A multiple-capillary electrophoresis system for small-scale DNA sequencing and analysisJianzhong Zhang, Karl O. Voss, Diana F. Shaw, K. Pieter Roos, Darren F. Lewis, Juying Yan,Rong Jiang, Hongji Ren, Joan Y. Hou, Yu Fang, Xiaoling Puyang, Hossein Ahmadzadeh andNorman J. Dovichi*
Department of Chemistry, University of Alberta, Edmonton, Alberta T6G 2G2, Canada
Received May 20, 1999; Revised September 20, 1999; Accepted October 10, 1999
ABSTRACT
A five-capillary system has been developed for DNAsequencing and analysis. The post-column fluores-cence detector is based on a sheath-flow cuvette.The instrument provides uniform and continuousillumination of the samples. The cuvette virtuallyeliminates cross-talk in the fluorescence signalbetween capillaries. Discrete single-photon countingavalanche photodiodes provide high efficiency lightdetection. The instrument has detection limits (3 σσσσ) of130 ±±±± 30 fluorescein molecules injected onto eachcapillary. Over 650 bases of sequence at 98.8% accu-racy were generated in 100 min at 50 °°°°C fromM13mp18. Separation and detection of short tandemrepeats proved efficient and accurate with the use ofinternal standards for direct comparison of migrationtimes between capillaries.
INTRODUCTION
Large-scale DNA sequencing projects require instruments thatgenerate high throughput and high sequencing accuracy at lowcost (1). Capillary electrophoresis provides low-cost, easilyautomated and rapid DNA sequencing (2–15). The firstmultiple-capillary instrument was reported in 1990. Zagurskymodified a commercial DuPont Genesis 2000 sequencer tooperate with 500-�m ID capillaries (16). In that instrument, anargon ion laser beam was scanned across the capillary array.The instrument operated at 50 V cm–1; 9.5 h were required toseparate fragments 500 bases in length. Sequencing accuracywas <97% for fragments ranging from 29 to 512 bases inlength. Mathies reported a similar scanning instrument toimage a capillary array (17). That instrument operated with100-�m ID capillaries and produced sequencing informationup to 320 bases in length.
Duty cycle is an important parameter in specifying adetector’s performance. In scanning systems, the opticalsystem probes each capillary in sequence. Duty cycle is thefraction of time that a sample is illuminated. Duty cycle isimportant because DNA fragments migrate from the capillaryundetected during the period when a capillary is not
illuminated. In scanning systems, the duty cycle decreasesproportion to the number of capillaries.
In contrast, several systems have been developed to contally monitor fluorescence from a capillary array; these systeinherently have a much higher duty cycle than scannisystems. Yeung’s group reported a multiple-capillary DNsequencer in which a ribbon of capillaries was illuminated wia line-focused laser beam. Fluorescence was collected at rangles and imaged onto a CCD camera. The use of the Ccamera ensured that all capillaries were monitored simultaously (18,19). Similarly, an eight-capillary DNA sequencewas reported based on the use of individual fiber-opticsdeliver a laser beam to each capillary (20). A discrete lasbeam is required to excite fluorescence from each samplesecond set of fibers transmits fluorescence to an imagspectrograph and CCD detector. The sequence from 400 tobases was generated in 1 h. These designs are relatively incient in their use of excitation light. For example, if 10 mW olight is required to excite fluorescence from each capillarthen 1 W of light is required to excite fluorescence from 10capillaries.
In a further improvement, both Yeung and Kambara hareported a capillary array approach with side-illumination, ocolumn fluorescence detection (21,22). These instrumeprovide continual illumination of the samples with one lasbeam, providing a much higher duty cycle compared to a scning instrument and requiring a much lower laser power thused in the line-focused or fiber-optic excited systems. The ocolumn detection scheme is useful for small numbers of caplaries but appears to be difficult to scale to larger arrayKambara showed another approach by applying post-colufluorescence detection in a sheath-flow cuvette (23,2Sequencing length of 303 bases was achieved in 111 min.
