations(insertions, inversions, deletions, and duplications) in speciation,evolution, disease, and cancer (1–3). Consequently, new near-term technologies for low-cost genome sequencing and eventualultra-low-cost sequencing technologies (4) must be evaluated inthe context of their ability to generate accurate short-range aswell as long-range genome assemblies.

Emerging near-term technologies include two cyclic arraymethods that produce highly parallel, short-read data (5, 6).Shendure et al. (5) used DNA ligation to generate 26-base readsthat were aligned to 70.5% of the 3.3-Mb Escherichia coligenome with high confidence. Using a pyrophosphate-basedapproach, Margulies et al. (6) generated read-lengths of �110bases for sequencing and de novo assembly of 90.4% of the2.1-Mb Streptococcus pneumoniae genome. Significantly, it wasnecessary to exclude repetitive DNA sequences from the assem-bly. Such repeat sequences are vital to genome structure andregulation, given that over half of the human genome is com-posed of repetitive DNA (7); repetitive sequences are ubiquitousand functionally important in prokaryotic and eukaryotic ge-nomes (8), and gene duplications are evident in disease (9),speciation (2) and cancer (10). Longer read-lengths with en-hanced accuracy would increase sequence coverage and improvede novo genome assemblies, but extending cyclic sequencingreads is challenging because the process is serial, requiring alinearly increasing number of enzymatic steps. Necessarily,genomes sequenced thus far by using cyclic methods are morethan three orders of magnitude smaller than the human genome.

Because Sanger sequencing is the only demonstrated technol-ogy for complete sequencing of plant and animal genomes (7,11–13), it is, thus, critical to ask: What are the ultimate limits ofSanger sequencing and how can these limits be achieved? The

feasibility of performing completely integrated, nanoliter-scaleSanger sequencing from a femtomole of template has beendiscussed in ref. 14. A significant challenge in creating such asystem is the integration of diverse biochemical processes andhigh-performance electrophoretic separation onto a single mi-crodevice. Sequencing separations alone were demonstrated ona single-channel glass microdevice in 1995 and subsequentlyscaled to 96 and 384 channels (15–17). However, in each case,sample production and purification was performed by usingconventional microliter-scale processing that generates a 1,000-fold more product than is needed for analysis (18). Because asignificant cost in Sanger sequencing is associated with reagentconsumption (19) as well as equipment expense and labor,conversion to an integrated lab-on-a-chip technology (20–23)could revolutionize Sanger sequencing by efficiently linkingminiaturized analysis to ultra-low-volume sample generation andpurification.

Inspiration for solving this process integration challenge isprovided by the sample analysis systems developed recently forpathogen and extraterrestrial life detection (24, 25), which usea hybrid glass–polydimethylsiloxane (PDMS) microdevice con-struction (26, 27). Incorporation of a microvalve-formingPDMS membrane into a multilayer glass wafer structureenables a level of complexity, integration, performance, andfunctionality previously unattainable in microfabricated sys-tems. Here, we explore the practical limits of Sanger sequenc-ing by using this hybrid glass–PDMS microdevice structure tocreate an efficient, nanoliter-volume bioprocessor capable ofread-lengths equivalent to those used in the initial sequencingof the human genome (7).

ResultsThe bilaterally symmetric bioprocessor shown in Fig. 1A containstwo independent DNA sequencing systems. Device elements aremicrofabricated on three 100-mm-diameter glass wafers andassembled with a featureless PDMS membrane. The top twoglass wafers are thermally bonded to form enclosed all-glassreaction chambers and capillary electrophoresis (CE) channels.Contact bonding between the fused top wafers, PDMS mem-brane, and bottom manifold wafer forms three-dimensionalinterconnects and microvalves (Fig. 1B). The microvalve struc-tures comprise a discontinuous fluidic channel on the bottomside of the resistive temperature detector (RTD)�valve wafer,the PDMS membrane, and a displacement chamber on the topside of the manifold wafer (26). Vacuum applied to the manifoldlines deflects the PDMS membrane into the displacement cham-ber, creating a fluidic connection.

