U.S. patent application number 16/397932 was filed with the patent office on 2019-08-22 for rapid detection of viable bacteria system and method.
The applicant listed for this patent is The Curators of the University of Missouri. Invention is credited to Hsueh-Chia Chang, Sachidevi Puttaswamy, Shramik Sengupta.
Application Number | 20190256886 16/397932 |
Document ID | / |
Family ID | 43823457 |
Filed Date | 2019-08-22 |
![](/patent/app/20190256886/US20190256886A1-20190822-D00001.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00002.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00003.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00004.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00005.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00006.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00007.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00008.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00009.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00010.png)
![](/patent/app/20190256886/US20190256886A1-20190822-D00011.png)
View All Diagrams
United States Patent
Application |
20190256886 |
Kind Code |
A1 |
Sengupta; Shramik ; et
al. |
August 22, 2019 |
RAPID DETECTION OF VIABLE BACTERIA SYSTEM AND METHOD
Abstract
An improved system and method is provided for detecting viable
bacteria in a suspension sample. A sample of a suspension in which
bacterial presence is suspected is collected from a source and a
portion of the sample transferred to a microfluidic unit. A series
of analysis signals at different frequencies are applied to the
sample portion. An impedance is measured via a signal analyzer for
the sample portion for each of the analysis signals to define an
impedance data set. An initial bulk capacitance value is determined
for a model circuit based on the impedance dataset. After a
predetermined time period, a new bulk capacitance value is
determined for another portion of the sample. The difference
between the new bulk capacitance and the initial bulk capacitance
value is compared to a threshold value to determine if viable
bacterial is present in the sample.
Inventors: |
Sengupta; Shramik;
(Columbia, MO) ; Puttaswamy; Sachidevi; (Columbia,
MO) ; Chang; Hsueh-Chia; (Granger, IN) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
The Curators of the University of Missouri |
Columbia |
MO |
US |
|
|
Family ID: |
43823457 |
Appl. No.: |
16/397932 |
Filed: |
April 29, 2019 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
14099110 |
Dec 6, 2013 |
10273522 |
|
|
16397932 |
|
|
|
|
12896188 |
Oct 1, 2010 |
8635028 |
|
|
14099110 |
|
|
|
|
61278156 |
Oct 2, 2009 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G01N 33/5438 20130101;
B01L 3/502715 20130101; C12Q 1/04 20130101; C12Q 1/06 20130101 |
International
Class: |
C12Q 1/06 20060101
C12Q001/06; G01N 33/543 20060101 G01N033/543; C12Q 1/04 20060101
C12Q001/04 |
Claims
1. A non-transitory computer-readable medium encoded with a
bacteria detection instructions stored thereon that, when executed
by a computing device cause the computing device to perform
operations for detecting viable bacteria in a sample of a
suspension, the operations comprising: activating a signal
generator to generate a series of analysis signals to apply to a
portion of a particular sample in response to an initial analysis
request received from a user interface; each of the series of
analysis signals being generated at a different frequency;
activating a signal analyzer to generate an initial impedance data
set for the particular sample by determining an impedance of the
particular sample during application of each of the series of
analysis signals; activating the signal generator to generate
another series of analysis signals to apply to another portion of
the particular sample, each of the other series of analysis signals
being generated at the different frequency; and activating the
signal analyzer to generate a new impedance data set for the
particular sample by determining the impedance of the particular
sample during application of each of the other series of analysis
signals; determining at least one initial parametric value of a
model circuit based on the initial impedance data set, the at least
one initial parametric value comprising an initial impedance
parameter value and an initial confidence interval value; and
determining at least one new parametric value of the model circuit
based on the new impedance data set, the at least one initial
parametric value comprising an new impedance parameter value and a
new confidence interval value; determining if the initial
confidence interval value and the new confidence interval value
overlap; and generating a positive result to indicate that viable
bacteria is present when the initial confidence interval value and
the new confidence interval value do not overlap; and generating
the positive result for display.
2. The non-transitory computer-readable medium of claim 1 further
comprising: retrieving pre-determined time interval data for the
particular sample from a data source, the pre-determined time
interval data comprising a corresponding pre-determined time
interval comprising at least one member selected from a group
consisting of a minimum doubling time of an expected bacteria type
in the suspension and a finite period of time defined by a user;
and initiate generation of another analysis request after
expiration of the corresponding pre-determined time interval; and
generating a negative result to indicate that viable bacteria is
not present when the initial confidence interval value and the new
confidence interval value overlap.
3. The non-transitory computer-readable medium of claim 1 wherein
the operations further include initiating the generation of the
other analysis request by generating a notification request to
notify a user to generate the other analysis request at the user
interface after expiration of the corresponding pre-determined time
interval.
4. The non-transitory computer-readable medium of claim 1 wherein
the operations further include initiating the generation of the
other analysis request by automatically generating the other
analysis request after expiration of the pre-determined time
interval.
5. The non-transitory computer-readable medium of claim 2 wherein
the data source further comprises a maximum processing time
defining a maximum amount of time for attempting to detect viable
bacteria in the particular sample.
6. The non-transitory computer-readable medium of claim 5 wherein
the operations further comprise: automatically generating a second
other analysis request in response to the negative result after
expiration of the corresponding pre-determined time interval if the
maximum processing time has not expired; and generating the
negative result for display if the maximum processing time has
expired.
7. The non-transitory computer-readable medium of claim 5 wherein
the operations further comprise: generating another notification to
notify the user to generate a second other analysis request at the
user interface in response to the negative result after expiration
of the corresponding pre-determined time interval if the maximum
processing time has not expired; and generating the negative result
for display if the maximum processing time has expired.
8. The non-transitory computer-readable medium of claim 1 wherein:
the at least one initial parametric value of the model circuit
comprises an initial magnitude of a Constant Phase Element; and the
at least one new parametric value of the model circuit comprises a
new magnitude of the Constant Phase Element.
9. The non-transitory computer-readable medium of claim 1 wherein
the series of analysis signals comprises at least one member
selected from a group consisting of voltage signals and current
signals.
10. The non-transitory computer-readable medium of claim 1 wherein
the plurality of suspension types comprises at least one member
selected from a group consisting of a bodily suspension, a food
product suspension, and a non-food product suspension.
11. A method for detecting viable bacteria in a sample of a
suspension, the method comprising: generating a series of analysis
signals at a signal generator to apply to a portion of a particular
sample in response to an initial analysis request received from a
user interface; each of the series of analysis signals being
generated at a different frequency; generating an initial impedance
data set at a signal analyzer for the particular sample by
determining an impedance of the particular sample during
application of each of the series of analysis signals; generating
another series of analysis signals at the signal generator to apply
to another portion of the particular sample, each of the other
series of analysis signals being generated at the different
frequency; and generating a new impedance data set at the signal
analyzer for the particular sample by determining the impedance of
the particular sample during application of each of the other
series of analysis signals; determining at least one initial
parametric value of a model circuit at a processor based on the
initial impedance data set, the at least one initial parametric
value comprising an initial impedance parameter value and an
initial confidence interval value; and determining at least one new
parametric value of the model circuit at the processor based on the
new impedance data set, the at least one initial parametric value
comprising a new impedance parameter value and a new confidence
interval value; determining if the initial confidence interval
value and the new confidence interval value overlap at the
processor; and displaying a positive result indicating that viable
bacteria are present when the initial confidence interval value and
the new confidence interval value do not overlap.
