llllllllllllll|||llllllllll|||l|llllllllllllllllllllllllllllllllllllllllll

llllllllllllll|||llllllllll|||l|llllllllllllllllllllllllllllllllllllllllll
llllllllllllll|||llllllllll|||l|llllllllllllllllllllllllllllllllllllllllll
US00555 6790A
United States Patent {191
[11]
Patent Number:
Pettit
[45]
Date of Patent:
[54]
5,556,790
Sep. 17, 1996
METHOD FOR AUTOMATED DNA
“AOTF Overview: Past, Present and Future”, by P. Katzka,
SEQUENCING
Acousto-Optic. Electro-Optic and Magneta-Optic Devices
[76] Inventor: John W. Pettit, 7808 Potters Mill Ct.,
Derwood, Md. 20855
[21] Appl. No.: 353,311
[22] Filed:
and Applications, SPIE vOL. 753, pp. 22-28 (1987).
“Large-Scale and Automated DNA Sequence DNA
Sequence Determination”, by T. Hunkapiller, et al., Sci~
encevol. 254, pp. 59-67 (Oct. 4, 1991).
“DNA sequenching: present limitations and prospects for the
future”by B. Barrell, The FASEB Journal, vol. 5, pp. 40-45
Dec. 5, 1994
(Jan. 1991).
“Genesis 2000 DNA Analysis System User Manual”,
Molecular Genetics Customer Support Group (Jan. 1992).
[51]
Int. Cl.6 .............................. ................... .. G01N 33/68
[52]
[58]
U.S. Cl. ........... ..
436/172; 436/91; 436/94
Field of Search ................................ .. 436/91, 94, 56,
“Reel Time Automated Simultaenous Double-Stranded
436/172
DNA Sequencing Using Two-Color Fluorophate Labeling",
by H. Karnbara, et a1. Biotechnology, vol. 9, No. 7, pp.
[56]
References Cited
U.S. PATENT DOCUMENTS
4,345,463
4,731,732
4,758,408
5,068,798
5,093,269
5,235,843
8/1982
3/1988
7/1988
11/1991
3/1992
Wilson et a1. .
Warchol et al. .
Krawetz et a1. .
Heath et a1. .
Leichnitz et al. .
8/1993 Langhorst .
OTHER PUBLICATIONS
“HPLC Analysis for 5-Methldeoxycytidine in Cellular DNA
Obtained Directly from the Culture Flask”, by S. P. Thacker,
et a1. Bio Techniques, vol. 16, No. 2, pp. 218-219 (1994).
“Acousto-optic Tunable Filters Spectrally Modulate Light”,
by X. Wang, Laser Focus World (May 1992) “Near-IR
Acousto-Optic Tunable Filter”, Crystal Technology Inc.
(Apr. 1993).
“Merging Spectroscopy and Digital Imaging Enhances Cell
Research”, by C. C. Hoyt, et al., Photonics Spectra, pp.
‘92-96 (Nov. 1992).
“Biophysical and Biochemical Aspects of Fluorescence
Spectroscopy”, T. G. Dewey, et al., Plenum Press, (pp.
73-104).
648-651 (Jul. 1991).
Primary Examiner—Lyle A. Alexander
Attorney, Agent, or Firm-Wigman, Cohen, Leitner &
Myers
[57]
ABSTRACT
An automated DNA sequencing apparatus and method is
disclosed in which a laser beam having a predetermined
wavelength is sequentially focused on a plurality of lanes of
DNA fragments migrating in a polyacrylamide gel and in
which the DNA fragments have been tagged with ?uorescent
compounds. The resulting ?uorescent light given off by each
of the tagged DNA fragments is collected and di?racted by
means of an acousto-optic tunable ?lter and the level of the
diffracted light is then detected and analyzed in order to
determine the DNA sequences of the analyzed DNA frag
ments. The automatic DNA sequencing system of the
present invention operates under computer control to repeat
a series of measurements of the ?uorescent data at a multiple
number of wavelengths such that the entire spectrum of
?uorescence emissions from all four ?uorescent tagged
DNA bases can be measured at any number of wavelengths.
“Acousto-optic Tunable Filters”, by I. C. Chang, Optical
Engineering, vol. 20, No. 6, pp. 824-829 (Nov./Dec. 1981).
22 Claims, 5 Drawing Sheets
40
2342
44
4B
US. Patent
Sep. 17, 1996
Sheet 1 of 5
5,556,790
FIG.1
FIG.2A
DETECTIEIN AMPLITUDE
VERSUS WAVELENGTH
% [INE DATA RECURU
SYNTHESIZER FREQUENCY mvELENm'm
FIG.2B
EMISSIEN UUTPUT
SPECTRA FUR
FUUR TAGS Tl-T4
T1
T3
WAVELENGTH
US. Patent
Sep. 17, 1996
Sheet 2 of 5
5,556,790
REcuRB
ME
NLMBER\ 1
A
sEBuENcE\
G
E
T
T2
T3
T4
[3 10
T;
15
T 20
A 25
~-*’
30
C 34
K/
T1
FIG- 3
DETEETIUN AVPLITLIDE
FIG.5A
FIG.5D
ENTER sTART Am ~50"
C?
EN: vAvELENBTHs
I
ENTE NMBER DF ~5°2
“'"’
(
508
EXAMINE BANB PBsITInN
BATA FILES
‘
51o
WAVELENGTH sTEPs
‘
ENTER MJUNT EIF DELAY ~5°4
BETUEEN DATA RECERDS
l
CDMPUTE sTABTANB ENB ~5°5
(
REmvE BAsE FRIJM BATA FILE
AN] SHIFT REMAINING BAsEs
IN ME FILE BuvN BY BNE
512
FREBuENcIEs ANJ
END “5 DATA
FREBuENcY INCREMENT
-
é
514-’ EN]
US. Patent
Sep. 17, 1996
Sheet 3 of 5
5,556,790
601
FIG.4
U.S. Patent
Sep. 17, 1996
5,556,790
Sheet 4 of 5
515
INITIALIZE CDTPUTER MEHIRY
AN] [PEN DATA FILES
CREATE DATA FILE REIIRDS
T
DLITPUT DIGITAL VDRD
v
518
TD SYNTHESIZER
T
DELAY
T
IN’LIT SIGNAL FRDM
522
DETECTDR ELECTRDNICS
A
EDWERT IN’UTTED
DATA HDRD
\224
A
FIG.5B
526
STDRE DATA IN MEMERY '
A
INCREMENT DATA RECERD
DATA ELEMENT PDINTER
T
528
V
AH] INCREMENT TD DIGITAL HERD
TD BE DUTPLIT TD SYNTHESIZER
530
(534'
INCREMENT DATA
REEDRD PDINTER
T
RESET DIGITAL
[IJTPUT VDRD
‘ DELAY
CLDSE RECDRD FILE
US. Patent
Sep. 17, 1996
T
5,556,790
542
CREATE BAN] DATA FILE
T
PERFORM LEAST
Sheet 5 of 5
\5,“
SQUARES CALCULATIDN
T
INTEGQATE EMISSIEN £45
LEVELS
END DF
DATA RECDRD
FIG.5C
‘7
YES
550
CLDSE FILES
A
[PEN FDUR BAND
PDSITIDN DATA FILES
/ 552
T
EXAMINE FIRST DATA
‘5,54
ELEMENT DF EACH RECDRD
T
READ IN DATA CDRRESPENDING TD
~
"’ UNE BAND PASSING THROUGH DETECTOR ‘—"'"
A
558
PERFDRM LEAST SuJARES H
FIT CALCULATIONS
EXAMINE NEXT DATA
ELEMENT DF EACH
55;‘ Tasman IN aAm FILES
5,556,790
1
2
NIETHOD FOR AUTOMATED DNA
to resolve fragments with a resolution of one base pair, and
that resolution is necessary for sequencing. Each fragment is
SEQUENCING
BACKGROUND OF THE INVENTION
5
The present invention relates generally to the investiga
tion of the sequencing of DNA. More particularly, the
present invention relates to a method of and apparatus for
automating the sequencing of DNA which increases the rate
at which DNA can be sequenced as well as improving the
reliability and accuracy of the sequencing determination.
labeled with a radioactive element that typically gives o? a
beta particle, such as radioactive phosphorus, P-32. Each of
the four samples are then separated in size in their own lane
in the gel. The four lanes are typically side by side. After
electrophoresis, a piece of x-ray ?lm is placed next to the gel
for a number of hours, often a couple of days, to expose the
?lm with the radioactive emissions from the P-32 phospho
rus. When developed, the fragments show up as dark bands
on the ?lm and the sequence can then be read from the order
in which the bands appeared, from the bottom to the top of
DNA sequencing is essential to the practice of biotech
nology, genetic engineering and many other disciplines that
the ?lm.
rely on the need to determine the genetic information
contained in DNA. The sequencing of DNA is the process of
determining the sequence of nucleic acid bases that com
prise a strand of DNA. There are four bases, denoted A for
Automating DNA sequencing involves automating the
process of detecting the fragments on the electrophoresis gel
and then automatically determining the DNA base sequence
from the sequence of detected fragments using the above
algorithm implemented in a microprocessor. Because of the
adenine, G for guanine, C for cytosine, and T for thymine,
that comprise the DNA. The sequence of these bases
uniquely describes each piece of DNA. Sequencing is a
time needed to expose the x-ray ?lm to the beta radiation of
20
the P-32 phosphorus, and other considerations involving the
crucial step in genetic engineering and biotechnology, since
it provides the precise code of genetic information contained
use of radioisotopes, new methods of tagging and sequenc~
in a sample of DNA.