Conflict of interest statement: R.A.M. is a consultant with Microchip Biotechnologies Inc.(MBI) and has a financial interest in MBI. MBI is working on the commercialization ofmicrochip sequencing technologies and may benefit from the results of this research.

Abbreviations: CE, capillary electrophoresis; PDMS, polydimethylsiloxane; RTD, resistivetemperature detector.

§To whom correspondence should be addressed at: Department of Chemistry, MS 1460,University of California, Berkeley, CA 94720. E-mail: rich@zinc.cchem.berkeley.edu.

© 2006 by The National Academy of Sciences of the USA

7240–7245 � PNAS � May 9, 2006 � vol. 103 � no. 19 www.pnas.org�cgi�doi�10.1073�pnas.0602476103

Dow

nloa

ded

by g

uest

on

Apr

il 20

, 202

0

Fig. 2A shows one half of the bioprocessor filled with distinctdyes to indicate reagent locations during operation. Each DNAsequencing system comprises three functional units correspond-ing to the three steps in Sanger sequencing. The thermal cyclingunit consists of a 250-nl reactor with RTDs (Fig. 2B), microv-alves (Fig. 2C), and a surface heater. Via holes (Fig. 2D) allowreagent to move between the enclosed all-glass layer and theglass–PDMS microvalve layer. Three microvalves for each reac-tor perform the dual role of sealing the chamber during thermalcycling and forming a self-priming diaphragm pump system forpostprocessing. Thermal-cycled sequencing reagent is processedby the capture�purification unit (Fig. 2E Right) before injection(Fig. 2E Left) into the CE separation unit. Tapered turns (Fig.2F) fold the 30-cm capillary onto the compact device whileminimizing turn-induced dispersion (28).

To demonstrate bioprocessor functionality, we performedintegrated sequencing on 1 fmol of a 750-bp pUC18 amplicon.

In the sequencing process, energy-transfer dye-terminatorsequencing reaction reagent containing DNA template and�47 M13 sequencing primer is rapidly thermal-cycled in the250-nl reaction chamber (35 cycles of 95°C for 12 s and 60°Cfor 55 s). The mixture is then pumped by the microvalves intothe upper capture�purification chamber (Fig. 3A). A 33-V�cmelectric field applied between the capture outlet and inlet portsdrives the charged reaction mixture into the capture gel (Fig.3B). The gel comprises a sparsely cross-linked polyacrylamidematrix containing 40 �M covalently bound oligonucleotidethat is complementary to the DNA sequence immediately 3� ofthe sequencing primer (29). Sanger extension fragments con-taining 12 or more complementary bases selectively hybridizeat 30°C and concentrate at the gel�buffer interface, whereasresidual salts, primer, template, and unincorporated nucleo-tides that would interfere with sample analysis pass through tothe capture inlet port (Fig. 3C).

After capture is complete, the microvalve pump system isturned off, and the electric potential is maintained to allow all

Fig. 1. Integrated nanoliter-scale nucleic acid bioprocessor for Sanger DNAsequencing. (A) Top view of the assembled bioprocessor containing two setsof thermal cycling reactors, purification�concentration chambers, CE channels(black), RTDs (red), microvalves�pumps (green), pneumatic manifold channels(blue), and surface heaters (orange). (B) Expanded view, showing microdevicelayers. Rim colors indicate the surface on which the respective features arefabricated. The top two glass wafers are thermally bonded and then assem-bled with a featureless PDMS membrane and manifold wafer.

Fig. 2. Bioprocessor components. (A) Photograph of the microdevice, show-ing one of two complete nucleic acid processing systems. Colors indicate thelocation of sequencing reagent (green), capture gel (yellow), separation gel(red), and pneumatic channels (blue). (Scale bar, 5 mm.) B–F correspond to thefollowing component microphotographs. (B) A 250-nl thermal cycling reactorwith RTDs. (Scale bar, 1 mm.) (C) A 5-nl displacement volume microvalve. (D)A 500-�m-diameter via hole. (E) Capture chamber and cross injector. (F) A65-�m-wide tapered turn. (Scale bars, 300 �m.) All features are etched to adepth of 30 �m.