12. The method of claim 11 further comprising: retrieving
pre-determined time interval data for the particular sample from a
data source, the pre-determined time interval data comprising a
corresponding pre-determined time interval comprising at least one
member selected from a group consisting of a minimum doubling time
of an expected bacteria type in the suspension and a finite period
of time defined by a user; and initiating generation of another
analysis request after expiration of the corresponding
pre-determined time interval.
13. The method of claim 12 further comprising initiating the
generation of the other analysis request by generating a
notification request to notify a user to generate the other
analysis request at the user interface after expiration of the
corresponding pre-determined time interval.
14. The method of claim 12 further comprising initiating the
generation of the other analysis request by automatically
generating the other analysis request after expiration of the
pre-determined time interval.
15. The method of claim 12 further comprising; retrieving a maximum
processing time from the data source, the maximum processing time
defining a maximum amount of time for attempting to detect viable
bacteria in the particular sample; and generating a negative result
to indicate that viable bacteria are not present when the initial
confidence interval value and the new confidence interval value
overlap and the maximum processing time has not expired.
16. The method of claim 15 further comprising: generating another
notification to notify the user to generate a second other analysis
request at the user interface in response to the negative result
after expiration of the corresponding pre-determined time interval
if the maximum processing time has not expired; and displaying the
negative result if the maximum processing time has expired.
17. The method of claim 15 further comprising: automatically
generating a second other analysis request in response to the
negative result after expiration of the corresponding
pre-determined time interval if the maximum processing time has not
expired; and displaying the negative result if the maximum
processing time has expired.
18. The method of claim 11 wherein: the at least one initial
parametric value of the model circuit comprises an initial
magnitude of a Constant Phase Element; and the at least one new
parametric value of the model circuit comprises a new magnitude of
the Constant Phase Element.
19. The method of claim 11 wherein the series of analysis signals
comprises at least one member selected from a group consisting of
voltage signals and current signals.
20. The method of claim 11 wherein the plurality of suspension
types comprises at least one member selected from a group
consisting of a bodily suspension, a food product suspension, and a
non-food product suspension.
Description
FIELD
[0001] The present invention relates to a system and method for
detecting the presence of viable bacteria in suspensions. More
specifically, the invention relates to a system and method for
detecting viable bacteria in a suspension based on changes in
impedance induced by the proliferation of bacteria in the
suspension.
RELATED APPLICATIONS
[0002] This application is a continuation of U.S. patent
application Ser. No. 14/099,110, filed Dec. 6, 2013, entitled Rapid
Detection of Viable Bacteria System and Method, which is a
continuation of U.S. patent application Ser. No. 12/896,188, filed
Oct. 1, 2010, entitled Rapid Detection of Viable Bacteria System
and Method, which claims priority to U.S. Patent Application No.
61/278,156, filed Oct. 2, 2009, and entitled Rapid Detection Method
For Viable Bacteria, the entire contents of which are incorporated
herein by reference.
BACKGROUND
[0003] The process of pasteurization involves heating liquid food
products like milk, juices, etc. to kill harmful organisms such as
viruses, bacteria, molds, and yeast. However, some amount of
bacteria may survive the pasteurization process or may be
inadvertently introduced during further processing. Such bacteria
typically causes spoilage of food products and has been estimated
to cause economic losses of $1 billion each year. Moreover, if the
surviving bacteria are pathogenic, outbreaks of food borne
illnesses may occur among consumers who assumed that the food
product was risk-free since it had been pasteurized. In the United
States alone, it has been estimated that approximately 76 million
food borne illnesses occur per year. It has also been estimated
that such illnesses result in up to 5000 deaths and have an adverse
economic impact of $6.5-$34.9 billion each year.
[0004] Detecting and quantifying bacteria that survive treatments
such as pasteurization is an important step in assuring food
quality and food safety and in complying with standards set by
appropriate governing bodies or trade organizations. For instance,
the United States Pasteurized Milk Ordinance requires "Grade A"
pasteurized milk to have a total bacterial count of .ltoreq.20,000
colony forming unit (CFU)/ml and a coliform count of .ltoreq.10
CFU/ml. As a consequence, those who produce and/or market food
products have to perform microbiological tests to satisfy
themselves, and the governing bodies, regarding the efficacy of
their processes designed to keep the numbers of bacteria within the
stipulated range. It is important to their economic operation that
they do so with the least possible expenditure of resources
(material and labor).
[0005] There are presently several ways to detect bacteria in
liquid samples like milk and juice. They can be broadly classified
into three broad classes: (a) traditional methods such as plate
cultures and biochemical assays, (b) DNA and antibody based
methods, often involving micro/nano particles and fluorescence, (c)
other "automated" techniques that rely on monitoring the effects of
bacterial metabolism on the medium. Of these, traditional methods
are the most extensively used, and often serve as the standard to
which other techniques are compared. However, such traditional
methods are tedious, labor intensive, and require very long times
to detect bacteria, which can range from overnight to weeks
depending on the type of the organism and medium used.
[0006] DNA and antibody based methods overcome many of the
disadvantages of the traditional methods. They are rapid, require
less reagents and labor, and are able to identify the
species/strain of the bacteria present relatively easily. However,
DNA and antibody based methods cannot distinguish between viable
and dead bacteria, and hence their applicability in many situations
(such as that described earlier) is limited.
[0007] The commercially available automated methods include devices
such as the Bactec.TM. that detects the amount of radio-labeled
carbon dioxide released, Coli-Check.TM. swabs that use Bromocresol
Purple as an indicator to measure the decrease in pH due to
bacterial metabolism, and the Bactometer.TM. (Bactomatic Ltd.),
Malthus 2000.TM. (Malthus Instruments Ltd.) and RABIT.TM. (Don
Whitley Scientific Ltd.) systems, that use electrical impedance. A
summary of various automated methods already commercialized, and
the times to detection ("TTD") for these methods (for various
mentioned initial loads) are given in Table 1.
TABLE-US-00001 TABLE 1 Summary of Existing Automated Methods
Commercial name Method employed Initial load Microorganisms TTD
RABIT (Don Change in solution 1 CFU/ml coliforms 16.1 hrs Whitley
Scientific conductance Ltd., Shipley, UK) Bactometer (Bio Impedance
>10.sup.5 CFU/ml Mainly E. coli 4 hours Merieux, microbiology
Nuertingen, Germany) Malthus systems Conductance 100 CFU/ml C.