DNA is double stranded and hence, the term base pairs is
often used, since each base of one strand is opposed by its
and Biochemical Aspects of Fluoresene Spectroscopy, edited
by T. Gregory Dewey, Plenum Press, 1997; “Large Scale and
Automated Sequence Determination,” by T. Hunkspillar, et
al., Science, Vol. 254, pages 59-67 (1991) and “DNA
Sequencing: Present Limitations and Prospects for the
Future,” by B. Barrell, the FASEB Journal, Vol. 5, page
40-45 (1991).
ing based on fluorescence were developed. See Biophysical
25
complimentary base on the other strand. There are an
enormous number of bases that need to be sequenced in
order to read a piece of DNA. Even a simple piece of DNA
from a bacteria cell would likely comprise several thousand
bases. The Human Genome Project, a large, multi-year,
United States Government funded national project to
sequence the DNA in humans, is attempting to sequence the
approximately 109 bases found in human DNA.
DNA sequencing is a very labor intensive process and,
Fluorescence tagging of the fragments involves the
with the large amounts of DNA that are needed to be 35
sequenced for the biotechnology industry to progress, meth
ods and apparatus to automate this process are very desir
able. Much has been written about DNA sequencing and
genetic engineering and the reader is referred to the many
references in this subject which will provide additional
optical wavelength that is optimum for that molecule. Fluo~
40
background information.
Two methods of DNA sequencing have been developed.
The ?rst is by Maxam and Gilbert, and is described in Proc.
Nail. Acad. Sci. USA by AM. Maxam and W. Gilbert, Vol.
74, page 560 (1977). The second method is described in
Proc. Natl. Acad. Sci. USA, by F. Sangen, S. Nicklen and A.
R. Coulson, Vol. 74, page 5463 (1977). Both of those
attachment of a ?uorescent compound, or ?uorophore, to
each fragment analogously to the attachment of the radio
active label to each fragment. These ?uorescence labels
were found to not adversely affect the process of gel
electrophoreses or sequence.
Fluorescence is an optical method that involves stimulat
ing the ?uorescent molecule by shining light on it at an
rescent light is then given off by the molecule at a charac
teristic Wavelength that is typically slightly longer than the
stimulation wavelength. By focusing the light at the stimu
lating wavelength down to a point on the gel and then
detecting the presence of any optical radiation at the char
45
methods involve performing a number of steps before the
fragments of DNA are ready to be detected to yield a
sequence. Those steps will not be reviewed here, as they are
detailed in the two references noted above. Although the two
acteristic wavelength of light from the ?uorescent molecule,
the presence at that point of fragments of DNA tagged with
that ?uorescent molecule may be determined.
Two methods of implementing an automated DNA
sequencing instrument are known in the art. One, reported
by Smith et al., in Nature, Vol 321, pages 674-679 (1986),
puts a different ?uorescent tag on each of the four samples
of fragments described above. Thus, the sample of frag
techniques differ, eventually one arrives at four samples of
ments that end in the base A are tagged by one ?uorophore;
DNA fragments that end at a given base. For instance, one
the sample of fragments that end in the base G are tagged by
sample contains fragments that end in base A; another other
another ?uorophore, and so on for the other two samples.
Each ?uorophore can be distinguished by its own stimula
contains fragments that end in base G, and so forth.
The task is to separate those fragments by size and see
what order they are in. If the shortest fragment in all of the
four samples is one that ends in T, then the ?rst base of the
sequence is T. If the second shortest one ends in C, then the
next base in the sequence is C, and so forth, until all of the
tion and emission wavelengths of light.
In the Smith et al. method, all four samples are electro
phoresed in the same lane together and the differences in
their tags are used to distinguish them. That has the advan
tage that four separate lanes are not used, since the progres
sion of fragments in different lanes is often not consistent
with one another and di?iculties often arise in determining
fragments are separated in order of increasing length and the
sequence is determined.
In order to perform this size separation and fragment
detection, the ?rst methods of manual DNA sequencing
utilized polyacrylamide gel electrophoresis techniques to
separate the fragments. Polyacrylamide gels have the ability
the sequence as a result.
65
Another method, reported by Ansor et al., in J. Biochem
Biophys. Methods, Vol. 13, pages 315-323 (1986) and
Nucleic Acids Res., Vol 15(11), pages 4593-4602 (1987),
5,556,790
3
4
uses one ?uorescent tag for all fragments, but employs four
separate lanes of gel electrophoresis in a manner that is
disadvantage in DNA sequencing which is overcome by the
present invention.
Another prior art system for automated DNA sequencing
was developed by DuPont of Wilmington, Del., and mar
similar to radioactive labeled sequencing. That approach has
the potential disadvantage that four lanes, with different
fragment migration rates caused by local temperature varia
keted as the Genesis 2000. It uses four ?uorescent tags, one
tions and other inconsistencies within the gel, could limit the
for each base, and runs them in a single lane. This system
reliability of the sequence determination.
Fluorescence tagging and the detection of natural ?uo
uses two ?xed wavelength optical interference ?lters, each
affixed to a photomultiplier tube light detector. The center
wavelengths and bandwidths of these ?lters are chosen such
rescence in molecules is a method of analytical chemistry
and biology that is well known in the art. The methods 10 that one of them is to the short wavelength side of the
described above have been developed for DNA sequencing
spectrum of wavelengths from the ?uorescent tags and the
by the creation of ?uorescent tags that can be bound to
other is on the long wavelength side. By looking at the ratio
fragments of DNA. The instruments used to detect ?uores
of the two detected signals, a determination can be made as
cence consist of the following parts. A light source with a
to which base is being detected at a given time. For instance,
broad optical bandwidth such as a light bulb or a laser is used 15 if the signal from the short wavelength detector is much
as the source of the stimulating light. An optical ?lter is used
greater than the other one, it is inferred that the ?uorescent
to select the light at the desired stimulation wavelength and
tag with the shortest wavelength is present. Likewise, if the
beam it onto the sample. Optical ?lters are available at
signal from the long wavelength ?lter is much greater, then
essentially any wavelength and are typically constructed by
the ?uorescent tag with the longest wavelength emission is
the deposition of layers of thin ?lm at a fraction of the
assumed to be present. If the two signals are close together
in magnitude, then the ?uorescent tag tending to that side of
wavelength of the desired transmission wavelength. The
light that exits the optical ?lter is then applied to the sample
the two remaining ones in the middle, is assumed to be
present. This technique, although workable, does not offer
much resolution and it is reported by Dewey to no longer be
to stimulate the ?uorescent molecule.
The molecule then emits light at its characteristic ?uo
rescent wavelength. This light is collected by a suitable lens
and is then passed through a second optical ?lter centered at
25
commercially available.
Yet another prior art system for automated DNA sequenc
the characteristic wavelength before being brought to a
ing has been developed by LI-COR, Inc., of Lincoln, Nebr.,
detection device such as a photomultiplier tube, a photo
conductive cell, or a semiconductor optical detector. There
which is marketed as the Model 4000 L Automated DNA
fore, only light at the desired characteristic wavelength is
one ?orescent tag that emits radiation in the near infrared
detected to determine the presence of the ?uorescent mol
ecule.
portion of the spectrum. The use of such a ?orescent tag is
Sequencer. This LI-COR system apparently operates using
to produce less background ?orescence since the regular
glass that the gel plates are made of, which is called “?oat
The Ansorge method involves only a single light source,
one stimulation optical ?lter, one ?uorescent radiation opti
cal ?lter and one optical detector. This apparatus is mechani
cally scanned across the gel to detect the presence of the
fragment in the four lanes. Mechanical scanning is a disad
glass” does not emit its own ?orescence in the infrared
35
cence in the visible wavelength range, which is a problem
that has been noted in the prior art.
In addition, one method for reducing or eliminating the
vantage due to the slow rate at which the apparatus can be
scanned, the alignment of the scanning and the repeatability
and durability of scanning mechanisms.