Blazej et al. PNAS � May 9, 2006 � vol. 103 � no. 19 � 7241

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Apr

il 20

, 202

0

residual reagents to electrophorese out of the capture chamber(Fig. 3D). Fluorescence quantitation indicates that 25 fmol ofextension fragments are retained in a �1.2-nl band of capture gelat �20 �M, a concentration factor of 200�. Given a startingquantity of 1 fmol of template in a 35-cycle linear amplificationreaction, total efficiency for extension fragment production andpurification is 71%. This efficiency is consistent with observedPCR efficiencies and is dictated primarily by the sequencingreaction because capture is quantitative.

Extension fragment affinity for the capture gel is reversed byincreasing the temperature beyond the melting point of theduplex (Tm � 54.7°C). Heating the microdevice to �67°Ccompletely releases the hybridized extension fragments andraises the separation capillary to an optimal temperature forDNA sequencing. A �200-�m-long plug of concentrated andpurified extension fragments is cross-injected onto the separa-tion capillary (Fig. 3E). Sample separation is initiated by apply-ing a 167-V�cm field between the cathode and anode (Fig. 3F),and small reverse potentials at the waste and capture inlet portsprevent uninjected material from leaking into the separationcapillary (see Materials and Methods). Approximately 3.6 fmol ofproduct is injected, as shown in Fig. 3F, giving an injectionefficiency of 15% and a total sample-delivery efficiency of 10%for 1 fmol of template.

Four-color sequence data were collected on a radial scanner(18) in �35 min and processed with the CIMARRON 3.12 base

caller and PHRED 0.020425.C to generate base calls with accuracyestimates (30). Although PHRED quality scores are calibrated forcommercial sequencing instruments, they were in good agree-ment with observed accuracies in this study and are useful forrelative evaluation of run conditions and other prototype in-struments. Fig. 4 presents high-quality sequence data generatedon the integrated bioprocessor. Primer and capture sequencesare internal to the 750-bp amplicon, resulting in a maximumpossible read length of 564 bases. Comigrating unincorporatedfluorescently labeled terminators are absent in the sequencingtrace, indicating their complete removal by the capture�purification process. Using a 99% accuracy cutoff with theknown pUC18 sequence, CIMARRON and PHRED base callsproduced a read length of 556 bases.

A range of injection potentials, electrophoretic separationfields, separation temperatures, and separation gels were inves-tigated in 18 separate experiments to optimize run parameters(see Materials and Methods). The average 99% accuracy PHRED-generated read length for all runs was 427 bases (� � 85).Consistent increases in read length expected with increasedtemperature or decreased separation field strength (31) were notobserved, indicating an additional, overriding factor affectingread length. Inexact positioning of the capture band relative tothe cross-injector with concomitant changes in injection timingis believed to be a significant source of this read-length vari-ability. Our best sample-injection conditions generated five runsof �500 bases (� � 540, � � 17), with separation potentialsranging from 150 to 167 V�cm and temperatures ranging from70°C to 73°C.

Fig. 5 graphs raw PHRED quality scores for each base (gray)and predicted cumulative read accuracy (black) for the datapresented in Fig. 4. Base call accuracy rises rapidly after the first25 bases to a predicted 1 in 1,000 error rate. Poorly resolvedsequential Cs early in the run are correctly assigned low scores,causing dips in quality. Between 100 and 450 bases, error ratesaverage 1 in 100,000. Accuracy drops at base 477, where PHREDis unable to make a base call in a typically problematic ho-mopolymeric domain. Subsequent calls are accurate until base541, after which point, signal strength drops substantially, mak-ing further base calling increasingly difficult. Aggregate PHREDerror rates correctly estimate a read length of 556 bases with anaccuracy of 99%.