Our research group reported the first capillary electrphoresis instrument based on post-column fluorescence detion in a sheath-flow cuvette (25). These detectors, borrowfrom the optical chamber used in flow cytometry, provide velow background and very high sensitivity fluorescence detetion, which allows the detection of individual fluorescenmolecules migrating from a capillary electrophoresis colum(26).
We first reported the use of the sheath-flow cuvette for sepration of single-base termination DNA sequencing fragmenby capillary electrophoresis in 1990 (6). The system w
*To whom correspondence should be addressed. Tel: +1 780 492 2845; Fax: +1 780 492 8231; Email: [email protected]
e36 Nucleic Acids Research, 1999, Vol. 27, No. 24
tedte
e
be
cenedil-
oneencethe
tic1-d-
nmtheectorserofattered
ray
the
expanded to four-color operation in 1991 (9). We reported asingle-capillary electrophoresis instrument that operates atelevated temperatures with non-crosslinked polyacrylamide(15). The 0%C polymer has low viscosity and may be pumpedfrom the capillary and replaced with fresh material after eachrun. Sequencing fragments over 640 bases were separated in 2 h atan electric field of 150 V cm–1 and at a temperature of 60�C.This report was the first description of high temperature sepa-ration of DNA sequencing fragments in non-crosslinkedpolymer. The high temperature operation increased sequencingrate, decreased compression, and increased the sequencingread length compared to room temperature sequencing.
In this paper, we report a five-capillary version of the high-efficiency electrophoresis system (27). The instrumentprovides high sensitivity, which is important for the detectionof the very small amount of sample loaded onto the capillary.The instrument also can operate at high temperature, whichminimizes the formation of compressions, and produces fasterseparations and longer read-length (15). This instrument fills aniche between the single-capillary Applied Biosystems 310and the larger-scale Applied Biosystems 377 slab-gel systemand the Applied Biosystems 3700 capillary array instrument,and should be of interest to small-scale sequencing laborato-ries.
MATERIALS AND METHODS
Instrumental design
The overall scheme for the instrument is shown in Figure 1(27). Two lasers, a 1-mW helium–neon laser operating in thegreen at 543.5 nm and a 4-mW argon ion laser operating in theblue at 488 nm, were alternately chopped with a sector wheelto provide sequential excitation. The beams were recombinedwith a dichroic beam splitter and focused with a 1� microscopeobjective into the locally designed multiple-capillary sheath-flow cuvette.
The rectangular sheath-flow cuvette (Fig. 2) was construcfrom high quality quartz by NSG Precision Cells. The cuvethad a 150�m � 750 �m inner chamber, four 1-mm thickwindows and a height of 2 cm. The two narrow walls of thchamber were tapered so that their spacing was 50�m wider atthe top than at the bottom. The taper forced the capillaries tosqueezed together as they were inserted into the cuvette.
Sheath fluid was pumped through the interstitial spabetween the capillaries. Either 10 mM borate (for free-zoelectrophoresis) or 1� TBE (for gel electrophoresis) was useas the sheath fluid. This buffer drew the sample from the caplaries, creating a set of independent sample streams,sample stream centered beneath each capillary. Fluorescfrom each sample stream was excited simultaneously bylaser beam that was focused ~100�m below the ends of thecapillaries.
A microscope objective (20� and 0.5 NA) collected fluores-cence. A filter wheel was located between the collection opand the detectors. This filter wheel was equipped with fourinch-diameter interference filters. For sequencing, the banpass of the filters was centered at 540, 560, 580 and 610with a 10-nm bandwidth. For short tandem repeat analysis,bandpass was centered at 530, 545, 560 and 580 nm. The swheel, which alternately transmitted the blue and green labeams, was synchronized with the filter wheel by usestepper motors driven by a common controller. Excitation488 nm was synchronized with detection at the two shorwavelength filters. Excitation at 543.5 nm was synchroniz
Figure 1. Four-spectral-channel laser-induced fluorescence detection. Thesector wheel alternately transmits the two laser beams and the filter wheel syn-chronizes the transmission of fluorescence in four bands chosen to match theemission spectra of the four dyes.