40
The Smith method uses only one lane of electrophoresis,
so it does not have the disadvantage of mechanically scan
ning the apparatus across the four lanes. However, there is
a plurality of stimulation wavelengths and detection wave
lengths that must be implemented. To do this, a mechanical
wheel with four optical ?lters attached to it is rotated in the
beam of the stimulation optical radiation. That selects in
sequence the four stimulation wavelengths from the broad
45
Recently, methods of preparing DNA fragments for
sequencing using capillary hair techniques have been
described. The present invention can be used with such
55
ing methods. A commercially available DNA sequencer that
SUMMARY AND OBJECTS OF THE
INVENTION
operates in a similar manner to that described above is
available from Applied Biosystems, Inc., of Foster City,
mechanical operation of its four color optics. That is a big
methods of preparing DNA fragments for sequencing, as
well as any additional methods or techniques of preparing
DNA fragments, as long as ?uorescent tags are used which
can be caused to emit light in response to an impinging light
beam.
with respect to maintaining synchronism and optical align
DNA sequencer results in relatively long run times due to the
long emission lifetimes, while the ?orescence tags them
selves have short emission lifetimes. Thus, the interfering
emissions from the ?oat glass do not respond to the rapid
chopping of the laser stimulation light and becomes a
subtractible background signal. The present invention over
optic tunable ?lter or acousto-optic modulator.
ment are severe. This fact is pointed out in chapter 3 of the
Dewey text, which is a review of automated DNA sequenc
California, and is designated the ABI model 373A. Hunk
pillar, in his April 1992 article in Science, points out that this
background ?orescence produced by the ?oat glass when
irradiated with the light in the invisible wavelength range is
to “chop” the laser light. Such a method works because the
?orescence emissions of the constituents in the glass have
comes this background ?orescence problem in a more
elegant and effective manner through the use of an acousto
band optical radiation source. At the same time, a second
mechanical wheel is ?tted to the detection optical path to
select the correct detection wavelength corresponding to the
?uorophore being stimulated at the same time by the stimu
lation ?lters. Therefore, two mechanical rotating devices
must be implemented, and operated in synchronism, in order
to produce the correct result.
The complexities of such an electro-mechanical device
wavelength range. Float glass does, however, emit ?ores
In view of the foregoing, it should be apparent that there
65
still exists a need in the art for a method of and apparatus for
an improved DNA sequencer which has an extended range
of the number of bases that can be sequenced in a single run
5,556,790
5
6
compared to prior art devices. It is, therefore, a primary
object of this invention to provide a method of and apparatus
for DNA sequencing which utilizes solid state, digitally
controlled equipment and is capable of operating several
orders of magnitude faster than the known automated DNA
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic block diagram showing the appa
ratus of the present invention;
FIG. 2A is a drawing of a graph of a data record showing
the detection amplitude versus the wavelength for a single
sequencing technology.
data record;
FIG. 2B is a drawing of a graph illustrating the emission
output spectra for Tags T1-T4 for the data record of FIG.
More particularly, it is an object of this invention to
provide an improved DNA sequencing system as aforemen
tioned which utilizes solid state and reliable electronic
components, which is not subject to mechanical failures and
which does not require frequent alignment nor costly com
2A;
FIG. 3 is a diagram of the band data generated using the
system of the present invention showing the record number,
ponents.
Still more particularly, it is an object of this invention to
base sequence and detection amplitude versus record num
provide an improved automated DNA sequencing system
ber for each of the Tags T1-T4 from the data record shown
in FIGS. 2A-2B.;
FIG. 4 is a block schematic diagram showing an alternate
and preferred embodiment of the apparatus of the present
which uses a unique combination of electrically tumble
optical ?lters which are employed to wavelength analyze the
light emitted by the DNA fragments under intense laser light
stimulation.
Another object of the present invention is to provide a
reliable and relatively inexpensive automated DNA
sequencing system for use in analyzing DNA bases.
Brie?y described, these and other objects of the invention
invention;
20
FIG. 5A is a diagram of a ?ow chart of a portion of the
DNA sequencing software used for inputting parameters in
connection with the operation of the apparatus of the present
invention;
are accomplished by providing a solid state digitally con
trolled optical system which focuses a beam of laser light
FIG. 5B is a diagram of a ?ow chart of a portion of the
DNA sequencing software used for collecting band data in
connection with the operation of the apparatus of the present
onto the gel containing multiple lanes of DNA fragments or
into the capillary hairs containing such DNA fragments. In
invention;
order to provide increased rejection of unwanted ?uores~
FIG. 5C is a diagram of a ?ow chart of a portion of the
cence and to stimulate ?uorescence tags at their optimum
DNA sequencing software relating to analyzing data records
wavelength to reduce stimulation of undesired ?uorescence
tags, the laser light may be chopped or modulated, before it
and band data used in connection with the apparatus of the
impinges upon the gel. The wavelength of the laser beam is
present invention; and
such that it stimulates the ?uorescent molecules with which
FIG. 5D is a diagram of a portion of the DNA sequencing
software relating to determining the base sequence of a
the DNA fragments have been tagged. The ?uorescent light
given off as a result of the stimulation of the DNA fragments
by the laser light beam is collected by a lens system and fed
fragment of DNA used in connection with the operation of
35
to a computer controlled acousto-optic tunable ?lter which
is operated in such a manner that the ?uorescent light
DETAILED DESCRIPTION OF THE
PREFERRED EMBODIMENT
entering the acousto-optic tunable ?lter is wavelength ?l
tered such that light at the selected wavelength exits from the
acousto-optic tunable ?lter at an angle of approximately
7—l0 degrees from the main light beam exiting from the
acousto-optic tunable ?lter. The acousto-optic tunable ?lter
One of the objectives of the inventive DNA sequencing
system is to extend the range and the number of bases that
are sequenced in a single run. However, the migration rate
of longer fragments does not differ by much when their
is controlled by means of a digital synthesizer which itself
is controlled by the computer.
The selected light exiting from the acousto-optic tunable
the apparatus of the present invention.
45
length differs by only one nucleotide. In addition, when the
fragment lengths reach the range of about 300 to 400 bases,
the difference in migration distance in the polyacrylamide
?lter is diffracted in both directions relative to the main
gels causes the bands to overlap and the sequence can no
beam of light leaving that ?lter. Each or both of the
longer be determined. That is because it can not reliably be
di?racted beams at the desired selected wavelength are
detected in order to provide an electrical measurement of the 50 determined which band is further ahead of another when the
detected signals blur together. Accordingly, a technique with
?uorescent emissions of the DNA fragments at the selected
greater resolution in detecting and determining the position
wavelengths. The output from the detector is fed to the
of the bands within the gel would extend the range of the
computer and may be stored for later analysis.
base sequence determination for a given run. A noticeable
The computer operates the instant automated DNA
sequencing system by repeating a series of measurements of
55
increase in that capability is of great utility to the biotech
nology industry because of the time required to prepare and
?uorescent data at a multiple number of wavelengths. In that
perform a gel separation.
manner, the entire spectrum of ?uorescence emissions from
The present invention utilizes a unique combination of
all four ?uorescent tags is then measured at any number of
wavelengths. The ?uorescence emissions measured at each
electrically tunable optical ?lters, optical techniques and
of the selected wavelengths are then de?ned as a data record, 60 signal processing algorithms to signi?cantly improve the
ability to determine DNA base sequences over the prior art
from which the DNA sequences can be determined.
With these and other objects, advantages and features of
the invention which may become hereinafter apparent, the
nature of the invention may be more clearly understood by ,
reference to the following detailed description of the inven
tion, the appended claims and to the several drawings
attached herein.
65
methods and equipment discussed above.
Referring now in detail to the drawings wherein like parts
are designated by like reference numerals throughout, there
is illustrated in FIG. 1 a diagram of the apparatus of the
automated DNA sequencing invention which is used to
generate and collect the ?uorescent data from the bands of
5,556,790
7
8
DNA in an electrophoresis gel. A pair of parallel glass plates
lens 30 is preferably a converging lens while the lens 32 is
preferably a diverging lens such that their combined eifect is
to reduce the beam of light emerging from the converging
10 that hold the gel material 12 therebetween are illustrated
in which a plurality of wells 14 in the gel which are used to
load the tagged DNA fragments to create lanes are shown in
collimating lens 28 by approximately the ratio of the focal
length of the converging lens 30 and diverging lens 32.
which the DNA fragments 16 are migrating down the gel 12.
The ?uorescent light collected by the converging colli
mating lens 28 then enters the acousto-optic tunable ?lter 34
in order to be wavelength ?ltered. The acousto-optic tunable
?lter 34 may be an Acousto-Optic Tunable Filter available
The gel 12 is loaded with fragments of DNA prepared in
the prescribed manner by the sequencing techniques known
in the art and described above and then tagged with four
different ?uorescent molecules, one for each base type. All
four base types are then loaded into one of the plurality of
wells 14 and become one lane on the gel.
A laser 24 generates a laser beam 18 which is capable of
being moved in a horizontal direction in order to collect data
from Crystal Technology, Inc. of Palo Alto, Calif.