DiscussionThe DNA sequencing bioprocessor presented here incorporatesa range of advanced lab-on-a-chip technologies, including min-iaturized temperature sensing, nanoliter-scale Sanger extensionreactions, microvalves�pumps, DNA affinity-capture, and high-performance CE. Several innovations were necessary duringdevelopment to integrate the disparate elements of the Sangersequencing process into a single microdevice. The use of com-bined electrophoretic, pneumatic, and thermal sample controlenabled efficient manipulation of minute reagent quantities.Earlier design iterations that relied solely on electrophoreticcontrol were inefficient because of slow movement of sequencingsample in weak-field regions present in interlayer via holes.Transitioning from a monolithic substrate to a hybrid glass–PDMS assembly was critical. Multilayer construction enablesgreater design complexity and allows crossing of fluidic andpneumatic lines (25), a capability that is important for scalingfeature density and parallel processing. At the same time, thedurable hybrid construction preserves the all-glass structuresnecessary for single-base-resolution DNA separation. A singlemicrodevice has been used for �40 complete runs without anyindication of microvalve or CE channel degradation.

Development of optimized separation and capture gels andimproved sample injection methods will be central in furtherimproving read length and reproducibility. Previous work with

Fig. 3. False-color fluorescence images of the capture, purification, andinjection steps. Cooler colors indicate lower intensity; warmer colors indicatehigher intensity. Fluidic channels are outlined in white. Relative electricpotentials are indicated by � and �. (Scale bars, 300 �m.) (A) Thermally cycleddye-terminator sequencing reagent is pumped into the buffer-filled upperchamber (a). Simultaneously, an electric field drives the sequencing mixtureinto the lower, capture-gel-filled chamber. (B) Fluorescently labeled DNAextension fragments selectively hybridize and concentrate at the gel�bufferinterface (b). (C) Desired extension fragments (c) are bound to the capture gelinterface, and contaminants (d) continue to electrophorese out of the gel. (D)Purified and concentrated sample is ready for injection. (E) Extension frag-ments are released from the capture gel at 70°C and injected into the channeljunction (e). Pinching potentials constrain the plug size. (F) A �1-nl sampleplug is injected onto the separation capillary at (f), and reverse potentials pullaway uninjected material (g).

7242 � www.pnas.org�cgi�doi�10.1073�pnas.0602476103 Blazej et al.

Dow

nloa

ded

by g

uest

on

Apr

il 20

, 202

0

the affinity-capture gel system used 2 �l of off-chip-preparedsequencing sample (29). Scaling the sequencing reaction down to250 nl eliminated sample excess, making the injection timingmore sensitive to thermal gel expansion and electrophoreticmovement of the charged capture gel. Sparsely cross-linking thecapture gel and reducing the capture temperature amelioratedthese effects but resulted in increased purification�concentra-

tion times. As expected, the resolution of commonly useddye-terminator samples proved more challenging than idealizeddye-primer samples typically tested in prototype systems. Exten-sion of the separation channel from 16 to 30 cm increased theresolving power of the system, resulting in error rates of �1 ina million between 100 and 300 bases. Appropriately tunedseparation gels should extend the �99% accuracy range byuniformly resolving peaks over greater amplicon lengths (32).

The core DNA sequencing engine demonstrated here estab-lishes a performance benchmark from which to evaluate thefeasibility of miniaturizing high-throughput Sanger sequencing.Practical limits to reaction volume reduction are determined bythe sensitivity of the detection system and the efficiency ofsample processing. In a capillary array scanning system, 1 millionfluorophores per band significantly exceed the limit of detection(18) and could be reduced 10-fold with the improved scanningand detection technologies now available. By comparison, in thisstudy, each band contained �4 million fluorophores. A reactioncontaining 1 fmol of template generates �26 times more productthan is needed for detection (6 � 108 template molecules � 35cycles � 71% efficiency�564 bands � 26 � 106 molecules perband). More efficient injection techniques under investigationwould eliminate the wasteful cross-injection methods used in thisstudy, reducing template requirements 10-fold to 100 attomoles.Combined with scanner sensitivity improvements, the ultimateminimal template quantity extrapolates to �10 attomoles.

If we conservatively adopt a 10-fold reduction in template from1 fmol to 100 attomoles, a sequencing reaction performed at

Fig. 4. High-quality sequence data generated on the integrated bioprocessor. Sanger sequencing extension fragments from a 750-bp pUC18 PCR amplicon areresolved at 73°C, 167 V�cm with linear polyacrylamide separation gel (3.5% linear polyacrylamide�5% DMSO�3 M urea�1� TTE buffer) in 34 min. Automatic basecalls by the program PHRED and base numbers are indicated above the electropherogram. Elapsed time in minutes from injection is indicated parenthetically every50 bases. The base position where no call could be made is assigned ‘‘N.’’ Compared with the known pUC18 sequence, total read length is 556 bases with anaccuracy of 99%. Black arrows indicate base call errors; correct calls are given below. X indicates an inserted base.