Figure 2. The sheath-flow cuvette fluorescence detection chamber for an arof five capillaries. The chamber is tapered, being 50�m wider at the top thanat the bottom. A single laser beam is used to illuminate fluorescence fromfive sample streams isolated by the sheath flow fluid.
ii
Nucleic Acids Research, 1999, Vol. 27, No. 24e36
odle
tedn.M/v)
as
ad atas
orrmleleom-
asblel;
ed.
d.ng
ch
rcho-
l
n
ersngles
7%
The
a
with detection at the two longer wavelengths. The filter wheelwas equipped with sensors to signal the identity of the filter inthe optical path.
It was not possible to image the fluorescence directly ontothe photodetectors; their relatively large size was incompatiblewith the spacing of the fluorescence image generated by thecollection optic. Instead, the fluorescence was imaged onto aset of five gradient refractive index (GRIN) lenses. TheseGRIN lenses were 1.8-mm diameter, 0.25 pitch and arranged at3-mm center spacing; there was one GRIN lens per fluorescentspot. Each GRIN lens coupled the fluorescence to a fiber optic,which was pig-tailed to an avalanche photodiode (APD). Thesephotodiodes operated in the single-photon-counting mode andprovided low dark count, high quantum efficiency photo-detection. The output of the single-photon-counting modulesconsisted of a train of pulses. The pulse train was converted toa voltage with a frequency-to-voltage converter, which wasmonitored with an analog-to-digital board. A MacintoshQuadra 700 was used for data acquisition.
Capillary electrophoresis
A 30-kV power supply was used to drive electrophoresis. Thesamples or running buffers were encased in a Plexiglas boxequipped with a safety interlock. The capillaries were held in alocally constructed heater based on Peltier devices and aproportional temperature controller, which held the tempera-ture of the capillaries with an accuracy of� 0.5�C. The sheathflow was provided either by a precision syringe pump or bysimply siphoning with the height difference of 5–2 cm betweensheath flow reservoir and waste container. Sheath flow ratewas typically 0.3 ml h–1.
Reagents and solutions
Fluorescein was a high purity standard from Molecular Probes.Borax, boric acid and EDTA were analytical-reagent grade. Trisbase and urea were ultra pure reagents. Acrylamide andN,N,N�,N�-tetramethylethylenediamine (TEMED) were electro-phoresis-purity reagents. Ammonium persulphate was ultra pureelectrophoresis grade. [�-(methacryloxy)propyl]trimethoxy-silane was reagent grade.
A stock borate solution was prepared by dissolving 0.478 gof borax in 50 ml water. It was diluted to a final concentrationof 10 mM borate (pH 9.2). A stock solution of ~1 mMfluorescein was made by dissolving fluorescein in ethanol. Aseries of four concentrations of fluorescein were prepared from2 � 10–11 to 2 � 10–12 M by diluting the stock fluoresceinsolution with 10 mM borate.
A 1% [�-(methacryloxy)propyl]trimethoxysilane solution wasprepared by adding 10�l of [�-(methacryloxy)propyl]trimeth-oxysilane in 990�l 95% ethanol–5% water solution (pH 4.8). Astock 10� TBE (pH 8.3) solution was prepared by dissolving108 g Tris, 55 g boric acid and 40 ml of 0.5 M EDTA in water toa final volume of 1 liter.
Capillary polymer preparation
We coated the capillaries using a method reported earlier (15).By using a water aspirator, the capillaries were first filled with1% [�-(methacryloxy)propyl]trimethoxysilane solution; after20 min, the silane solution was replaced with degassed 5%acrylamide, 7 M urea, 0.07% (w/v) ammonium persulphateand 0.07% (v/v) TEMED solution prepared in a TBE buffer
(1� TBE final). The polymerization reaction was allowed tproceed overnight. The 5%T polyacrylamide solution-fillecapillary was subjected to a 30-min pre-run before sampinjection.