The acousto-optic tunable ?lter 34 is an electrically
tunable optical ?lter which acts as a controllable optical
from each of the many lanes 16 which are loaded in the same
manner. Due to the compact size of the detection apparatus 15
of the present invention as well as its speed of operation,
many more parallel lanes can be run with the apparatus of
the present invention than can be run using previously
described methods and apparatus. The method of and appa
ratus for the present invention thus provides a distinct
advantage over such prior art methods and apparatus,
because parallel processing of the multiple lanes on a single
dioxide (TeO2) in which an acoustic wave is launched into
the crystal to create an interaction between the acoustic
wave and the light passing through the crystal.
The compression and rarefractions of the crystalline mate
rial due to the launched acoustic wave causes a diffraction
grating-like effect to take place throughout the volume of the
crystal. The optical wave whose wave vectors satis?es the
wave matching condition with the acoustic wave is accu
gel 12 increases the throughput of the automatic DNA
sequencing system.
wavelength ?lter such that a desired wavelength of light can
be passed through the ?lter while all other wavelengths are
blocked. Such acousto-optic tunable ?lters can be imple
mented by an optically birefringent crystal, such as tellun'um
25
mulatively diffracted and exits the crystal at a different angle
from the rest of undiifracted light. Using such a device, a
suitable arrangement of stops and apertures can be utilized
to select a desired wavelength that has been diffracted for
As described above, the laser light source 24 generates the
laser light which is used to stimulate the ?uorescent mol
ecules contained in the DNA fragments. The light from the
laser 24 is ?rst wavelength ?ltered by a ?xed optical ?lter 22
which allows only light at the desired wavelength for
stimulating the ?uorescent molecules to pass through it.
Alternatively, an acousto-optic tunable ?lter can be used in
frequencies utilized with such devices are in the range of
30
tens to hundreds of megahertz in frequency, those waves can
be established very quickly, on the order of microseconds or
place of the ?xed optical ?lter 22. As discussed hereafter in
less. Such acousto-optic tunable ?lter de,ices are therefore
detection and subsequent processing. Since the acoustic
capable of rapid wavelength tuning.
connection with FIG. 4, an infrared ?uorophore can be
utilized to decrease or eliminate the background ?orescence
A second type of optical ?lter utilizes pneumatic liquid
crystals whose birefrengence may be altered by varying the
produced by the glass plates 10 which carry the gel material
12. As described more fully hereinafter, such background
?orescence can also be eliminated by high~speed modulation
techniques. The thus ?ltered laser light 18 is then focused
onto the gel 12 at the point where detection is to occur as a 40
point source that is as small as can be attained. That is
typically a diffraction limited point of light, which is pro
duced by a lens 20.
The ?uorescent light 26 given oif as a result of the
stimulation of the DNA fragments 16 by the laser light beam
45
electric ?eld that surrounds them. That has the effect of
controlling the long chain-like molecules which in turn
alters their birefrengence. That electric ?eld can be estab—
lished and controlled by varying a voltage applied to a
parallel plate which contains the liquid crystals. An optical
?lter is created by using several zones of liquid crystals in
order to implement a Lyot interferometer.
The acousto-optic tunable ?lter 34 is tuned using a digital
synthesizer 36 which creates a radio frequency signal which
18 appears to come from a point source due to the small spot
corresponds to the desired optical wavelength to be analyzed
size of the laser produced ?uorescent excitation light beam
at each moment in time. The output from the digital syn
18. The ?rst lens 28 is a converging collimating lens which
collects the ?uorescent light. It is preferably chosen with a
thesizer 36 is ampli?ed by a radio frequency ampli?er 38
such that the power output from the digital synthesizer 36 is
large diameter so that it collects as much of the ?uorescent 50 increased to approximately 0.5 to 1.0 watt. The ampli?ed
light 26 as possible. The converging collimating lens 28 is
positioned and aligned such that its focal point is exactly at
the point of emission of the emitted ?uorescent light 26. The
?uorescent light 26 therefore comes out of the converging
collimating lens 28 as parallel rays.
55
A second lens 30 upon which the parallel rays of light
exiting from the converging collimating lens 28 impinge and
radio frequency signal is then fed from the output of the
radio frequency ampli?er 38 to the control input of the
acousto-optic tunable ?lter 34. Light at the selected wave
length then exits from the acousto-optic tunable ?lter 34 at
an angle of approximately 7 to 10 degrees from the rest of
the light beam at 42.
The selected light 40 and 44 from the acousto~optic
a third lens 32 operate together to reduce the diameter of the
tunable ?lter 34 is diffracted in both directions relative to the
beam of parallel rays of ?uorescent light 26 output by the
converging collimating lens 28 so that the resultant beam
main beam 42 leaving the acousto-optic tunable ?lter 34.
contains parallel rays and has a diameter that is compatible
with the entrance aperture of the acousto-optic tunable ?lter
34. The use of those two lenses 30 and 32 is important
because the entrance aperture of the acousto-optic tunable
?lter 34 is small, typically on the order of 3-7 mm and it is
desirable to pass as much of the collected ?uorescent light
as possible through the acousto-optic tunable ?lter 34. The
60
The two diifracted beams 40 and 44 diifer in polarization
from one another. Each or both of the diffracted beams 40
and 44 at the selected wavelength can be detected by light
detectors to provide a measurement of the ?uorescent emis
sions of the DNA fragments at the selected wavelength. For
65
the purposes of illustration, only a single detector 46 which
is positioned to detect the diffracted light beam 44 is shown
in FIG. 1.
5,556,790
9
10
The diffracted light representative of the ?uorescent emis
sions which emerges from the acousto-optic tumble ?lter 34
utilize a second acousto-optic tunable ?lter or acousto~optic
modulator to chop the laser light before the laser light
impinges on the gel to stimulate the ?uorescence molecules.
Such a method allows electronic modulation of the stimu
lation and provides a mechanism for accomplishing syn
chronous detection or matched ?lter detection of the light at
the detector. This greatly increases the rejection of the
can be detected by various types of devices, such as a
photomultiplier which is cooled to decrease its noise back
ground, or by a solid-state detector such as a PN junction or
PIN diode. The detector should have good sensitivity with
low noise characteristics in order to accommodate the low
signal levels to be detected while at the same time it should
background ?uorescence emitted by the ?oat glass. An
be relatively fast to keep up with the speed that acousto-optic
acousto-optic modulator, which is similar to the acousto
optic tunable ?lter, is designed to only pass or not to pass the
light and is used essentially to turn the light on or off under
electronic control.
A matched ?lter arrangement can alternatively be utilized
where the stimulation laser light is modulated with a signa
ture by use of an acousto-optic tunable ?lter and acousto
tunable ?lters can be scanned in order to realize the potential
of the automated DNA sequencing system of the present
invention.
The output from the detector 46 is fed to a computer 50
through a suitable interface device, such as an analog-to
digital converter (not shown) interface to the computer 50.
Prior to supplying its output to the computer 50, the output
from the detector 46 is electrically ?ltered using ?lter 48 in
order to optimize the signal-to-noise ratio of the measured
?uorescence value without degrading the speed of response
of the system.
The computer 50 or an equivalent logic implementing
unit, commands the logic and operation of the DNA
sequencing system and also records the measured ?uores
cent data. The computer 50 is also connected to the digital
optic modulator and the matched ?lter response only to a
signal that matches that signature. Such techniques take
advantage of the operating characteristics of acoustic-optic
20
as a practical matter, operate far too slowly to be of much use
synthesizer 36 in order to control the generation of the radio
frequency signal which corresponds to the desired optical
wavelength to be analyzed. The computer 50 may be an IBM
or compatible type AT or more powerful personal computer
equipped with appropriate memory, hard and ?oppy disk
drives, video monitor and other peripherals. Preferably, the
30
computer 50 may be formed from a microprocessor or
microcontroller having the appropriate memory, computing
power and other characteristics.
The computer 50 operates the instant automated DNA
sequencing system by repeating a series of measurements of
fluorescent data at a multiple number of wavelengths. The
entire spectrum of ?uorescence emissions from all four
to eliminate this background ?uorescence.
Turning now to FIG. 4, there is shown a block schematic
diagram of an altered and preferred embodiment of the
apparatus of FIG. 1. The apparatus shown in FIG. 4 provides
increased rejection of unwanted background ?uorescence as
well as stimulating ?uorescence tags at their optimum
wavelength in order to reduce stimulation of undesired
?uorescence tags.