Fig. 5. Base-call accuracies and sequence read length as predicted by PHRED.Percent accuracy is related to the PHRED quality score by: 100 � (1 � Pe), wherePe, the probability that the base call is incorrect, is equal to 1�10Q/10. Aone-in-a-hundred error rate is indicated by the dashed line. The gray line plotsPHRED quality scores at each base position. The black line charts predicted readaccuracy at each base position: 100 � (Basei � Pei)�Basei.

Blazej et al. PNAS � May 9, 2006 � vol. 103 � no. 19 � 7243

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Apr

il 20

, 202

0

standard concentrations in an easily fabricated 25-nl reactor rep-resents a 400-fold reduction in current sequencing reagent con-sumption [Department of Energy Joint Genome Institute (2004)www.jgi.doe.gov�sequencing�protocols�prots�production.html]with 800-fold less DNA template. Such ultra-low-volume�templatereactions should make direct Sanger sequencing of clonal PCRamplicons (33) on 5-�m-diameter controlled pore glass microbeadspossible. To achieve this ultimate limit of Sanger sequencing, we areworking toward a Microbead Integrated DNA Sequencer(MINDS) that parses PCR-colony beads into discrete thermalcycling chambers coupled to purification and electrophoretic sep-aration to produce a fully integrated genome-center-on-a-chip.

In addition to DNA sequencing, the bioprocessor functionalunits are capable of performing three fundamental biomedicaland biological research operations: DNA amplification, nucleicacid concentration�purification, and DNA fragment-size deter-mination. Consequently, the bioprocessor is suitable for a rangeof applications including genotyping, microsatellite analysis,single nucleotide polymorphism screening, and loss of heterozy-gosity detection. Integrated, low-volume analysis could be vitalin a wide range of applications, such as tumor biopsy screening,forensic analysis, or pathogen detection, where sample is limitedand error-free operation is essential.

Sanger sequencing has undergone a remarkable evolutionsince its radioactive slab-gel beginnings �30 years ago. Theprevalence and longevity of this core technique results from itsunique ability to generate both long reads and highly accuratebase calls. It is essential that we drive the cost, efficiency, andscalability limits of this robust sequencing technology to itsultimate limit. Based on our results, template�reagent require-ments can be reduced an additional 100-fold, and a fully inte-grated microfluidic genome sequencing system should also leadto significant infrastructure and labor savings. The bioprocessordemonstrated here is a key step toward the development of abench-top platform capable of de novo sequencing of plant andanimal genomes. The ability to determine long-range genomestructure and not simply local sequence uniqueness is critical incomplex genome sequencing and is likely to be central in fullyunderstanding the subtle effects of genome organization ondevelopment, speciation, and disease.

Materials and MethodsFabrication and Device Control. The bioprocessor is constructedfrom three patterned 100-mm-diameter glass wafers (D263;Schott, Duryea, PA) and a featureless 254-�m-thick PDMSmembrane (Bisco, Schaumberg, IL). Patterns are photolitho-graphically transferred to the glass wafers and etched to a depthof 30 �m in 49% HF by using a 2,000-Å-thick amorphous siliconhard mask (34). Holes are drilled for fluidic, electric, pneumatic,and interlayer access. A 1.1-mm-thick wafer containing channeland chamber features etched on the bottom side forms the topof the bioprocessor. Four-point RTDs are fabricated on the topside of a 550-�m-thick wafer coated with 200 Å�2,000 Å Ti�Pt.The bottom side of this wafer is patterned with fluidic lines andmicrovalve seats in register with the RTDs by using backsidealignment lithography (KS Aligner; Suss MicroTec, Santa Clara,CA). The two wafers are thermally bonded in a vacuum furnaceat 570°C for 6 h, forming all-glass enclosed features. The bottomside of the bonded stack is UV-ozone-cleaned for 3 min. Similartreatment is performed on the top surface of the PDMS mem-brane, and the two are assembled immediately. The final layerof the device is a 550-�m-thick wafer containing pneumaticchannels and microvalve displacement chambers (26). The waferand PDMS membrane are UV-ozone-cleaned and assembled asbefore. The fully assembled device is left overnight under a 2-kgweight, ensuring a permanent bond between the glass layers andthe PDMS membrane. Access holes are punched in the PDMS

membrane and 1�16-inch barbed pneumatic ports are epoxied tothe top surface.