The four-color sequencing reaction products were separausing a non-crosslinked dimethylacrylamide polymer solutioThe polymer was prepared from 6% dimethylacrylamide, 7urea, 0.07% (w/v) ammonium persulphate and 0.07% (vTEMED solution prepared in a TBE buffer (1� TBE final)under an argon atmosphere. The polymerized solution wloaded into a 10-ml syringe and stored at 4�C before use. Thepolymerized solution was injected into the capillary usinghomemade manifold system. The separation was performe50�C, the capillary length was 60 cm and the electric field w185 V cm–1.
Sequencing sample preparation and injection
All sequencing templates were M13mp18. The four-colsample was a cycle-sequencing product (15). A SequitheLong-Read Cycle Sequencing kit was used for cycsequencing with fluorescently labeled primers. Cycsequencing reactions were carried out according to the recmended protocols of the manufacturer. Cycle sequencing wperformed using 30 cycles on a PRT-100 ProgrammaThermal Controller equipped with a hot bonnet without oieach cycle consisted of 15 s at 95�C and 90 s at 70�C. Reactionproducts were pooled and immediately ethanol precipitatThe dried pellet was resuspended in 1.5�l of formamide. Theproducts were injected at 100 V cm–1 for 40 s.
Short tandem repeat (STR) analysis
Human blood samples were collected and DNA extractePolymerase chain reactions (PCR) were performed using 25genomic DNA. Each PCR contained dNTPs (2 mM eadATP, dCTP, dGTP and dTTP), 5� N buffer (50 mM KCl,10 mM Tris–HCl pH 8.3, 170�g ml–1 bovine serum albumin,0.05% Tween 20, 0.05% NP-40 and 1.5 mM MgCl2) and 0.5 UTaq polymerase. STR primers were ordered from ReseaGenetics or Integrated DNA Technologies Inc., and were flurescently labeled with 6-FAM, TET or HEX. The total PCRvolume was 15�l. The mixture was first subjected to an initiadenaturing step of 94�C for 1 min. PCR was performed for 30cycles consisting of 94�C for 1 min, 56�C for 2 min and 72�Cfor 1 min. The mixture was then treated to a final elongatiostep of 72�C for 7 min. PCR products were purified withMicrocon-30 columns to remove excess salts and primaccording to the manufacturers’ instructions. The final elutiowas in formamide. TAMRA-labeled Genescan-500 sizinladder was added to the purified PCR products and sampwere denatured at 94�C for 2 min immediately prior to injec-tion.
The PCR products were separated by use of a denaturinglinear polyacrylamide solution in 1� TBE buffer with 6 Murea. The polymer was prepared as described above.fused-silica capillaries were 43 cm long, 140�m OD and50�m ID.
Electrokinetic sample injection was 100 V cm–1 for 30 s. Thesheath flow buffer was 1� TBE. Electrophoresis wasperformed at 150 V cm–1 and 45�C. Subsequent to the run, datwere analyzed using Igor Pro and MatLab.
iii
e36 Nucleic Acids Research, 1999, Vol. 27, No. 24
wInenley.nta
eemathette.tertetlye
Asarect,he
-ur
e,-
ating.is
cesdther
ck-
nce
Data processing
A simple base-calling algorithm was used to analyze the data.The routine was written in MatLab and run on a G3 powerMacintosh and will be described in detail elsewhere. Theroutine has several components. The data were first convolutedthrough a Gaussian filter and then each of the traces was base-line corrected. Next, a response matrix was constructed basedon the relative fluorescence intensities in each spectralchannel. The data were multiplied by the inverse of this spec-tral response matrix to convert from spectral-space to dye-space (28). A mobility shift routine was incorporated toaccommodate differences in mobilities for fragments labeledwith the different dyes. Local maxima were identified andsequence was called based on the maximum dye response.Peak area was also calculated and used to identify and resolvemultiplets late in the run.
RESULTS AND DISCUSSION
Hydrodynamics
In our sheath-flow cuvette, an array of capillaries is snugly fitinto a rectangular quartz flow chamber. A simple siphonpumps the fluid through the interstitial spaces between thecapillaries and draws the analyte streams, one per capillary, inthe open region below the capillaries.
A single laser beam excites fluorescence from all samplestreams simultaneously. The highly transparent sheath fluidand the vanishingly low concentration of the DNA sampleproduces a negligible attenuation of the laser beam across thecuvette. Figure 3 presents a photograph of the sample streams.Each spot is ~50�m in diameter, which equals the inner diam-eter of the capillaries. The spots are uniformly spaced by~150�m, which equals the outer diameter of the capillaries.The fluorescent spots are well separated in the photograph,generating negligible cross-talk.
The design of a successful multiple-capillary sheath-flocuvette requires careful attention to hydrodynamic focusing.particular, it is necessary to have uniform sheath flow betweeach capillary. An unbalanced flow will cause the sampstreams to deflect towards the region of lower flow velocitThis deflection of the sample stream results in misalignmewith the optical system and can result in the failure to recordsignal from that capillary.
Uniform hydrodynamic flow is achieved if the capillaries aruniformly spaced within the cuvette. While it is possible to usmicromachined cuvettes to hold the capillaries on uniforcenters, our five-capillary instrument employs a somewhsimpler design that ensures uniform capillary spacing. Tcapillaries are inserted into a rectangular sheath flow cuveThe narrow dimension of the cuvette is matched to the oudiameter of the capillary (Fig. 2). The inner walls of the cuvetare slightly tapered so that the top of the cuvette is slighwider than the sum of the capillaries’ diameters while thbottom of the cuvette is slightly narrower than that distance.a result, the capillaries are squeezed together as theyinserted into the cuvette; since the capillaries are in contatheir spacing is very uniform, as are the sheath flow and tsample streams.
Optics
Microscope objectives are efficient collection optics for fluorescence detection in capillary electrophoresis (29). In osystem, a 20� 0.5 numerical aperture microscope objectivwhich provides a collection efficiency of 6.7%, collects fluorescence from the sheath flow cuvette. With 50-�m ID and150-�m OD capillaries, the objective produces an image thconsists of 1-mm diameter fluorescence spots at 3-mm spac
A set of single-photon counting avalanche photodiodesused for fluorescence detection. These very rugged deviprovide extremely high quantum efficiency (>50%) anreasonable dynamic range. The APDs are housed in ra
Figure 3. (A) Image of the GRIN lens array. The image was a back-illumination of the optical system. (B) Superimposed image of the fluorescence spots and bailluminated spots seen through an auxiliary microscope placed opposite the sheath-flow cuvette from the collecting optic. Capillaries were 50�m ID, 150�m ODand 37.0 cm long, filled with 10 mM borate, pH 9.2. Fluorescein concentration was 10–7 M. The argon ion laser power was 4.0 mW. Green spots are fluorescefrom sample streams, while the brown spots are scattering from the capillary tips.
iv
Nucleic Acids Research, 1999, Vol. 27, No. 24e36
l-edh-al
ngse-
sateill
tione
ive23kerser-of
veandzeantica-
-n-
cted
bulky containers, which contain low-noise amplifiers, Peltiercoolers and single-photon detecting electronics. The APDs aretoo bulky to be used directly to detect the fluorescence imagesfrom the capillary array; we cannot pack them closely enoughto simultaneously monitor fluorescence from each capillary.