Sporogenes 15.5 hrs (Malthus change of the fluid Instruments Ltd.,
Crawley, UK) BacTrac (Sylab) Impedance 100 CFU/ml P. Aeruginosa 30
hours analyzer
[0008] The common underlying feature of these techniques, including
those which use electrical impedance, is based on bacterial
metabolism to produce a discernible change in a material property
of the medium (such as pH, optical density, amount of carbon
dioxide dissolved, electrical conductivity). The amount of
metabolite processed by an individual bacterium is extremely small.
Hence, there has to be a sufficiently large number of bacteria
present (either a priori or arising due to proliferation from the
smaller number initially present) before the signal generated
(change in the material property of the suspension) can be
effectively measured. If the bacterial count in the original
suspension happens to be small (1000 CFU/ml or lower), one must
wait for cells to proliferate to an appropriately high number
(often .about.10.sup.6 CFU/ml or greater) before a discernible
change in the physical properties of the medium (such as pH,
O.sub.2 /CO.sub.2 concentration, conductivity, etc.) can be
noticed. Thus, for low initial loads, current commercial automated
systems take almost as long as the plate cultures (overnight or
longer) to provide the desired result.
[0009] Recently, there have been efforts to increase the ease of
handling, cut costs, and most importantly, reduce TTDs by using
microfluidic systems to miniaturize the automated methods. For
example, chip-based micro-devices have been developed in which the
pH and impedance of a sample contained therein are monitored in
order to detect bacterial metabolism, and various additional
modifications like the use of interdigitated microelectrodes, and
arrays of microelectrode based biosensors have been tried in order
to increase the sensitivity of measurements (with respect to
conventional electrodes), and thus further decrease the TTD. While
these efforts were successful in the sense that their TTDs are
lower than those of the commercially available devices, they
continue to be limited by the amount of time it takes for bacterial
metabolism to significantly alter the composition of the medium
when bacterial loads are low. One method previously attempted to
overcome this drawback involved concentrating the bacterial cells
from dilute samples to a small volume by using dielectrophoresis
(DEP) prior to culture, and then detecting changes in medium
composition as before. While the culture time needed for detection
was reduced, one needs to take into account the time needed for
concentration using DEP (an additional 2-3 hours) as well to get
effective TTDs. Again, while successful, the actual method of
detection still relies on bacterial metabolism, with its inherent
limitations (as discussed earlier).
[0010] Therefore, there is a need to provide a new and improved
method to detect viable bacteria in a suspension based on the
changes of capacitance of the suspension due to the bacteria
proliferation. There is another need to provide a new system to
detect viable bacteria in a suspension based on the changes of
capacitance of the testing suspension.
SUMMARY
[0011] The invention provides a new and improved method to detect
viable bacteria and its proliferation in a suspension based on the
changes in the capacitance of material in the interior bulk of the
suspension. According to one embodiment of the invention, the
detection method includes the steps of 1) incubating a suspension
containing viable bacteria 2) obtaining impedance measurements at
multiple pre-determined frequency to obtain a parameter (using a
novel theoretical electronic circuit model) that reflects the
amount of charge stored in the interior bulk of the suspension and
repeating this after pre-determined intervals of time and 3)
analysing changes in the value of the obtained parameter with time
to infer the presence, or lack thereof, of viable bacteria in the
suspension.
[0012] The invention also provide a detection system to acomplish
the inventive method. The detection system composed of 1) a
microfluidic channel with defined geometric properties having
electrodes on its two ends, wherein a testing suspension may be
injected into the channel, 2) an impedance mesuring device to
obtain impedances of the testing suspension at multiple frequencies
and different time intervals, and 3) a data analysis means to
analyzing the impedances and obtaining the parameter of the
interest based on the related circuit model.
DESCRIPTION OF DRAWINGS
[0013] FIG. 1A is a block diagram of an emplary computing
environment for implementing a viable bacteria detection
system.
[0014] FIG. 1B is a block diagram that depicts an exemplary viable
bacteria detection system.
[0015] FIG. 1C illustrates a method for detecting viable bacteria
in fluid sample in accordance with an aspect of the viable bacteria
detection system.
[0016] FIG. 2A depicts of a microfluidic unit with electrodes on
either end loaded with suspension harboring bacteria.
[0017] FIG. 2B is an exemplary model circuit representation of the
impedance components of a microfluidic unit harboring bacteria.
[0018] FIG. 3A shows another exemplary equivalent representation of
the impedance components of a microfluidic unit harboring
bacteria.
[0019] FIG. 3B is a Cole Plot of Resistance (Z') on the x-axis
against Reactance (Z'') on the y-axis.
[0020] FIGS. 3C and 3D are plots of the same data shown in FIG. 3B
plotted as magnitude of impedance (|Z|) (3C) and phase angle
(.theta.) (3D) as functions of frequency.
[0021] FIG. 4 is a plot showing the increase in a calculated CPE-T
parameter calculated by along with actual increase in the
concentration of the bacteria in the suspension.
[0022] FIG. 5 shows the plot of CPE-T values and actual
concentration of bacteria in the sample at various points in time
for a system consisting of E. coli suspended in Tryptic Soy Broth
(TSB).
[0023] FIGS. 6A-6I are CPE-T v/s time plots for some representative
samples with different initial bacterial loads.
[0024] FIGS. 7A-7C are the consolidated plots showing the variation
of the Time to Detection (TTD) as a function of the initial
bacterial load for multiple experiments with E. coli in Tryptic Soy
Broth (left), E. coli in milk (center), and Lactobacillus in apple
juice (right)
[0025] FIGS. 8A and 8B compares TTDs obtained using the present
invention (solid lines) to (A) those of the commercial systems
already on the market, and (B) other, especially microfluidic,
systems under development (dashed lines).
DETAILED DESCRIPTION OF INVENTION
[0026] The present invention provides a new and improved method for
detecting viable bacteria in suspensions, such as fluid food
products, blood samples, and environmental water samples, and other
liquid media samples. Unlike the existing metablism based methods
used to detect bacteria, the inventive method is based on the
ability of viable bacteria to store electric charge. As the number
of viable bacteria in a suspension increases due to the
reproduction of the previously existing viable bacteria, the charge
carrying capacity of the particular suspension as a whole (its bulk
capacitance) increases. The inventive method is designed to magnify
the effect at measureable frequencies (<100MHz) and apply the
inventive data analysis to filter out other effects such as change
in temperature that can affect measured impedance values. Employing
the inventive methods, the bacterial proliferation in liquid media
may be detected much faster (about 4 to 10 times) than existing
methods.
[0027] The present invention also provides a new and improved
system for rapid detection of bacterial proliferation in
suspensions. The inventive system includes 1) a microfluidic
testing channel unit with electrodes at its opposite ends, whereas
a testing suspension may be injected into the testing channel at a
pre-determined amount and interval, 2) an impedance detecting means
to measure the impedances of the testing suspension at a series of
pre-determined frequencies ranging from about 10 KHz to about 100
MHz, whereas the impedance detecting means is in electrical
communication with the electrodes, and 3) a data analysis means
that processes the impedances.
[0028] The present inventive method and system may be employed in
various applications. For example, the inventive method may be
applied in food quality testing. Producers of bottled products like
pasteurized milk, juice, etc. as well as operations like meat
processing factories need to ensure that the procedures they adopt
are effective to eliminate harmful bacteria from their products.