Thermal cycling and microvalve actuation are controlled by aLABVIEW graphical interface (National Instruments). A 10-mm-diameter surface heater (7.8 ; Minco, Minneapolis) is mountedunderneath the thermal cycling reactor. RTDs are contacted byusing electrical pins, and 1 mA is passed through the outer leads.The voltage drop across the inner leads as a function oftemperature is measured in real-time by using a signal condi-tioner (5B31-01; Analog Devices) and a data acquisition board(National Instruments). The output drives a proportional inte-gral derivative (PID) controller that regulates 0–8 V for heatingand pressurized air for active cooling. Microvalves are closed oropened by applying 5 psi (1 psi � 6.89 kPa) pressure or vacuum,respectively, to the pneumatic ports.

Bioprocessor Preparation. Before operation, the channel andchamber walls are derivatized after a modified Hjerten coating(35). The system is primed by first filling capture gel [5% linearpolyacrylamide�1 � 10�4% bis-acrylamide�40 �M methacry-late-modified DNA�5�-Acrydite-ACTGGCCGTCGTTTTA-CAA-3� (Operon Technologies)�1� TTE buffer] to the point ofchamber constriction through the capture inlet port. The oligo-nucleotide sequence is complementary to the 19 bases immedi-ately 3� of the sequencing primer. Separation gel [in-house-prepared matrix (3.5% linear polyacrylamide�5% DMSO�3 Murea�1� TTE buffer), MegaBACE Long Read Matrix (Amer-sham Pharmacia), or CEQ Separation Gel LPA-1 (Beckman)],is pushed into the CE unit through the anode port at 600 psi byusing a high-pressure loader. Excess gel f lows out the waste andcathode ports as well as into the top portion of the capturechamber. Water flushed through the sample inlet port and outthe capture outlet port at 73°C removes excess separation gelfrom the capture chamber. Anode, cathode, waste, capture inlet,and capture outlet ports are filled with 5 �l of 1� TTE buffer.

DNA Sequencing. Energy-transfer dye-terminator sequencingreagent (1� DYEnamic ET mix (Amersham Pharmacia)�250-nM �47 M13 forward sequencing primer (5�-CGC-CAGGGTTTTCCCAGTCACGAC-3�)�4 nM 750-bp pUC18amplicon�250 ng/�l BSA) is loaded into the 250-nl thermalcycling chamber through the sample port. The chamber issealed by applying 5 psi pressure to the three pneumatic portsand rapidly thermal cycled for 35 cycles of 95°C (12 s) and 60°C(55 s). Three serial valves used for input, displacement, andoutput form a diaphragm pump system that is cycled as follows:open input and close output valves (4 s), open displacementvalve (200 ms), close input valve (200 ms), open output valve(200 ms), close displacement valve (200 ms). As the reactionmixture enters the upper portion of the capture chamber, acontinuously applied 33-V�cm electric field between the cap-ture inlet and outlet ports drives charged extension fragments,template DNA, excess primers, unreacted nucleotides, andsalts into the lower, capture-gel-filled chamber held at 30°C.After capture is completed in 20 min, the pumping is turnedoff, and the electric field is maintained for an additional 5 minto wash the captured products free of residual reagents.Extension fragments are released from the capture gel at67–75°C and electrokinetically injected for 3–4 s by applying750 V to the waste port while maintaining the cathode, anode,and capture inlet ports at 650, 250, and 450 V, respectively.Fragment separation occurs at 125–167 V�cm, and 75- and100-V pull-back potentials at the waste and capture inlet,respectively, prevent leakage of uninjected product. Bands aredetected by laser-induced f luorescence on the Berkeley con-focal rotary scanner (18).