A set of five GRIN-lenses, coupled to fiber optics, is used totransmit fluorescence from the image-plane of the microscopeobjective to the APDs. The core of the optical fiber is only100�m in diameter, which is much smaller than the 1-mmfluorescence spot. We use a set of GRIN lenses, 1.8 mm diam-eter and 0.25 pitch, at the image plane at 3-mm center spacingto couple fluorescence into the optical fibers (Fig. 2). Theseinexpensive, compact optical elements efficiently couple fluo-rescence into the optical fibers.
The use of optical fibers–GRIN lenses has proven to be valu-able in alignment of the system. The optical fibers can bedisconnected from the APDs. The detection end is illuminatedwith a lamp, creating back-illumination of the optical system.When viewed through an alignment microscope placed on theopposite side of the sheath flow cuvette from the collectionoptics, the illuminated optical fibers transmit light through theGRIN lenses to the 20� microscope objective into the sheath-flow cuvette (Fig. 3). Alignment is achieved by flowing dilutefluorescent dye through the capillaries; the relative position ofthe cuvette and the laser beam are adjusted until the fluorescentspots from the dye and the back-illuminated spots from theGRIN lenses are superimposed. The optical fibers are then re-connected to the APDs and a final tweaking of the opticalsystem is performed to maximize the signal from the APDs.
Detection limits
The limit of detection was evaluated in free-zone-electro-phoresis mode. The 37.0-cm capillaries were filled with a10 mM borate buffer. Electrophoresis was performed with anelectric field of 300 V cm–1 across the capillaries. A 1.1-nl plugof fluorescein was injected electrokinetically (1.0 kV, 5 s).
Figure 4 presents an electropherogram of a 2� 10–12 M solu-tion of fluorescein (2.2 zeptomol or 1300 molecules injected).The five traces were recorded simultaneously and presented asphoton counts per 200 ms window. Migration of the fluores-cein solution from the capillary generated one peak per capil-lary. The difference in migration time reflects the differencesin electro-osmotic flow between the capillaries; the capillarywalls were not coated for this experiment. The average peakarea corresponds to 12 000 photons above the backgroundsignal level; each molecule generated an average of ninedetected photons. Blank injections were performed by dippingthe capillaries into the dye solution without application ofinjection potential; no peaks were observed from these blanks.Detection limits (3�) were 130� 30 molecules (2� 10–13 M)injected onto the capillaries (30). These detection limits reflectthe good light collection efficiency of the optical system, thelow background signal generated in the sheath-flow cuvette,and the high quantum efficiency and low-noise of the APDs.Figure 4 also demonstrates the other important feature of thedesign—the fluorescence detection is free of cross-talk; a peakin one capillary did not generate a signal in an adjacentcapillary. The sheath flow not only lowered the background influorescence detection, but also provided excellent physicalisolation for the separation channels even when the capillarieswere in contact.
DNA sequencing at 50°°°°C
By incorporating two-laser-line excitation and four-spectrachannel detection (9) into the five-capillary system, we turnthe five-capillary instrument into a modest throughput, higperformance DNA sequencer. Figure 5 shows a typicsequencing separation performed at 50�C at a moderate elec-tric field strength of 185 V cm–1 in a capillary filled with 6%non-cross-linked polydimethylacrylamide. The sequenciaccuracy was 98.8% for 650 bases when using a simple bacalling algorithm written using MatLab. The softwareperformance was limited by difficulties in handling multipletlate in the run; all errors were associated with an inaccurestimate of these multiplets. Clearly, improved software wresult in improved sequencing accuracy.
Microsatellite analysis
Markers on chromosome 7 were chosen to test the applicaof microsatellite methodology to capillary electrophoresis. Thchromosome is ~184 cM (sex-averaged) in genetic length; fmicrosatellite markers (D7S479, D7S500, D7S501, D7S5and D7S554) were used to test the technology. These marcover almost half of the long arm of the chromosome. Genally, markers spaced at ~20 cM intervals allow the detectionlinkage to a distance of 10 cM on either side of any putatidisease-causing gene. Figure 6 presents five loci for a childits parents, along with the signal from a commercial sistandard. Although the D7S479 locus generated significstutter bands, the patterns are clearly resolved and identiftion of the STR pattern is trivial.