That is, producers need to verify that processes, such as
pasteurization, irradiation, etc. are effective to eleiminate
viable bacteria left in the product or that the amount of viable is
below a certain threshold. Plating techniques, or some of the
"automated" technique, such as RABIT, are examples of techniques
that are currently used to detect viable bacterial. The inventive
method can perform much better than the current automated
methods.
[0029] The inventive method and system may be employed in the rapid
detection of slow glowing pathogens like mycobacteria for animal
and human health applications. Mycobacteria are a class of bacteria
that are responsible for a number of important diseases both in
animals and humans. For example, such mycobacteria can cause
Johne's Disease in cattle, which is estimated to cost the U.S.
cattle industry over $2 Billion a year. As another example, similar
mycobacteria can cause Tuberculosis in humans. Mycobacteria are
characterized by a uniquely thick cell wall and slow metabolism
that enable to survive many conditions that kill almost all other
bacteria. Their unique physiology makes detecting them in samples
such as sputum and fecal extracts an extremely time consuming
process. Typically, the biological sample is first subjected to
conditions that kill other microorganisms--and then it is cultured
(mycobacteria are allowed to grow and proliferate). Automated
techniques such as the TREK-VET system relies on a decrease in the
concentration of dissolved oxygen to detect viable Mycobacterium
avium ssp. Paratuberculosis, which is the causative organism for
Johne's Disease. Another automated technique, such as BACTEC
system, uses the detection of radioactive CO.sub.2 released from
radiaactive solid nutrients to detect the presence of viable M
tuberculosis. Both of these systems can typically take weeks (e.g.,
40 days) to provide results. Estimately, the inventive method and
system can cut down the time to detection by a factor of 4 to
10.
[0030] Additionally, the inventive method and system may be
employed to assist blood culture analyses. For example, septicemia
or sepsis is the infection of pathogenic microroganisms into the
bloodstream. There are over 200,000 cases of sepsis in the United
States each year. Typically, when a patient begins to demonstrate
clinical symptoms of the disease, pathogens are present at less
than 10 cfu/ml of blood. To detect these pathogens, about 3-10 ml
of blood is cultured (incubated at 35.degree. C. under
aerobic/anareobic conditions) for 2 to 7 days. Estimately, the
inventive method/system may bring the TTD of this procedure down by
a similar factor. Since the rapid diagnosis of sepsis enables the
clinician to commence proper treatment quickly, which will make a
significant impact in reducing the fatality rate associated with
sepsis (which is currently close to 30%).
[0031] Furthermore, the inventive method/system may be applied in
environmental water quality testings. Currently it takes more than
2 days to ascertain presence, or lack thereof, of viable pathogens
(such as coliforms) after reports of supected infections in areas
such as recreational water bodies (E.g. beaches, lakeshores etc).
The inventive method/system may cut down the TTDs for these cases
as well.
Example of Bacterial Proliferation Testings
[0032] The present invention also provides several examples of
bacterial proliferation testings using the inventive
method/system.
[0033] Sample preparation and inoculation of bacteria into samples:
Escherichia coli K12 (ATCC 23716), and Lactobacillus acidophilus
(Nature's Life.TM. Apple-honey Lactobacillus acidophilus probiotic)
were used in this study. In order to obtain load cultures, E. coli
K12 was incubated overnight at 37.degree. C. in Tryptic Soy Broth
(TSB) (Bacto.TM., BD), Lactobacillus acidophilus was incubated at
30.degree. C. for about 48 hrs in MRS Broth (Difco.TM., BD). These
were then used, in appropriate dilutions, to seed the samples in
which bacterial proliferation was monitored using the inventive
method. These samples included those of TSB loaded with E. coli (to
compare the present technique to previous work), and two
representative liquids to study the ability of the method to detect
bacteria in food samples, [2% reduced fat milk (Prairie Farms.TM.)
for E. coli and preservative free apple juice (Florida's
Natural.TM.) for Lactobacillus acidophilus].
[0034] To facilitate growth of lactobacilli in the apple juice, its
pH was adjusted to about 6 by adding potassium hydroxide (about 1
ml of 10M KOH to 50 ml of Apple Juice). The media and the food
samples were all autoclaved at 121.degree. C., 15 psi to ensure no
presence of live bacteria in them. This ensures the right
concentration of the bacteria in the sample when it is artificially
inoculated it a bacteria of interest. The samples are allowed to
cool down to room temperature before bacterial inoculation. The
bacterial suspension after being incubated for specified time
periods was initially assumed to contain approximately 10.sup.9
CFU/ml bacteria. 1 ml of E. coli K12 and 1 ml of Lactobacillus
acidophilus were taken in separate eppendorf tubes and centrifuged
for 8 minutes to settle the bacteria down as pellet. Then the
supernatant was discarded and pellets were re-suspended in equal
volume of food samples in which they were to be detected. Then the
suspension was serially diluted and inoculated into the liquid
samples to have different initial concentration of bacteria in them
and also simultaneously the samples were plated onto petriplates to
get the actual initial concentration of the inoculated bacteria in
the sample.
[0035] Experimental Design: 4 sets of 9 ml of each of the liquid
samples (TSB, milk or apple juice) were taken in the incubating
tubes. Each tube was inoculated with the bacteria to be detected
such that the final concentrations of the bacteria in the tubes
were approximately 1, 10, 100 and 1000 CFU/ml respectively. The
tubes were then allowed to incubate for a time period of 8 hours
for 1, 10, 100 CFU/ml concentrations and 5 hours for 1000 CFU/ml
concentration. At regular time intervals (30 min for 1000 CFU/ml
and 1 hour for 1, 10, 100 CFU/ml) small volume (.about.250 .mu.l)
of the sample was taken out, injected into the cassettes and
impedance measurements were made using the Agilent 4294A impedance
analyzer (Agilent technologies, Calif., USA) over the frequency
range of 1 kHz to 100 MHz. Simultaneously at every time interval,
100 .mu.l of the sample was taken, diluted appropriately and plated
onto petri-dishes to give actual concentration of bacteria at that
hour in the sample. The entire process was repeated independently
at least 3 times for each targeted initial load of the system (1,
10, 100 or 1000 CFU/ml) and for all liquids (TSB, milk and apple
juice).
[0036] The microfluidic cassettes used for the measurement was
fabricated using liquid phase photo-polymerization of a
commercially available UV curable polymer (Loctite 363.TM., a
process that has been described elsewhere in detail. The cassettes
were sterilized in an autoclave at 121.degree. C. before use. After
each of the experiment, the electrical connectors were replaced;
cassettes were washed thoroughly with soap, bleach, alcohol and
water, and then autoclaved.
Viable Bacteria Detection System
[0037] FIG. 1A is a block diagram of an exemplary computing
environment 100 for detecting the presence of viable bacteria in a
fluid sample. The computing environment 100 includes a microfluidic
unit 102, an input device, 103, and a viable bacteria detection
system (VBDS) 104.