7244 � www.pnas.org�cgi�doi�10.1073�pnas.0602476103 Blazej et al.

Dow

nloa

ded

by g

uest

on

Apr

il 20

, 202

0

Sequence Analysis. Raw electropherogram data were processedwith the CIMARRON 3.12 base-caller (NNIM, Sandy, UT). Addi-tional data processing to obtain base-call accuracy estimates wasperformed with PHRED 0.020425.C (30). The ‘‘ET Terminators’’chemistry profile was specified, causing PHRED to derive errorprobabilities from an internal lookup table calibrated for theMolecular Dynamics MegaBACE instrument. PHRED qualityscores relate to the base-call error probability as �10 �log10(Pe), where Pe is the probability that the base call isincorrect. Percent accuracy plots were generated by extrapolat-ing Pe from the PHRED quality scores and graphing: (Basei ��Pei)�Basei vs. Basei. Base calls were aligned to the knownpUC18 sequence by using the Needleman–Wunsch algorithm

(�3 for a correct match reward, �1 for a mismatch penalty, �3for a gap initiation penalty, and �2 for a gap extension penalty).Read lengths were determined by finding the maximum numberof aligned bases with 99% identity.

We acknowledge discussions with Brian Paegel on sample purification;with William Grover and Alison Skelley on microvalve design; and withCharles Emrich, Teris Liu, and Nicholas Toriello on microdevice fab-rication. Microfabrication was carried out at the University of California,Berkeley, Microfabrication Laboratory. This work was supported byNational Institutes of Health (NIH) Grants HG01399 and HG003583 viaMicrochip Biotechnologies, Inc. R.G.B. was supported by NIH TraineeFellowship HG00047 from the Berkeley Program in Genomics.

1. Zhao, S. Y., Shetty, J., Hou, L. H., Delcher, A., Zhu, B. L., Osoegawa, K., deJong, P., Nierman, W. C., Strausberg, R. L. & Fraser, C. M. (2004) Genome Res.14, 1851–1860.

2. Cheng, Z., Ventura, M., She, X. W., Khaitovich, P., Graves, T., Osoegawa, K.,Church, D., DeJong, P., Wilson, R. K., Paabo, S., et al. (2005) Nature 437,88–93.

3. Check, E. (2005) Nature 437, 1084–1086.4. National Institutes of Health (February 12, 2004) Near-Term Technology

Development for Genome Sequencing (Natl. Inst. of Health, Bethesda), RFA-HG-04-002.

5. Shendure, J., Porreca, G. J., Reppas, N. B., Lin, X. X., McCutcheon, J. P.,Rosenbaum, A. M., Wang, M. D., Zhang, K., Mitra, R. D. & Church, G. M.(2005) Science 309, 1728–1732.

6. Margulies, M., Egholm, M., Altman, W. E., Attiya, S., Bader, J. S., Bemben,L. A., Berka, J., Braverman, M. S., Chen, Y. J., Chen, Z. T., et al. (2005) Nature437, 376–380.

7. Lander, E. S., Linton, L. M., Birren, B., Nusbaum, C., Zody, M. C., Baldwin,J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al. (2001) Nature 409,860–921.

8. Shapiro, J. A. & von Sternberg, R. (2005) Biol. Rev. 80, 227–250.9. Gonzalez, E., Kulkarni, H., Bolivar, H., Mangano, A., Sanchez, R., Catano, G.,

Nibbs, R. J., Freedman, B. I., Quinones, M. P., Bamshad, M. J., et al. (2005)Science 307, 1434–1440.

10. Lengauer, C., Kinzler, K. W. & Vogelstein, B. (1998) Nature 396, 643–649.11. C. elegans Sequencing Consortium (1998) Science 282, 2012–2018.12. Arabidopsis Genome Initiative (2000) Nature 408, 796–815.13. Adams, M. D., Celniker, S. E., Holt, R. A., Evans, C. A., Gocayne, J. D.,

Amanatides, P. G., Scherer, S. E., Li, P. W., Hoskins, R. A., Galle, R. F., et al.(2000) Science 287, 2185–2195.