Figure 4. Injection of 1300 fluorescein molecules. The capillaries were 50�mID, 150�m OD and 37.0 cm long, filled with 10 mM borate, pH 9.2. Fluorescein, 2� 10–12 M, was injected at 1 kV for 5 s. The electrophoresis was coducted at an electric field strength of 300 V cm–1. Argon ion laser power was4.0 mW at 488 nm. Each data point was a 0.2 s count. The data were subjeto a binomial smoothing before plotting.
v
e36 Nucleic Acids Research, 1999, Vol. 27, No. 24
ineticersls.elyxt
to-n
Also
nt,ay,r,
re noted
The microsatellite allele sizes for each family member deter-mined by capillary electrophoresis were compared to thoseproduced using traditional radioactive labeling and slab gelseparation. The allele sizes generated by the two techniquescorresponded (data not shown), indicating that DNA fragmentsizes can be accurately determined and directly compared withthe use of an internal standard. Separation of DNA fragmentsby capillary electrophoresis was rapid, with DNA of 500 bp lengthdetected within 200 min with single base pair resolution.
CONCLUSIONS
We describe a five-capillary DNA sequencer based on asheath-flow cuvette. The instrument uses avalanche photo-diodes and a sheath-flow cuvette to produce extremely highdetection sensitivity, which is important when analyzing smallamounts of fluorescently labeled DNA. The instrument can
operate at 50�C, which is valuable in reducing compressionsDNA sequencing. The instrument can also be used for genmapping, where the use of fluorescently labeled size markfacilitates comparison of genotyping patterns between individua
The instrument currently relies on manual refilling of thcapillaries with sequencing matrix between runs. Rough2.5 h were required to refill the capillaries and analyze the nesample. This turnaround time would be improved with an aumated capillary refilling system, such as that found ocommercial instruments.
The instrument fills a niche between single-capillary DNsequencers and 96-capillary DNA sequencers. We have aconstructed 16- and 32-capillary versions of this instrumewhich will be described elsewhere (H.J.Crabtree, S.J.BD.Lewis, L.Coulson, G.Fitzpatrick, D.J.Harrison, S.DelingeJ.Z.Zhang, and N.J.Dovichi, paper submitted).
Figure 5. DNA sequencing run of an M13mp18 sample. The separation was performed in 6% non-crosslinked polydimethylacrylamide at 50�C at an electric fieldstrength of 185 V cm–1. The base-calls were performed using an algorithm written in MatLab. The called sequence is given above each peak. Errors abeneath the called sequence.
vi
Nucleic Acids Research, 1999, Vol. 27, No. 24e36
ip.
.
)
2)
.J.
ACKNOWLEDGEMENTS
This work was supported by the Canadian Genetic DiseasesNetwork, the Canadian Bacterial Diseases Network, theNatural Sciences and Engineering Research Council of Canada
(NSERC) and Sciex. K.V. acknowledges a graduate fellowshfrom the Alberta Heritage Foundation for Medical Research
REFERENCES1. Hunkapiller,T., Kaiser,R.J., Koop,B.F. and Hood,L. (1991)Science, 254,
67.2. Dovichi,N.J. (1990) In Landers,E. (ed.),CRC Handbook of Capillary
Electrophoresis. CRC Press, Boca Raton, FL, pp. 369–387.3. Swerdlow,H. and Gestland,R. (1990)Nucleic Acids Res., 18, 1415–1419.4. Drossman,H., Luckey,J.A., Kostichka,A.J., D’Cunha,J. and Smith,L.M
Figure 6. STR analysis of five loci from a child and its parents. The blue peaksare size-markers that were used to align the patterns. The data were treated witha color-inversion matrix to correct for spectral overlap between the dyes. Datafrom the migration period containing each locus is plotted in the five panels.