In the embodiment of the invention shown in FIG. 4, an
acousto-optic tunable ?lter 68 is inserted in the path of the
laser beam 18 prior to the laser light beam 18 impinging
upon the DNA fragments 16. The acousto-optic tunable ?lter
35
?uorescent tags can then be measured at any number of
wavelengths under command from the computer 50. Pref
erably, a minimum of four di?ferent wavelengths are utilized,
tunable ?lters. On the other hand, prior art spinning ?lter
wheels provide no control over those signal properties and,
40
68 is controlled by means of a digitally controlled radio
frequency synthesizer 62 whose output is connected to a
chopping or modulating circuit 64. The output from the
chopping or modulating circuit 64 is ampli?ed by means of
the radio frequency power ampli?er 66 which then feeds its
output to the input of the acousto-optic tunable ?lter 68. The
acousto-optic tunable ?lter 68, the digitally controlled syn»
one wavelength centered on the ernissionof each ?uorescent
thesizer 62 and the radio frequency power ampli?er 64 may
molecule, to a maximum of as many wavelengths as are
desired which may, as a practical matter, have an upper limit
be the same elements as described in connection with the
?rst acousto-optic ?lter 34, the digital synthesizer 36 and the
radio frequency ampli?er 38 shown in both FIGS. 1 and 4
of approximately 20 individual wavelengths per each ?uo
rescent molecule.
and as described in more detail in connection with the
The ?uorescence emissions measured at each of the
selected wavelengths can be de?ned as a data record. The
description of FIG. 1. The modulating or chopping circuits
computer 50 steps the acousto-optic tunable ?lter 34 to each
waveform generators 54 and 56, respectively.
64 are controlled by means of the digital oscillators or
of those wavelengths in turn and records the amount of 50
The digital oscillators or waveform generators 54 and 56
?uorescent light detected at the detector 46. That informa
generate a pulse waveform at separate frequencies that feed
into the chopping circuits. Each chopping frequency is
tion is stored in the non-volatile memory of the computer 50.
When one record or measurements is completed, time
selected for optimum rejection based upon the ?uorescent
stamped and then recorded in the memory of the computer
time constance, detector response times and system through
50, the next record may then be measured immediately, or a
put requirements. The digitally controlled radio frequency
small time delay may be implemented in order to avoid
generating too much data too quickly. The process of
recording records of ?uorescent data continues until the run
of the migrating DNA fragments contained within the pair
synthesizer is connected to and controlled by the computer
50 such that the second acousto-optic tunable ?lter 68 is
commanded to pass only the optimum stimulation wave
length for the ?uorescent tag being stimulated at any
10 of parallel glass plates is complete.
moment in time.
As discussed previously, the ?oat glass plates 10 utilized
The signal that is fed into the acousto-optic tunable ?lter
68 (or alternatively, an acousto-optic modulator) to modu
to contain the gel 12 emits ?uorescence in the visible light
wavelength range which can sometimes create problems in
late the laser beam 18 can also be fed into a synchronous
detection circuit in order to implement a technique which is
accurately detecting the ?uorescence emissions 26 given off
as a result of the stimulation of the DNA fragment 16 by the
laser light beam 18. A more elegant and effective method for
reducing or eliminating this background ?uorescence is to
65
essentially the same as a phase-locked ampli?er. The chop»
ping signal can be phase adjusted as it enters the detection
circuit so that the signals from the ?uorescent tags are in
5,556,790
11
12
phase with the chopping signal. The synchronous detection
sources 54 and 56 that are of separate frequencies in order
to create one combined chopping signal which is fed into the
circuit then ampli?es only signals that are matched in phase
with the chopping signal. Such a scheme ampli?es the
wanted signal, the DNA tags, While not amplifying the
unwanted signal from the ?uorescence caused by the ?oat
synchronous demodulator 60, permits only the light that has
been processed through both of the acousto-optic tunable
?lters 34 and 68 to be effectively ampli?ed. That produces
the result of rejected the unwanted background ?uorescence
and other extraneous light from being ampli?ed. The desired
glass.
In addition, the ?rst acousto-optic tunable ?lter 34 can
also be chopped and the detection function of the apparatus
of FIG. 4 can be doubly synchronous in order to provide a
still further rejection of unwanted signals. By properly
utilizing the acousto-optic tunable ?lter 68 through which
10
nous demodulators such as element 60 are known in the art
the laser beam 18 passes, the optimum stimulation wave
length for each ?uorophore can be selected in turn in order
to provide its stimulation when it is desired to measure that
?uorophore. That reduces the emission of ?uorescence from
the other tags by not stimulating them at their optimum
light from the ?uorescence molecule at interest at any
moment in time is therefore greatly ampli?ed over any other
light or noise that may enter the detector 46.
As is known to those of ordinary skill in the art, synchro
and are alternatively termed phased-locked ampli?ers when
used with appropriate reference signals. While synchronous
15
detection has been suggested in the art for use with acousto
optic tunable ?lters, the use of dual chopping sources and the
wavelength. The rejection of unwanted signals is important
combination of their chopping signals in order to synchro
because the desired signal being measured is very small in
comparison and precise detection is important for the per
formance of the inventive apparatus. As discussed earlier,
nously detect a signal from a detector which has light
entering it from two acousto-optic tunable ?lters used in this
LI-COR, Inc. utilizes a ?uorophore that emits in the near
infrared wavelength range which eliminates the ?uorescent
emissions of the ?oat glass. The present invention can be
utilized with all wavelength ranges of ?uorescent emissions
In the event that the selection of a different wavelength for
the stimulation light for each ?uorescent tag is not needed,
than the acousto-optic tunable ?lter 68 can be replaced with
an acousto-optic modulator which does not need the second
by the ?uorophores by selecting the desired acousto-optic
manner to detect ?uorescent tag on DNA is new and novel.
25
tunable ?lter that operates in the wavelength range of the
?uorescent emissions to be detected. The acousto—optic
tunable ?lter technology readily works in the infrared range
in order to be able to handle the ?uorophores utilized by
digitally controlled radio frequency synthesizer 62 nor the
software operating on the computer 50 that controls the
operation of that radio frequency synthesizer 62. A ?xed
radio frequency is generally used with an acousto-optic
modulator such that it is switched on and o?" at the desired
30
LI-COR, as well as those utilized by others.
In order to provide the double synchronous function, a
chopping frequency.
second chopping circuit 58 is inserted between the digitally
controlled synthesizer 36 and the radio frequency power
ampli?er 38 that chops or modulates the signal output by the
digitally controlled synthesizer 36 prior to its ampli?cation
by the radio frequency power ampli?er 38. The second
chopping circuit 58 is commanded by the ?rst digital oscil
by the automated DNA sequencing system of the present
35
lator or waveform generator 54 while the second chopping
circuit 64 is commanded by the second digital oscillator or
waveform generator 56. The digital oscillators or waveform
generators 54 and 56 generate a pulse waveform at separate
frequencies that are fed to respective chopping circuits 58
45
and the chopping circuit 58, as described above to the
circuitry of FIG. 1, the light received at the detector 46 is
modulated in a distinct manner only unto the light at the
wavelength that the acoustic-optic tunable ?lter 34 is select
ing at any moment of time. When that light is detected in
synchronism with its chopping source, as described herein,
a large processing gain is realized which ampli?es only the
desired selected light and rejects all other light.
By adding the digital oscillator or waveform generator 26
and the second chopping circuit 64 to the circuitry shown in
FIG. 1, the light exiting from the acousto-optic tunable ?lter
cream a signal that is fed into the synchronous detector 60.
The combination of the outputs from the two chopping
detection signal from the Tag T4 showing a band tail.
FIG. 2B is a graph showing the emission output of the
detected ?uorophore spectra versus its wavelength for the
four tags Tl-T4. Each of the curves E-H corresponds to the
emission spectrum for the Tags T1-T4, respectively.
55
The data consisting of records of measurements at the
selected number of wavelengths is processed in a number of
steps in order to determine a DNA sequence. First, each
record of data is evaluated in order to determine the amount
60
synchronously detected, only light from the ?uorescence
tags stimulated by the chopped excitation source will be
ampli?ed and all other light will be rejected to a great extent.
The combiner 52 is used to combine the chopping signals
generated by the waveform generators 54 and 56 in order to
the detection amplitude synthesizer frequency or wave
length for each of the four ?uorophores and forms one data
record. The data in this graph indicates at data point A, a
small detection signal from the Tag T1 showing a band tail.
Data point B indicates a strong detection signal from the Tag
T2 showing a band peak. Data point C indicates that no
detection signal was detected from the Tag T3, showing that
there is no band present. Data point D indicates a small
50
68 that impinges on the gel 12 has a distinct signature at the
chopping frequency of the waveform generator 56. When
invention. FIG. 3 illustrates the band data generated from the
data record shown in FIGS. 2A and 2B that represents the
progression of bands down the polyacrylamide gel 12. As a
band of tagged fragments comes to the point of detection
where the laser beam 18 excitation spot is focused, a rise in
the signal at the characteristic wavelength for that ?uoro
phore is observed.
FIG. 2A illustrates various data points within a graph of
and 64. Each chopping frequency is selected for optimum
rejection based upon ?uorescent time constants, detector
response times and system throughput requirements.