14. Paegel, B. M., Blazej, R. G. & Mathies, R. A. (2003) Curr. Opin. Biotechnol.14, 42–50.

15. Woolley, A. T. & Mathies, R. A. (1995) Anal. Chem. 67, 3676–3680.16. Paegel, B. M., Emrich, C. A., Weyemayer, G. J., Scherer, J. R. & Mathies, R. A.

(2002) Proc. Natl. Acad. Sci. USA 99, 574–579.17. Aborn, J. H., El-Difrawy, S. A., Novotny, M., Gismondi, E. A., Lam, R.,

Matsudaira, P., McKenna, B. K., O’Neil, T., Streechon, P. & Ehrlich, D. J.(2005) Lab Chip 5, 669–674.

18. Shi, Y. N., Simpson, P. C., Scherer, J. R., Wexler, D., Skibola, C., Smith, M. T.& Mathies, R. A. (1999) Anal. Chem. 71, 5354–5361.

19. Nakane, J., Broemeling, D., Donaldson, R., Marziali, A., Willis, T. D., O’Keefe,M. & Davis, R. W. (2001) Genome Res. 11, 441–447.

20. Dittrich, P. S. & Manz, A. (2005) Anal. Bioanal. Chem. 382, 1771–1782.21. Gardeniers, J. G. E. & van den Berg, A. (2004) Anal. Bioanal. Chem. 378,

1700–1703.22. Hansen, C. & Quake, S. R. (2003) Curr. Opin. Struct. Biol. 13, 538–544.23. Krishnan, M., Namasivayam, V., Lin, R. S., Pal, R. & Burns, M. A. (2001) Curr.

Opin. Biotechnol. 12, 92–98.24. Lagally, E. T., Scherer, J. R., Blazej, R. G., Toriello, N. M., Diep, B. A.,

Ramchandani, M., Sensabaugh, G. F., Riley, L. W. & Mathies, R. A. (2004)Anal. Chem. 76, 3162–3170.

25. Skelley, A. M., Scherer, J. R., Aubrey, A. D., Grover, W. H., Ivester, R. H. C.,Ehrenfreund, P., Grunthaner, F. J., Bada, J. L. & Mathies, R. A. (2005) Proc.Natl. Acad. Sci. USA 102, 1041–1046.

26. Grover, W. H., Skelley, A. M., Liu, C. N., Lagally, E. T. & Mathies, R. A. (2003)Sens. Actuator B-Chem. 89, 315–323.

27. Karlinsey, J. M., Monahan, J., Marchiarullo, D. J., Ferrance, J. P. & Landers,J. P. (2005) Anal. Chem. 77, 3637–3643.

28. Paegel, B. M., Hutt, L. D., Simpson, P. C. & Mathies, R. A. (2000) Anal. Chem.72, 3030–3037.

29. Paegel, B. M., Yeung, S. H. I. & Mathies, R. A. (2002) Anal. Chem. 74,5092–5098.

30. Ewing, B., Hillier, L., Wendl, M. C. & Green, P. (1998) Genome Res. 8, 175–185.31. Kotler, L., He, H., Miller, A. W. & Karger, B. L. (2002) Electrophoresis 23,

3062–3070.32. Doherty, E. A. S., Kan, C. W., Paegel, B. M., Yeung, S. H. I., Cao, S. T.,

Mathies, R. A. & Barron, A. E. (2004) Anal. Chem. 76, 5249–5256.33. Dressman, D., Yan, H., Traverso, G., Kinzler, K. W. & Vogelstein, B. (2003)

Proc. Natl. Acad. Sci. USA 100, 8817–8822.34. Simpson, P. C., Roach, D., Woolley, A. T., Thorsen, T., Johnston, R.,

Sensabaugh, G. F. & Mathies, R. A. (1998) Proc. Natl. Acad. Sci. USA 95,2256–2261.

35. Hjerten, S. (1985) J. Chromatogr. 347, 191–198.

Blazej et al. PNAS � May 9, 2006 � vol. 103 � no. 19 � 7245

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Apr

il 20

, 202

0