[0038] According to one aspect, the microfluidic unit 102 receives
a portion of a particular suspension sample from a sample
collection device (not shown), such as a vial, vacutainer, or other
fluid sample container. For example, the sample collection device
may be a fingerstick collection device or a vacutainer that is used
to collect 50-200 .mu.l of a whole blood sample from a finger stick
and to subsequently transfer at least a portion of the sample to
the microfluidic unit 102. According to one aspect, the
microfluidic unit 102 is a disposable closed containment device
that contains reagents, fluidic channels, and biosensors. The
microfluidic unit 102 also includes electrodes 106, 108 that allow
input and/or output of electrical voltage and/or electrical current
signals, and may simultaneously serve as a measurement electrode
according to an aspect of the invention.
[0039] Referring briefly to FIG. 2A, a schematic representation of
a micro-channel 200 of the microfluidic unit 102 with electrode
terminals 202, 204 on either end loaded with a suspension 206
harboring bacteria 208 is depicted. When the microfluidic unit 102
is loaded with a portion of a particular suspension 206 being
investigated, the suspension 206 fills the micro-channel 200 and
contacts the terminals 202, 204 of the electrical electrodes 106,
108.
[0040] Referring back to FIG. 1A, the VBDS 104 includes an
interface 110 that enables the microfluidic unit 102 to be
connected and disconnected to the VBDS 104. The interface 110
comprises, for example, receptacles 112, 114 for receiving
electrodes 106, 108 of the microfluidic unit 102 such that the VBDS
104 can supply analysis signals to the sample and receive
measurement signals from the sample. According to one aspect, the
VBDS 104 comprises a signal generator 116 to generate voltage
and/or current signals at various frequencies and amplitudes to
apply to the electrodes the 106, 108 of microfluidic unit 102.
[0041] The VBDS 104 also includes a signal analyzer 118 to measure
parameters of a circuit created by the electrical interaction
between the electrodes 106, 108 and the fluid sample. According to
one aspect, the signal analyzer 118 is, for example, an Agilent
4294A Impedance Analyzer that measures the electrical impedance
between the electrodes 106, 108 at multiple frequencies (e.g.,
>500 different frequencies) between 1 KHz to 100 MHz. The signal
analyzer 118 measures the magnitude and phase of an AC current that
flows through the suspension upon the application of a sinusoidal
AC voltage of 500 mV (peak-to-peak) and then calculates the
impedance (i.e., resistance and reactance) from the measurements.
Since the current is not in-phase with the applied sinusoidal
voltage, the impedance, which can be considered as the AC analog of
the DC resistance, has both an in-phase component called the
resistance (R), and an out-of-phase component called the reactance
(X). Impedance is typically represented as a complex number and as
shown in equation 1.
Z=R+j X (1)
[0042] where j= -1
[0043] Alternatively, the impedance can also be represented
completely by its magnitude (|Z|) and its phase angle .theta.. The
magnitude and phase angle, respectively, of the impedance, are
related to the resistance and reactance by the equations
Z= (R.sup.2+X.sup.2) (2a)
.theta.=Tan.sup.-1(X/R) (2b)
[0044] The signal analyzer 118 measures impedance by measuring the
resistance (R) and reactance (X) for each sample, over the
frequency range of 1 kHz to 100 MHz and hence generates an
impedance data set containing the values of R and X for each of the
multiple frequencies.
[0045] By obtaining impedance measurements at multiple
pre-determined frequencies, a the value of the parameter in the
theoretical circuit model, which reflects the amount of capacitive
charge stored in the interior bulk of the suspension, can be
calculated. As discussed above, the presence of bacteria in a
suspension can be detected based on the changes in the bulk
capacitance of the suspension over time. Thus, by repeating the
process of obtaining impedance measurements at multiple
pre-determined frequency after pre-determined intervals of time,
the presence, or lack thereof, of viable bacteria in the suspension
can be determined.
[0046] According to one aspect, the user interface 103 is a
computer or processing device, such as a personal computer, a
server computer, or a mobile processing device. The input device
may include a display (not shown) such as a computer monitor, for
viewing data, and an input device (not shown), such as a keyboard
or a pointing device (e.g., a mouse, trackball, pen, touch pad, or
other device), for entering data. The user interface 103 is used by
a user to enter information about a particular sample to be
analyzed by the VBDS 104. For example, the user uses the keyboard
to interact with an entry form (not shown) on the display to enter
sample information data that includes, for example, fluid type,
fluid collection date and time, fluid source, etc.
[0047] The user interface device 103 can also be used by the user
to generate an analysis request 119 for a particular sample to be
analyzed by the VBDS 104. For example, after a portion of the
particular sample in a collection device has been transferred to
the microfluidic unit 102 and the microfluidic unit 102 is
connected to the VBDS 104, the user interacts with an entry form
(not shown) on the display of the user interface 103 to select, for
example, start analysis control to generate the analysis request
119. The user interface 103 provides the analysis request 119 to
the VBDS 104. The VBDS 104 initiates the operation of the signal
generator 116 and the signal analyzer 118 in response to the
received analysis request 119.
[0048] Subsequently, the user interface device 103 can also be used
by the user to generate another analysis request 119 for another
portion of the same particular sample. For example, after a
pre-determined time interval expires, the user interface device 103
notifies or alerts the user to transfer another portion of the
particular sample from the collection device to the microfluidic
unit 102 for analysis. The microfluidic unit 102 is again connected
to the VBDS 104 and the user again interacts with the entry form
(not shown) on the display of the user interface 102 to select the
start analysis control to generate another analysis request 119. As
described in more detail below, the pre-determined time interval is
a function of expected TTDs data for individual samples.
[0049] According to another aspect, the user interface device 103
can also be used by the user to define pre-determined time
intervals for collecting different portions of the sample. For
example, the user may define pre-determined time intervals, such as
15 minutes, 30 minutes, 1-hour, etc.
[0050] According to another aspect, the user interface device 103
can also be used by the user to define a maximum processing time
for attempting to identify viable bacteria in a particular sample.
For example, the user may define the maximum processing as equal to
8 hours, 24 hours, 48 hours, etc. As explained below, if the VBDS
104 does not determine that the sample has viable bacteria within
the maximum processing time, the sample is deemed not to contain
bacteria and the detection process is terminated.
[0051] The VBDS 104 executes a bacteria detection application (BDA)
120 to detect whether viable bacteria is present in the sample
based on a change in impedance measurements of the sample over a
period of time. According to one aspect, the BDA 120 determines
parametric values of a model circuit based on the impedance
measurements. The BDA 120 then determines whether one or more of
the parametric values, such as bulk capacitance, change more than a
threshold amount over the pre-determined time period. If the change
in one or more of the parametric values is more than the threshold
amount, then the sample is deemed to contain viable bacteria. If
the amount of change in the one or more of the parametric values
does not exceed the threshold amount, then the sample is not deemed
to contain viable bacterial. The BDA 120 then displays whether the
result of the analysis is positive or negative for viable
bacteria.