By adding the digital oscillator or waveform generator 54
FIGS. 2A and 2B show a typical record of data generated
65
of each ?uorophore present at the point of detection at the
point and time that the record was recorded. That isv done by
?tting the spectral emission curve of each ?uorophore to the
data. Alternatively, a method which consumes less computer
processing power would be to determine the ratio of signal
strength at the peak of each ?uorophore’s emission to the
signal strengths at the valleys between the emissions. In any
event, Whatever method is utilized, a value of the amount of
?uorophore present is determined for each of the four
5,556,790
13
14
possible ?uorophores in the gel 12 for each record of the
Alternatively, a Gaussian function can be used as a good
entire data set. In order to aid in processing, four separate
approximation to the distribution of fragments in the band
data sets are then created. Each of those data sets represents
the amount of one ?uorophore present at the time of each
16. Data from a number of bands 16 in the same vicinity are
used to determine a band to band separation in a given
record of initial data. That data set is depicted in FIG. 3.
portion of the gel 12, and a band separation decrease per
base. That creams a situation similar to establishing a timing
track for bases along the gel, in a manner analogous to
FIG. _ 3 illustrates the detection amplitude versus the
record number for each of the tags T1-T4. Data point I of
FIG. 3 shows a band peak at the record number 6 for the Tag
T1, which represents the base A. Data point shown J shown
in FIG. 3 indicates a band peak at the record number 24 for
the TAG T1, which represents the base A. The reference
letter K indicates the base sequence determined by the data
10
synchronous detection of digital data from a communication
link with added noise and disturbances. Once those relations
are computed from the collected data, the determination of
the band sequence becomes more accurate because one of
the possible four bases must exist at each timing mark and
a statistical prediction can be made. That continues until the
shown, namely AGCTAC.
statistical probability of accurately determining which base
The emission spectrum of each ?uorophore is known in
advance with respect to its center frequency, lineshape,
linewidth, and percent effect. A software program is used to
implement a least-squares ?t of each of these line shape
is next in the sequence becomes unacceptably low to the
researcher. The researcher can then make his own choice as
to when to stop the analysis because the algorithm gives the
statistical level of con?dence for each base analyzed.
A diagram of the ?ow chart of the DNA sequencing
software used in connection with the preferred embodiment
of the present invention is shown in FIGS. 5A-5D. Refer~
ring ?rst to FIG. 5A, there is shown, in ?ow chart form, a
?rst portion of the DNA sequencing software used in con
nection with the present invention for entering the initial
input parameters. At step 500, the start and end wavelengths
functions to the data collected to determine the amount of
each ?uorophore present at each instant of time. Suitable
software products are readily available to perform this
function. In practice, this procedure is performed at a rate
that is determined by the speed of the detection electronics,
consistent with the needed signal-to-noise ratio of the
detected signal and the processing power of the computer
fragments.
for the digitally controlled radio frequency synthesizer 62
are input. Then, the number of wavelength steps is input at
step 502. The minimum number of wavelength steps is 4,
Obtaining a spectral scan of the ?uorophore present and
matching the known lineshapes to the collected data is a
one for each ?orescent tag, however the maximum number
may be as high as 20. The maximum or upper limit of the
50. Several times a second to once every few seconds are 25
preferred rates due to the slow speed of migration of the
number of wavelength steps is determined by such factors as
the wavelength resolution of the optical system being uti
lized, the data processing demands imposed by taking more
data points and the point of diminishing return with respect
to determining the line shape of each ?orescent tag emission.
useful technique for measuring the amount of ?uorophore
present at each point on the gel 12. When the bands 16 get
crowded together as the fragments get longer, there is a
signi?cant overlap of the bands 16 due to the diffusion
effects of migration. The data collected by the spectral scan
allows a second processing to be implemented that provides
35
signi?cantly increased band position determination.
Each band is spread out in space along the direction of
migration through the gel 12. That is a consequence of the
statistical nature of migration. With the prior art crude
detection ability, the ability to resolve which band is ahead
data records in order to not generate an amount of data that
cannot be handled or that is redundant. The amount of the
of the other was limited when the leading edge of one band
began to overlap with the trailing edge of the band ahead of
it. At that point, the bands became too blurred together and
the sequence could not be determined any further.
The next processing step of the present invention is to
Next, at step 504, the amount of pause or delay between
data records is inputted. A delay is needed because the
optical techniques used in connection with the present
invention are much faster than the migration of the DNA
fragments in the gel. Therefore, a delay of about 0.01 to
approximately 3 seconds will be necessary between taking
delay is also a function of how fast the bands run down the
gel or capillary, which will in turn depend on parameters of
45
the testing such as the voltage used. In general, the advan
tage of the present invention is in its ability to allow the
process to run as fast as it possibly can in order to improve
perform at least squares ?t of a band di?’usion function to the
data collected in the previous step. That is done for each
?uorophore so that the band centroid, or center position of
each band, can be determined with a known amount of
statistical certainty. The bands 16 are then ranked in order of
the overall throughput of DNA sequencing.
Utilizing the start and end wavelengths and the number of
wavelength steps inputted in steps in 500-504, the start
frequency, end frequency and frequency increment are com.
puted at step 506 for the digital synthesizer 62. A synthesizer
frequency needed to select a desired optical wavelength is
computed from the characteristics of the optically non-linear
material that the acousto-optic tunable ?lter is made from,
progression until they are separated by an amount that is
within a given tolerance of the computed position certainty.
At that point, the present invention ends. A signi?cant
improvement in DNA sequence determination capability is
which is known in the art.
It should again be noted that it is desirable to utilize a
realized if even 50 or 100 more bases are determined by the
present invention over existing techniques.
After the positions have been assigned on the gel for each
band on each of the four data sets of the four ?uorophores,
digital synthesizer in the preferred embodiment of this
invention. However, an analog-based frequency generator
the DNA sequence may be determined by inspecting the data
set and observing which ?uorophore is next in progression
along the gel 12. The DNA sequence is then called out in this
fashion until the position in adjacent bands becomes less
lator in which the controlling voltage is swept with a ramp
generator. In that event, that start voltage, end voltage and
voltage ramp rate would replace the start frequency, end
can also be utilized, based upon a voltage controlled oscil
than the computed position uncertainty for the bands.
The band diffusion function is determined by analysis or
through the laws of diffusion to obtain a mathematical
function that the detected ?uorophore data can be ?tted to.
frequency and frequency step information input in steps
65
500-502.
FIG. 5B shows in flow chart form the collect band data or
data acquisition portion of the DNA sequencing software
5,556,790
15
16
utilized in connection with the present invention. The soft
nation is made at step 538 then the record ?les are closed at
ware causes a digital word to be outputted to the digital
step 540 and the software proceeds to analyze the data
records and band data, the ?ow chart of which is shown in
FIG. 5C. [John—please examine steps 532—540 closely to
synthesizer in order to command the synthesizer to the
desired frequency through a standard output port or in a
manner prescribed by the interface between the computer 50
and the digital synthesizer 62. It will be obvious to those of
ordinary skill in the art that there are many ways of achiev
ing that function known in the art. It will also be obvious that
determine that the ?ow chart is correct]
The object of the portion of the DNA sequencing software
which analyzes the data records is to use the data in each of
the data records to determine the amount of ?orescent tag
present of each tag in the sample of DNA fragments at the
utilizing a standalone personal computer for the computer 50
is only an example of a computer which can be used in l0 time that each data record was taken. Curve ?tting algo
rithms of the resonance line shaped function, sometimes
connection with this invention. Obviously, any type of
termed Lorentzian, or of the common bell shape curve,
digital processing unit, such a microprocessor, microcon
termed Gaussian, are used to obtain the parameters of line
troller or other type of computer would also adequately
serve the same function.
As shown in FIG. SE, at step 516, the computer memory
15
is initialized and data ?les are opened. A ?le of data records
is created, in which each data record has a data element used
to store the detector output at each frequency. That data is
stored in the format chosen for data processing. Therefore,
position, width and height for each ?orescent tag in the
sample for each data record in the data ?le. Methods for
performing that curve ?tting are well known in the art.
At step 542, a band data ?le is created which contains the
records of data of the band concentration for each ?orescent
tag. Each record in that ?le contains a data element repre—
there are from 4 to 20 data elements for each data record and 20 senting the amount of ?orescent tag present for each tag in
the sample of DNA fragments. Since there are typically 4
the ?le of data records would be as long as the data run
tags, there will be 4 data elements per data record. Step 542,
requires, typically thousands of data records. In addition, a
the band data ?le pointers are initialized, the data record ?le
data record counter and pointers to the data record ?le are
initialized and the ?rst word to the digital synthesizer 62 is
generated. The ?rst word to the digital synthesizer 62 is the
word that commands the synthesizer to select the start
wavelength on the ?orescence spectrum. The software pre
pares to output that word, and then goes to step 518.
At step 518, the DNA sequencing software outputs the
?rst digital word to the synthesizer, which commands the
synthesizer 62 to move to the frequency of the wavelength
desired at that point in time. The program then waits for a
small delay at step 520 for the electronics to settle and for
the signal generated by the detector electronics to stabilize.
is opened and the pointers for the opened data ?le are
initialized.
25
Then, at step 544, for each record of the data record ?le
accumulated in step 542, a least squares ?t of a curve
30
That delay is typically from a few to a few hundred 35
containing a height, width and position for each ?orescent
emission in the DNA samples is performed. At step 544, a
value is obtained for the height, width and position for the
emission of each ?orescent tag in the data record.
At step 546, for each emission line, the emission level is
integrated over the span of the data record in order to obtain
a value of the total emissions for each ?orescent tag. That
value is stored in the data element for that ?orescent tag in
the corresponding record of the band data ?le.
A determination is then made, at step 548, of whether the
end of the data record has been reached. If a negative
determination is made at step 548, then the program returns
to step 544 and continues to execute from that point until
each data record has been processed to obtain a value of the
total emission for each ?orescent tag at each data record
microseconds.
At step 522, the signal from the detector electronics is
input into the DNA sequencing program by means of an
analog-to-digital converter interfaced to the computer 50.
The inputted data word from the detector electronics 46, 48 40
is converted at step 524 into a convenient internal format for
data analysis, such as a ?oating point format. The data word
initially recorded.
is then stored in the computer memory at step 526 in a data
record at the element number associated with that particular
If an a?irmative detemiination is made at step 548, then
frequency step. The data record data element pointer is then
all of the ?les are closed at step 550 and the program
incremented by one at step 528 and a frequency increment
proceeds to the analyzing of band data portion of the DNA
is added to the present frequency to be output to the digital
sequencing software, the ?ow chart of which is shown in
synthesizer 62 at step 530.
steps 552-564.
At step 532, a determination is made of whether the
After executing step 550, a data ?le of band data repre
frequency to be output to the digital synthesizer 62 as
senting the concentration of each ?orescence tag at each
generated at step 530 is the end frequency. If a negative
point in time that data was recorded has been generated. By
determination is made at step 532, then the program returns
analyzing data for a single ?orescent tag, the data will reveal
to step 518 and outputs the current digital word (frequency)
a series of rises and falls that resemble a sequence of humps
to the digital synthesizer 62 and then continues on at step 55 or hills. Those are the bands with that tag passing by the
520 as previously described.
detector in time. A position is then assigned to the center of
each those hills or bands for each of the four ?orescent tags.
If an affirmative determination is made at step 532, that
A least squares curve ?tting process is used with a Gaussian
means that a data record has been completed. In that event,
line shape in order to determine the center position.
the data record pointer is incremented at step 534 and the
digital output word is reset to the selected start wavelength 60
At step 532, four band position data ?les are opened. Each
at step 536. Then, the software waits for a period of time at
data ?le consists of records with two data elements, one
step 537 equal to the delay inputted at step 504.
After the delay at step 537, a determination is made of
representing the position and the other representing the
position uncertainty of each band. Also, data pointers and
record counters are initialized.
whether the end of the run has been reached at step 538. If
a negative determination is made at step 538, then the 65
The ?rst data element of each record in the band data e is
program returns to begin repeating the steps described
then examined at step 554. That data element represents the
above, beginning with step 518. If an a?irrnative determi
concentration of the ?rst ?orescent tag at each point in time.
5,556,790
17
18
At step 556, the data corresponding to one band passing
through the detector is read in. That data is determined by a
which band arrived ?rst, then second, and so forth, will
properly determine the base sequence of the DNA fragment.
coarse examination of the data to discern a peak and a valley
in the data. Then, a least squares ?t of that section of data to
the Gaussian curve is performed at step 558. The center
Although the foregoing describes the basic steps for
processing the data produced by the automated DNA
sequencing system of the present invention, many enhance
position, width, height and position uncertainty is calculated
ments may be made to the processing steps. In addition, for
purposes of clarity, a few well-known details have been
from that ?t and recorded in the corresponding data record.
A determination of whether the end of the data ?le has
been reached is then made at step 560. If a negative
determination is made at step 560, then the program returns
omitted. Such details would include, for example, subtract
in g background noise, correcting for the relative ?uorophore
emission strengths, correcting for the relative ?uorophore
migration rates, removing the ?uorescence emissions pro
duced by the glass plates 10 or the gel 12 themselves,
and begins executing again starting with step 556.
If an a?irmative determination is made at step 560, then
the next data element in the band ?le is examined at step 562
removing the Rarnan scattering line from water, as well as
a few other extraneous factors. However, such details do not
and the data corresponding to that band passing through the
detector is then read in at step 556 such that the processing
of the next ?orescent tag is performed.
Prior to reaching step 562, after an af?rrnative determi~
alter the operation of the disclosed automated DNA sequenc
ing system.
As has been described previously, some automated DNA
sequencing systems utilize a single ?orescence tag because
of the di?iculty in accurately measuring the amount of each
nation at step 560, a determination is made at step 564 of
whether all of the florescent tags have been processed. If a
negative determination is made at step 564, then step 562 is
20
executed. If an a?irmative determination is made at step 562
then the DNA sequencing software program proceeds to the
tag present when performing multiple ?orescence tag
sequencing. The problem arises due to the overlapping
spectral output of the ?orescent emissions which makes the
signal of one ?orescence molecule ride of the skirt of an
adjacent one. That in turn gives rise to errors in the deter—
next processing step of calling bases. [John—please check
steps 554—564 of FIG. SC to determine that I have accurately
mination of the quantity of each ?orescence tag present.
drawn then on the ?ow chart]
25
If a band of DNA consisting of fragments labeled with one
When the DNA sequencing software program reaches the
?orescence molecule happens to follow closely behind a
calling bases portion of the program, the ?ow chart of which
band tagged with another molecule, as will often happen, the
is shown in FIG. 5D, four data ?les that contain positions of
emissions of the previous band will tend to raise the optical
bands for their respective ?orescence tags have been gen
detection baseline for the measurement of the present band.
erated. While this portion of the DNA sequencing software 30 Such effects will cause errors in the measurements and make
the technique break down sooner than could be realized with
can be implemented in numerous ways, it is important to
the present invention disclosed herein. Thus, the use of the
note that this portion of the DNA sequencing software is able
present invention with a single ?orescence tag will likewise
to perform its function due to the ability of the acousto-optic
provide an increase in speed of DNA sequencing.
tunable ?lters to generate data which is good enough to
allow the structure of the ?orescence data coming from 35
Also, as brie?y described herein, capillary methods for
multiple ?orescence tags to be utilized to obtain good
sequencing DNA fragments may be utilized in place of ?at
measurements of each ?uorophore and the precise center
gel techniques. The apparatus of the present invention will
position of each band.
operate in a similar manner when used with capillary
methods instead of ?at gel apparatus. The capillaries appear
The determination of the base sequence of the fragment of
as a strand of material and the mixture of tagged DNA
DNA being analyzed is performed as follows by the DNA
fragments are introduced at one end. The fragments are
sequencing program utilized in connection with the opera
made to travel along the strand and will separate in size, with
tion of the present invention.
the shortest fragments travelling faster than the longer ones.
At step 508, the band position data ?les are examined. The
The
capillary methods techniques can be thought of as a thin
?rst data record in each ?le is looked at. The ?le whose band 45
strand of material that the tagged DNA fragments are being
anived ?rst, or that has the closest position to the start is
made to travel through instead of the flat gel.
determined to be the ?rst base in the sequence. That is
Band separation in the strand occurs in a manner analo
declared a valid base call if the position of uncertainty
gous to that which occurs when using the ?at gel techniques.
between that base and the next closest one does not suggest
that those two bases overlap. If there is a suggestion that they 50 Excitation laser beam and detection optics are then focused
on a point on the strand and the passage of the bands is
overlap, the sequence has been determined to the maximum
measured in a manner similar to that with a gel.
extent possible. Therefore, that is the end of the sequence
determination.
Although only a preferred embodiment is speci?cally
illustrated and described herein, it will be appreciated that
At step 510, the base being examined is taken out of the
data ?le and the remaining bases in that base ?le are shifted 55 many modi?cations and variations of the present invention
are possible in light of the above teachings and within the
down by one so that the next base in that ?le takes the place
purview of the appended claims without departing from the
of the ?rst base that has been called out. A determination is
spirit and intended scope of invention.
then made at step 512 of whether the end of the data has been
What is claimed is:
reached. If a negative determination is made at step 512,
1. A method for automating the sequencing of DNA
meaning that all of the bases have not been called out, then
fragments, comprising the steps of:
steps 508 and 510 are repeated again. If an af?rmative
determination is made at step 512, then the DNA sequencing
tagging each of a predetermined plurality of DNA frag
software ends at step 514.
ments ending in a different base with a different ?uo
rescent tag;
.