[0052] The data source 122 is, for example, a computer system, a
database, or another data system that stores data, electronic
documents, records, other documents, and/or other data. The data
source 1506 may include memory and one or more processors or
processing systems to receive, process, and transmit communications
and store and retrieve data. The BDA 120 retrieves the
pre-determined interval data from the data source 122 to determine
when to notify the user to collect another portion of the sample
for analysis. According to one aspect, the data source 122 includes
a sample database 124 that stores pre-determined time interval data
for various fluid samples. The sample database 124 may also store
impedance data or the various parametric values of the model
circuit determined at different point in time for each of the
various samples.
[0053] Although, the data source 122 is illustrated in FIG. 1A as
being integrated with the VBDS 104, it is contemplated that in
other aspects the data source 122 may be separate and/or remote
from the VBDS 104. According to one such aspect, the VBDS 104
communicates with the data source 122 over a communication network,
such as the Internet, an intranet, an Ethernet network, a wireline
network, a wireless network, and/or another communication network,
to identify relevant images, electronic documents, records, other
documents, and/or other data to retrieve from the data source 122.
In another aspect, the VBDS 104communicates with the data source
122 through a direct connection.
[0054] FIG. 1B is a block diagram that depicts an exemplary BDA
120. According to one aspect, the VBDS 104 includes a processing
system 150 that executes the BDA 120 to detect whether viable
bacteria is in the sample based on a change in impedance
measurements detected at the signal analyzer 118 over a period
time. The processing system 150 includes one or more processors,
and the processing system 150 can reside on a computer or other
processing system.
[0055] The BDA 120 includes instructions or modules that are
executable by the processing system 150 to manage the retrieval of
pre-determined time interval data from the data source 122 and to
detect whether there is viable bacterial in the sample changes in
on or more of the determined parametric values of a model circuit.
The VBDS 104 includes computer readable media 152 configured with
the BDA 1512.
[0056] Computer readable medium (CRM) 152 may include volatile
media, nonvolatile media, removable media, non-removable media,
and/or another available medium that can be accessed by the VBDS
104. By way of example and not limitation, computer readable medium
152 comprises computer storage media and communication media.
Computer storage media includes memory, volatile media, nonvolatile
media, removable media, and/or non-removable media implemented in a
method or technology for storage of information, such as computer
readable instructions, data structures, program modules, or other
data. Communication media may embody computer readable
instructions, data structures, program modules, or other data and
include an information delivery media or system.
[0057] A data collection module 154 activates the signal generator
116 to generate a series of analysis signals to apply to the sample
at various frequencies in response to an analysis request 119
received from the user interface 103. The data collection module
154 also activates the signal analyzer 118 to obtain impedance
measurement data of the sample based on the applied analysis
signals in response to the received analysis request 119. The net
measured impedance (Z.sub.measured) is, as shown by equation 1 is
affected by not only by the presence of conductive and capacitive
(charge-storing) elements in the bulk, but also by such elements
present at the electrode-solution interface. As described above,
the signal analyzer 118 measures impedance by measuring the
resistance (R) and reactance (X) for each sample, over the
frequency range of 1 kHz to 100 MHz and hence generates the data
set containing the values of R and X at each of the multiple
frequencies.
[0058] A parameter calculation module 156 calculates parametric
values of a model circuit based on the impedance measurement data
sets received from the data calculation module 154. Each impedance
data set corresponds to a series of impedance measurements obtained
at various frequencies at during a particular measurement cycle.
Each measurement cycle is separated by a pre-determined time
interval. According to one aspect, parameter calculation module 156
employs, for example, commercial circuit analysis software (Z view)
to fit the values of resistance (R) and reactance (X) for a
particular impedance measurement data set to an equivalent circuit
model. The parameter calculation module 156 uses the circuit model
and the impedance measurement data set to estimate each of the
individual parameters (R.sub.e, C.sub.e, R.sub.b and C.sub.b) of
the circuit.
[0059] Referring briefly to FIG. 2B, an example of the circuit
model 210 that works well for estimating individual impedance
parameters (R.sub.e, C.sub.e, R.sub.b and C.sub.b) at low
frequencies. However, the model circuit depicted in FIG. 2B may be
not sufficiently accurate at estimating the parameters at higher
frequencies.
[0060] FIG. 3A depicts another circuit model 302 for estimating
impedance parameters. In this model circuit 302, the bulk
capacitance (C.sub.b) is replaced with a Constant Phase Element
(CPE) 304 and the model circuit provides a much better fit to the
data obtained, as shown in the FIG. 3B-3D. The CPE 304 a
non-intuitive circuit element that replaces a capacitor in a
circuit when the there is some type of non-homogeneity in the
system, delaying or impeding the movement of charge carriers. In
more mathematical terms, the impedance of a CPE 304 is given by the
equation
Z=0 j(1/(wQ).sup.n) (3)
[0061] As shown in equation 3, the impedance of the CPE 304 is
defined by two values: the magnitude component CPE-T (Q) that is
measured in farads and the phase component CPE-P (n). If CPE-P (n)
equals 1 then the equation is identical to that of a capacitor.
While bacteria can store a charge, it likely does not behave like
ideal capacitors. Thus, using a CPE 304 to compensate for the non
ideal charge storage capability of the bacterial is appropriate.
The CPE is used for the data analysis, as the arc of the Cole Plot
for the impedance data was a depressed semicircle or an arc of the
circle rather than a perfect semicircle as would be the case if the
bacteria behaved like ideal capacitors. The value of the CPE-P (n)
is not a constant but is different for different samples.
[0062] FIG. 3B depicts a plots of impedance obtained from Impedance
Analyzer (see analyzer plot 306) as compared to another plot of
impedance obtain from fitting via the circuit model with the
parameters on the right (see model circuit plot 308). FIG. 3C
depicts a plots of the impedance magnitude obtained from Impedance
Analyzer (see analyzer plot 310) as compared to another plot of
impedance magnitude obtain from fitting via the circuit model with
the parameters on the right (see model circuit plot 312). FIG. 3D
depicts a plots of the impedance phase obtained from Impedance
Analyzer (see analyzer plot 314) as compared to another plot of
impedance phase obtain from fitting via the circuit model with the
parameters on the right (see model circuit plot 316).
[0063] Referring back to FIG. 1B, the parameter calculation module
156 loads a particular impedance data set and the model circuit to
which it is fit is constructed. The parameter calculation module
156 then estimates an initial value each of the circuit parameters
(R.sub.e, C.sub.e, R.sub.b, CPE-T, and CPE-P) and numerically
optimizes the parameter values to obtain the best fit for the
system as a whole over the range of frequencies examined. The CPE-T
value determined by the parameter calculation module 156 provides a
measure of the charge-storing capability of the suspension being
investigated. Over a period of time, this quantity is expected to
increase with increase in the number of bacteria, and one can
conclusively state that there are viable bacteria in the sample
when one observes this quantity (CPE-T) to increase
significantly.
[0064] According to another aspect, the parameter calculation
module 156 determines a corresponding confidence interval for the
CPE-T value. A "significant" change is said to occur when a
confidence interval of the newer value (as specified by the
software fitting the impedance v/s frequency data to the proposed
theoretical circuit model of our system) does not overlap with the
reference value (usually the zero-hour value). The confidence
interval refers to, for example, a range or expected variance of
the calculated CPE-T value based on fitting the impedance v/s
frequency data to the proposed theoretical circuit model. For
example, the initial confidence interval for a calculated CPE-T
value of a sample may be 35+/-3 at a first point in time (e.g., 0
hour point) 39+/-2 at second point in time (e.g., 1 hour later),
and 45+/-3 at a third point in time (e.g., 2 hours later). In this
instance, because the CPE-T value can be as high as 38 at a first
point in time and as low as 37 at the second point in time, no
"significant" change deemed to have occurred between the 0 hour
point and the 1 hour point because the CPE-T values overlap As a
result, no viable bacterial is deemed to be present at the 1 hour
point. However, because the CPE-T value can only be as low as 42 at
the third point in time, a "significant" change is deemed to have
occurred between the 0 hour point and the 2 hour point because the
CPE-T values do not overlap and, thus, viable bacterial is deemed
to be present.
[0065] For instance, with reference now to FIG. 4, the reading
taken at 1-hour is not significantly different from the initial
(0-hour) value because there is overlap between a confidence
interval of the CPE-T reading at the zero (0)-hour point and the
one (1)-hour point. However, the CPE-T reading taken at the 2-hour
point is significantly different because there is no overlap
between CPE-T values at the zero (0)-hour point and the two
(2)-hour point. The time needed to make this observation
(significant increase in the value of CPE-T) is the Time to
Detection (TTD) for the present system. For the example depicted in
FIG. 4, the TTD is 2 hours). Thus, the pre-determined time interval
data may correspond to minimum time required to observe a
significant increase in the value of CPE-T for various types of
fluid samples.
[0066] According to another aspect, the parameter calculation
module 156 retrieves the pre-determined time interval data from the
data source 122 for the sample being analyzed. As discussed above,
the pre-determined time interval data may correspond to the
expected Times to Detection (TTDs) of bacteria for individual
samples. After the expiration of a time interval defined by the
pre-determined time interval data, the parameter calculation module
156 sends a notification to a display of the user interface 103 to
notify the user to transfer another portion of the sample to the
microfluidic unit 102 for analysis. After transferring another
portion of the sample to the microfluidic unit 102, the user
connects the microfluidic unit 102 to the VBDS 104 and generates
another analysis request 119.
[0067] It is contemplated that in other aspects, the parameter
calculation module 156 automatically initiates the collections of
another portion of the sample being analyzed without the
intervention of a user. For example, rather than transferring the
notification to the display, the parameter calculation module 156
transfers a transfer notification to a transfer mechanism (not
shown) that is configure to collect the other portion of sample.
The transfer mechanism may be further configured to connect the
microfluidic unit 102 to the VBDS 104 and to generate another
analysis request 119.
[0068] The data collection module 154 activates the signal
generator to generate another series of analysis signals to apply
to a different portion of the sample at various frequencies in
response to the other analysis request 119 received from the user
interface 103. The data collection module 154 also activates the
signal analyzer 118 to measure new impedance data of the sample
based on the in response to the other analysis request 119. The
parameter calculation module 156 then calculates new parametric
values of the model circuit based on the new impedance measurement
data sets received from the parameter calculation module 156.
[0069] The analysis module 158 compares the confidence interval of
the at least one of the new impedance parametric values and the
confidence interval of the at least one previously calculated
impedance parametric values If the two confidence intervals do not
overlap, then the analysis module 158 determines that viable
bacterial is present.
[0070] According to another aspect, if the analysis module 158
determines that the confidence intervals overlap, then the analysis
module 158 waits for the pre-determined interval to receive another
set of new parametric values of the model circuit from the
parameter calculation module 156. This may be an iterative process
by which the analysis module 158 performs a series of iterations
during a particular time period before determining that there is no
bacteria present in the sample. For example, the analysis module
158 may continue to wait for the pre-determined interval to receive
another set of new parametric values of the model circuit from the
parameter calculation module 156 until the maximum processing time
has expired.
[0071] According to another aspect, if the analysis module 158
compares a new impedance parametric value, such as a new confidence
interval associate with a new CPE-T value with a previously
determined confidence interval associate with a previously
calculated CPE-T value to see if the values overlap. As discussed
above when the new confidence interval value and the previous
confidence interval overlap, no viable bacterial is deemed to be
present. However, when the new confidence interval value and the
previous confidence do not interval overlap, viable bacterial is
deemed to be present.
[0072] An output module 160 generates an analysis result for
display. According to one aspect, the displayed result indicates
whether or not there is viable bacterial present in the sample.
According to one aspect, the displayed result may also indicate an
amount and/or a type of bacteria present in the sample.
Viable Bacteria Detection Method
[0073] FIG. 1C illustrates a method for detecting the presence of
viable bacteria in a fluid sample. At 170, a sample of a fluid
sample in which bacterial presence is suspected is collected from a
source. An initial portion of the sample is transferred to a
microfluidic unit 106 at 172. At 174, a series of analysis signals
at different frequencies are generated at the VBDS 104 and applied
to the microfluidic unit 106 in response to user input received at
the VBDS 104. The VBDS 104 determines an impedance data set by
measuring the resulting impedance of the first portion of the
sample for each of the analysis signals at 176. At 178, VBDS 104
determines an initial CPE-T value of the model circuit based on the
impedance measurements and an initial confidence interval. The VBDS
104 retrieves pre-determined time interval data from the data
source 122 for the sample being analyzed and after the expiration
of a time interval defined by the pre-determined time interval
data, the VBDS 104 sends a notification to a display of the VBDS
104 to notify a user to analyze another portion of the sample at
180.
[0074] The other portion of the sample is transferred to the
microfluidic unit 106 after the expiration of a time interval at
182. At 184, another series of analysis signals at the same
different frequencies are generated at the VBDS 104 and applied to
the microfluidic unit 106 in response to another user input
received at the VBDS 104. The VBDS 104 determines a new impedance
data set by measuring the resulting impedance of the second portion
of the sample for each of the analysis signals at 186. At 188, the
VBDS 104 determines a new CPE-T value, of the model circuit based
on the new impedance measurements and determines a new confidence
interval. The VBDS 104 compares the initial confidence interval to
the new confidence interval factor to determine if the values
overlap at 190. If the initial confidence interval and the new
confidence interval values do not overlap at 190, the VBDS 104
displays a positive result indicating that viable bacterial is
present in the sample at 192. If the initial confidence interval
and the new confidence interval values overlap at 190, the VBDS 104
checks to see if a maximum processing time has expired at 194. If
maximum processing time has not expired at 194, the VBDS 104 sends
a notification to a display of the VBDS 104 to notify the user to
analyze another portion of the sample at 180. If maximum processing
time has expired at 194, VBDS 104 displays a negative result
indicating that viable bacterial is not present in the sample and
ends processing at 196.
[0075] Ability of the calculated CPE-T value to track true
bacterial counts: As can be seen in FIG. 5, using the value of
CPE-T as an indicator of bacterial load in the system, closely
tracks the actual bacterial numbers present (as obtained using
plate counts) irrespective of whether the bacterial numbers hold
steady (as occurs in the lag and saturation/stationary phase),
rise, or decline. The decrease in CPE-T values as bacteria die off
seems to indicate that dead bacteria are not as capable as live
ones of storing charge.
Example Times to Detection
[0076] Times to Detection (TTD) as a function of initial bacterial
loads: Three types of samples (TSB, Milk and apple juice) were
inoculated with 4 different initial bacterial loads (targeted to be
1, 10, 100, 1000 CFU/ml), impedance measurements were taken at
specific intervals (half hour or one hour), and the impedance data
were analyzed using Z view software to obtain the CPE-T values as
described in the previous section. These values were used to obtain
TTDs using the criteria explained using FIG. 4. FIG. 4 is a plot
showing the increase in the bulk capacitance (e.g., see diamond
402) with actual increase in the concentration of the bacteria
(e.g., see square 400) in the suspension. The plot also indicates
the time to detection (shown by the arrow), For this sample, the
error bar of CPE-T value of 2 hours does not overlap with the error
bar of zero-hour reading and hence 2 hours is considered as the
time to detection. Some more of such typical plots of CPE-T vs time
which gives the TTDs for each sample are shown in FIGS. 6A-6I, with
the arrows indicating the TTD for that sample with respective
initial bacterial load.
[0077] In a few cases, mostly for L. acidophilus in apple juice, a
significant lag phase is observed. During this period, the actual
concentration of bacteria (e.g., square 502 of FIG. 5) did not
grow, and sometimes even die--as indicated by the plate count data.
In such cases, a better estimate of the capabilities of the present
system is obtained by subtracting the lag phase time (e.g., 2
hours) when calculating the Time to Detection for the given initial
load in the given system. For example, in the case shown in FIG. 5,
although the significant increase in CPE-T (e.g., diamond 504 of
FIG. 5) from the initial value is detected only at the fourth hour,
the TTD of the system is taken to be 2 hours since for the first
two hours, the bacteria in the suspension were in the lag
phase.
[0078] Each experiment with the targeted initial load of bacteria
in a specific sample is repeated three times to ascertain the
reproducibility of the method. A more accurate estimate of the true
value of the initial loads could only be obtained the next day,
once plate counts were obtained. Hence, twelve points each have
been taken for TSB, milk, and apple juice (some of these points
overlap very closely, and are hence not distinguishable). These
points are used to calculate a line of bet fit using linear
regression, and these lines are also shown in FIGS. 7A-7C. The
equations for these lines provide the best estimate of the time
that the present system will take to detect a given load of a
particular type of bacterium in a particular substrate.
[0079] As seen in from the plots 702, 704, 706 shown in FIGS.
7A-7C, respectively, there is an inverse relationship between the
(log) initial load of bacteria in a sample, and the TTD of
proliferating bacteria using the present system and method. In this
it is similar to methods that rely on detecting the effects of
bacterial metabolism such as Bactometer, RABIT etc. This is
expected since the presence of more bacteria (our method) also
leads to increased metabolite consumption/generation. Also, as in
the case of RABIT, Bactometer etc., one can generate a calibration
plot for a particular type of suspension, and the TTD can be used
to estimate the initial load of the system. The scatter observed is
qualitatively comparable to the data used to generate calibration
curves for RABIT etc. The scatter arises due to multiple reasons.
For example, two such reasons include uncertainties in the
estimates of the initial loads and differences in metabolic state
of members within and between populations seeded. In other words,
although the plot of TTDs against a "known" initial load is based
on plate counts, this "known" value itself is subject to some
degree of uncertainty--typically of the order of the square root of
the true number of particles present (Poisson distribution). Thus,
if the suspension being incubated had 100 CFU/ml of bacteria (true
value), it is expected that a 100 .mu.l sample introduced into a
microfluidic unit, or use for plating, will have 10 CFUs. However,
there is also a 33% chance that isolated sample will have either
less than 7 (10- 10), or greater than 13 (10+ 10) bacteria. In
addition, a certain fraction of the bacterial cells that constitute
the inoculum may remain in the lag phase slightly longer than
others. This may not be readily captured by the plate counts taken
to determine initial load (since in plates, they get adequate time
to grow). When operating at low concentrations (low numbers of
bioparticles), such sampling uncertainties have the potential to
introduce a greater relative error. However, despite these sources
of error, the TTD data still shows a clear trend in the manner
expected (inverse with respect to log initial load).
[0080] Another characteristic of the present method is that the TTD
is a function of the doubling time of the proliferating bacteria.
The faster that a given bacterium doubles, the shorter is TTD in
the present invention. For example, E. coli K-12 bacteria that have
a doubling time of 27 minutes at 37.degree. C., and Lactobacillus
acidophilus has a doubling time of 50-60 minutes at 30.degree. C.
Thus the doubling time of Lactobacillus is about 2 times that of E.
coli. The TTDs for Lactobacillus are also correspondingly longer (8
hours for 1 CFU/ml and 4.5 hours at 100 CFU/ml v/s 4.5 hours and 2
hours, respectively, for E. coli at the same initial loads). For
initial loads of 1000 CFU/ml or higher, proliferation was detected
in half an hour (the shortest time interval used) for E. coli (and
in one case, for Lactobacillus as well). Thus, at these relatively
higher loads, bacteria can be detected within one cycle of
division. For lower initial bacterial loads, at the points in time
where significant changes in CPE-T values are detected, their
concentration in the sample (as estimated from the plate counts) is
typically between 200-1000 CFU/ml. As a rough ballpark estimate,
the present invention detect bacteria in the act of doubling their
numbers when there are about 500 of them present per ml of
suspension.
[0081] As shown in FIGS. 8(a) and 8(b), the TTDs of the present
invention compare very well with automated techniques already on
the market and other automated techniques in development. FIG. 8(a)
is a graph 802 that provides the comparison of the present
detection method with that of previously available commercial
automated systems, such as RABIT.TM., Malthus 2000.TM.,
Bactometer.TM., and BacTrac.TM.. FIG. 8(b) is a graph 804 that
provides the comparison between the present system and some of the
systems currently under development. Virtually all of these systems
under development continue to rely on detecting the effects of
bacterial metabolism on the medium properties, such as changes in
pH, Conductivity, oxygen concentration, for detection. Some employ
features and capabilities available through microfluidic systems,
such as micro-interdigitated electrodes or pre-concentration using
dielectrophoresis to try and reduce the overall TTDs. While they
achieve low TTDs (3-9 hrs) for very high initial concentrations of
bacteria (.about.10,000 CFU/ml), they continue to have high TTDs
(10-14 hours) at low initial concentration of bacteria (1 CFU/ml).
Thus, for any given initial load, the inventive system is able to
detect bacteria at least 3 to 4 times faster than other
methods.
* * * * *