As will be obvious to those of ordinary skill in the art,
there are many alternate methods by which to determine the
sequence in which the bands arrived at the detector. Any
method that examines the band position data and determines
65
stimulating the emission of light at its characteristic
wavelength from each of said ?uorescent tagged plu
rality of DNA fragments;
5,556,790
20
19
a predetermined setting by said digital data processor depen
dent upon the expected wavelength of the emission of
?uorescent light by each of said ?uorescent tagged plurality
of DNA fragments.
focusing the emission of ?uorescent light from each of
said ?uorescent tagged plurality of DNA fragments as
substantially parallel rays of light onto an acousto-optic
tunable ?lter;
controlling the operation of said acousto-optic tunable
5
only light at a predetermined wavelength is di?'racted
10. The method of claim 8, further including the steps of:
modulating said beam of light passing through said sec
ond acousto-optic tunable ?lter;
modulating said diffracted light which exits said ?rst
acousto-optic tunable ?lter; and
synchronously detecting said modulated diffracted light in
and exits from said acoustic-optic tunable ?lter at at
least one angle to a main beam of light exiting from
said acousto-optic tunable ?lter; and
detecting at least one diffracted light beam exiting said
acoustic-optic tunable ?lter to produce an electrical
representation of said ?uorescent light emissions of
said plurality of DNA fragments.
2. The method of claim 1, further including the step of
storing said electrical representations of said ?uorescent
said detecting step.
15
light emissions of said plurality of DNA fragments.
3. The method of claim 1, wherein said step of controlling
the operation of said acousto-optic tunable ?lter includes
ter;
combining said ?rst and second modulating signals to
form a combined modulating signal; and
using said combined modulating signal to synchronously
fragments, comprising the steps of:
tagging each of a predetermined plurality of DNA frag
detect said modulated diffracted light in said detecting
step.
12. A method for automating the sequencing of DNA
ments to be sequenced ending in a different base with
a different ?uorescent tag;
fragments, comprising the steps of:
tagging each of a predetermined plurality of DNA frag
generating a beam of light for stimulating the emission of
?uorescent light from each of said ?uorescent tagged
ments identi?ed by a different predetermined charac
plurality of DNA fragments;
optimizing the wavelength of the stimulating beam of
teristic with a different ?uorescent tag;
stimulating the emission of light at its characteristic
wavelength from each of said ?uorescent tagged plu
light prior to its impingement on each of said ?uores
35
focussing the emission of ?uorescent light from each of
said ?uorescent tagged plurality of DNA fragments as
substantially parallel rays of onto a ?rst acousto-optic
tunable ?lter;
controlling the operation of said acousto-optic tunable
40
?lter under control using a digital data processor such
that only light at a predetermined wavelength is dif—
fracted and exits from said acoustic-optic tunable ?lter
only ?ght at a predetermined wavelength is diffracted
and exits from said acoustic-optic tunable ?lter at at
least one angle to a main beam of light exiting from
said acousto-optic tunable ?lter; and
detecting at least one diffracted light beam exiting said
acoustic-optic tunable ?lter to produce an electrical
representation of said ?uorescent light emissions of
from said acousto-optic tunable ?lter; and
detecting at least one diffracted light beam exiting said
acoustic-optic tunable ?lter to produce an electrical
sequentially stepping said acousto-optic tunable ?lter
rality of DNA fragments;
focusing the emission of ?uorescent light from each of
said ?uorescent tagged plurality of DNA fragments as
substantially parallel rays of light onto an acousto-optic
tunable ?lter;
controlling the operation of said acousto-optic tunable
?lter under control of a digital data processor such that
at at least one angle to a main beam of light exiting
representation of said ?uorescent light emissions of
said plurality of DNA fragments.
5. The method of claim 4, further including the step of
storing said electrical representations of said ?uorescent
light emissions of said plurality of DNA fragments.
6. The method of claim 4, wherein said step of controlling
the operation of said acousto-optic tunable ?lter includes
11. The method of claim 10, wherein said synchronously
detecting step includes the steps of:
generating ?rst and second modulating signals suitable for
modulating said diffracted light exiting from said ?rst
acousto-optic tunable ?lter and said beam of light
passing through said second acousto-optic tunable ?l
sequentially stepping said acousto-optic tunable ?lter
through a plurality of wavelengths, wherein each of said
plurality of wavelengths corresponds to a different one of
said ?uorescent tags.
4. A method for automating the sequencing of DNA
cent tagged plurality of DNA fragments;
9. The method of claim 8, further including the step of
modulating said beam of light passing through said second
acousto-optic tunable ?lter.
?lter under control of a digital data processor such that
said plurality of DNA fragments.
13. The method of claim 12, further including the step of
storing said electrical representations of said ?uorescent
?ght emissions of said plurality of DNA fragments.
14. The method of claim 12, wherein said step of con
55
trolling the operation of said acousto-optic tumble ?lter
includes sequentially stepping said acousto-optic tunable
through a plurality of wavelengths, wherein each of said
?lter through a plurality of wavelengths, wherein each of
plurality of wavelengths corresponds to a different one of
said ?uorescent tags.
said plurality of wavelengths corresponds to a different one
of said ?uorescent tags.
15. A method for automating the sequencing of DNA
7. The method of claim 4, further including the steps of:
modulating said diffracted light which exits said ?rst
acousto-optic tunable ?lter; and
synchronously detecting said modulated diffracted light in
said detecting step.
8. The method of claim 4, wherein said step of optimizing
comprises the step of ?ltering said beam of light through a
second acousto-optic tunable ?lter which is commanded to
60
fragments, comprising the steps of:
tagging each of a predetermined plurality of DNA frag
ments identi?ed by a different predetermined charac
teristic with a different ?uorescent tag;
65
generated a beam of light for stimulating the emission of
?uorescent light from each of said ?uorescent tagged
plurality of DNA fragments;
5,556,790
22
21
optimizing the wavelength of the stimulating beam of
19. The method of claim 15, wherein said step of opti
light prior to its impingement on each of said ?uores
mizing comprises the step of ?ltering said beam of light
cent tagged plurality of DNA fragments;
through a second acousto-optic tunable ?lter which is com‘
manded to a predetermined setting by said digital data
processor dependent upon the expected wavelength of the
emission of ?uorescent light by each of said ?uorescent
focussing the emission of ?uorescent light from each of
said ?uorescent tagged plurality of DNA fragments as
substantially parallel rays of light onto a ?rst acousto
tagged plurality of DNA fragments.
optic tunable ?lter;
controlling the operation of said acousto~optic tunable
20. The method of claim 19, further including the step of
modulating said beam of light passing through said second
acousto-optic tunable ?lter.
?lter under control using a digital data processor such
that only light at a predetermined wavelength is dif
21. The method of claim 19, further including the steps of:
modulating said beam of light passing through said second
acousto-optic tunable ?lter;
modulating said diffracted light which exits said ?rst
acousto-optic tunable ?lter; and
synchronously detecting said modulated diffracted light in
said detecting step.
22. The method of claim 21, wherein said synchronously
detecting step includes the steps of:
generating ?rst and second modulating signals suitable for
modulating said diffracted light exiting from said ?rst
acousto-optic tunable ?lter and said beam of light
passing through said second acousto~optic tunable ?l~
fracted and exits from said acoustic-optic tunable ?lter
at at least one angle to a main beam of light exiting
from said acousto-optic tunable ?lter; and
detecting at least one diffracted light beam exiting said
acoustic-optic tunable ?lter m produce an electrical
representation of said ?uorescent light emissions of
said plurality of DNA fragments.
16. The method of claim 15, further including the step of
storing said electrical representations of said ?uorescent
light emissions of said plurality of DNA fragments.
17. The method of claim 15, wherein said step of con
trolling the operation of said acousto-optic tunable ?lter
includes sequentially stepping said acousto~optic tunable
?lter through a plurality of wavelengths, wherein each of
said plurality of wavelengths corresponds to a different one
of said ?uorescent tags.
18. The method of claim 15 further including the steps of:
modulating said di?cracted light which exits said ?rst
acousto-optic tunable ?lter; and
synchronously detecting said modulated di?racted light in
said detecting step.
25
30
ter;
combining said ?rst and second modulating signals to
form a combined modulating signal; and
using said combined modulating signal to synchronously
detect said modulated diffracted light in said detecting
step.
UNITED STATES PATENT AND TRADEMARK OFFICE
CERTIFICATE OF CORRECTION
PATENT-NO. I
5’556’79O
DATED
September 17', 1996
:
INVENTOR(S)I
John W. Pettit
It is certi?ed that error appears in the above-indentified patent and that said Letters Patent is hereby
corrected as shown below:
CQLZO, Claim 15, line 6, change "generated" to --generating--.
Signed and Sealed this
Seventeenth Day of December, 1996
BRUCE LEHMAN
Arresting O?‘icer
Commissioner of Palems and Trademarks
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement