CONTROLS AND CALIBRATORS FOR TESTS OF NUCLEIC ACID AMPLIFICATION PERFORMED IN DROPLETS

[0001] This application is a continuation of U.S. patent application Ser. No. 12/586,626, filed Sep. 23, 2010.

[0002] U.S. patent application Ser. No. 12/586,626, in turn, is based upon and claims the benefit under 35 U.S.C.§119(e) of the following U.S. provisional patent applications: Ser. No. 61/194,043, filed Sep. 23, 2008; Ser. No. 61/206,975, filed Feb. 5, 2009; Ser. No. 61/271,538, filed Jul. 21, 2009; Ser. No. 61/275,731, filed Sep. 1, 2009; Ser. No. 61/277,200, filed Sep. 21, 2009; Ser. No. 61/277,203, filed Sep. 21, 2009; Ser. No. 61/277,204, filed Sep. 21, 2009; Ser. No. 61/277,216, filed Sep. 21, 2009; Ser. No. 61/277,249, filed Sep. 21, 2009; and Ser. No. 61/277,270, filed Sep. 22, 2009.

[0003] These priority applications are incorporated herein by reference in their entireties for all purposes.

[0004] This application incorporates by reference in their entireties for all purposes the following materials: U.S. Pat. No. 7,041,481, issued May 9, 2006; and Joseph R. Lakowicz, PRINCIPLES OF FLUORESCENCE SPECTROSCOPY (2^nd Ed. 1999).

[0005] Droplet-based tests for amplification generally need to be accurate. If inaccurate, these tests can generate erroneous results, that is, false negatives and false positives. Each type of erroneous result can have detrimental consequences. False negatives related to detection of a disease could mean that the disease is not treated early and is permitted to spread. In contrast, false positives could cause unnecessary alarm, potentially triggering an unnecessary response that may be costly and disruptive. To avoid problems associated with false negatives and false positives, inaccurate amplification tests must be repeated to improve their reliability, which increases cost and uses more sample and reagent, each of which may be precious.

[0006] FIG. 1 shows a graph 5710 illustrating an exemplary approach for using fluorescence to measure amplification of a nucleic acid target in droplets formed by partitioning a sample. The graph plots, with respect to time, fluorescence signals that may be detected from a flow stream containing the droplets. Each droplet may be detected as a transient change (e.g., a transient increase) in intensity of the fluorescence signal, such as a peak or spike 5712 (i.e., a wave) formed by the fluorescence signal.

[0007] To improve clarity, the illustrative data shown here and in other figures of the present disclosure, are presented in a simplified form: each peak has no width and projects from a constant background signal 5713 formed by detection of a continuous phase carrying the droplets. However, a signal peak may have any suitable shape based on, for example, the frequency of detecting signals, the shape of each droplet, the size and geometry of a channel carrying the flow stream, the flow rate, and the like. Moreover, the signal peaks may have any suitable temporal distribution, for example, occurring at relatively constant intervals, as shown here, or at varying intervals. A droplet signal provided by and/or calculated from the peak (e.g., a signal corresponding to peak height or peak area, among others) may be used to determine whether amplification occurred in the corresponding droplet, and thus whether the droplet received at least one molecule of the nucleic acid target when the sample was partitioned.

[0008] Each droplet signal may be compared to a signal threshold 5714, also termed a cutoff. This comparison may provide a determination of whether each droplet signal represents a positive signal (target is present) or a negative signal (target is absent and/or not detected), for amplification in the droplet. For example, droplet signals greater than (and, optionally, equal to) the threshold may be considered as representing positive droplets. Conversely, droplet signals less than (and, optionally, equal to) the threshold may be considered as representing negative droplets. (A positive droplet signal above threshold 5714 is indicated at 5716, and a negative droplet signal below threshold 5714 is indicated at 5718 in FIG. 1.) Comparison to the threshold thus may transform each droplet signal to a digital value, such as a binary value (e.g., a “1” for a positive droplet and “0” for a negative droplet). In any event, the fraction of droplets that are positive can be determined. For a given droplet size, the fraction of positive droplets can be used as an input to an algorithm based on Poisson statistics to determine the number of copies (molecules) of the nucleic acid target present in the initial sample volume. In some embodiments, more than one threshold may be used to categorize results (e.g., negative, positive, or inconclusive).

[0009] FIG. 2 shows an exemplary histogram 5720 of ranges of droplet signal intensities that may be measured from the flow stream of FIG. 1. The relative frequency of occurrence of each range is indicated by bar height. The distribution of positive and negative signal intensities may be larger than the modest difference in signal intensity produced by amplification (a positive droplet) relative to no amplification (a negative droplet). Thus, the distributions of droplet signals from positive droplets and negative droplets may produce a problematic overlap between the amplification-positive and amplification-negative droplet signals, indicated at 5724. Accordingly, as shown in FIG. 1, some amplification-positive droplets may provide relatively weak droplet signals, such as false-negative signal 5726, that are less than threshold 5714, resulting in incorrect identification of these positive droplets as negative. Conversely, some amplification-negative droplets may provide relatively strong droplet signals, such as false-positive negative signal 5728, that are greater than threshold 5714, resulting in incorrect identification of these negative droplets as positive. Since either type of erroneous result may be costly and harmful, it is desirable to minimize their occurrence.

[0010] There are many factors that can lead to variation in the fluorescence signal from droplets tested for amplification. Examples of physical parameters that may affect the fluorescence signal may include droplet position when detected (e.g., relative to the “sensed volume” of the detector), droplet volume and shape, optical alignment of detection optics (including excitation source, filters, and detector), detector response, temperature, vibration, and flow rate, among others. Examples of reaction chemistry parameters that may affect the fluorescence signal include the number of target molecules and/or the amount of background nucleic acid present in each droplet, amplification efficiency, batch-to-batch variations in reagent concentrations, and volumetric variability in reagent and sample mixing, among others. Variations in these physical and chemical parameters can increase the overlap in the distribution of positive and negative droplet signals, which can complicate data interpretation and affect test performance (e.g., affect the limit of detection). The variations can occur within a run and/or between runs, within a test on a target and/or between tests on different targets, on the same instrument and/or different instruments, with the same operator and/or different operators, and so on.

[0011] Thus, there is a need for improved accuracy and reliability in droplet-based amplification tests. For example, it would be desirable to have droplet-based controls for these tests, optionally, droplet-based controls that can be incorporated into test droplets or incorporated into control droplets that can be intermixed with test droplets. Such integrated controls may have the benefit of reducing cost by processing control reactions in parallel with test reactions, which may speed the analysis. It also would be useful to have one or more controls that can be used to verify hardware, reagent, and/or software (e.g., algorithm) performance.

[0012] The present disclosure provides a system, including methods and apparatus, for performing droplet-based tests of nucleic acid amplification that are controlled and/or calibrated using signals detected from droplets.

[0033] The present disclosure provides a system, including methods and apparatus, for performing droplet-based tests of nucleic acid amplification that are controlled and/or calibrated using signals detected from droplets.

[0034] The present disclosure provides a method of sample analysis.

[0035] Droplets may be obtained. The droplets may be generated on-line or at least a subset of the droplets may be pre-formed off-line. At least a subset or all of the droplets may include a partition of a sample to be tested and may be capable of amplification of at least one test nucleic acid target, if present, in the partition. In some embodiments, the droplets may be capable of amplification of a test nucleic acid target and a control nucleic acid target. The droplets collectively or each may include a dye, or at least a first dye and a second dye. In some embodiments, the droplets may be of at least two types, such as two or more types of test droplets, test droplets and calibration droplets, or test droplets and control droplets, among others. In some embodiments, the two or more types of droplets may be distinguishable based on distinct temporal positions of the droplets types in a flow stream, the presence of respective distinct dyes in the droplet types, distinguishable signal intensities of the same dye (or different dyes), or a combination thereof, among others.

[0036] Signals, such as fluorescence signals, may be detected from the droplets. The signals may include test signals, calibration signals, control signals, reference signals, or any combination thereof. In some embodiments, test signals and control signals may indicate respectively whether amplification of a test nucleic acid target and a control nucleic acid target occurred in individual droplets. In some embodiments, detection may include (a) exciting first and second dyes with a same wavelength of excitation light and (b) detecting emitted light from the first and second dyes at least substantially independently from one another in respective first and second detection channels.

[0037] The signals detected may be analyzed to determine a test result related to a presence (number, concentration, etc.), if any, of a test nucleic acid target in the sample. In some embodiments, analysis may include transforming test signals based on reference signals to reduce variation in the test signals. The test signals and the reference signals may be detected in respective distinct detection channels or in the same detection channel. In some embodiments, the reference signals may be provided by a second dye that is not coupled to an amplification reaction and thus serves as a passive reference. In some embodiments, the reference signals may be provided by control signals detected from a control amplification reaction. The control amplification reaction may measure amplification of an exogenous or endogenous template. In some embodiments, analysis may include (a) comparing test signals, or a transformed set of the test signals, to a signal threshold to assign individual droplets as positive or negative for a test nucleic acid target, and (b) estimating a number of molecules of the test nucleic acid target in the sample based on the comparison. In some embodiments, analysis may include (a) analyzing control signals to determine a control value corresponding to a number and/or fraction of the droplets that are amplification-positive for a control nucleic acid target, and (b) interpreting a test result, such as determining its validity, based on the control value.

[0038] The systems disclosed herein may offer improved instrument calibration and/or substantial improvements in the accuracy and/or reliability of droplet-based amplification tests. Exemplary capabilities offered by the present disclosure may include any combination of (1) correcting/minimizing variations in the fluorescence signal to increase the accuracy of droplet PCR results; (2) providing an internal indicator of whether nucleic acid amplification failed (e.g., PCR inhibition from interfering components in the sample, incorrect sample and reagent mixing, incorrect thermal cycling, incorrect droplet formation); (3) providing measurement of droplet volumes without having to add additional hardware components; (4) providing measurement of changes in the baseline fluorescence signal (i.e., baseline drift); (5) providing calibration of a droplet detector before and/or during a run; (6) monitoring the performance of quantitative droplet PCR measurements and data processing algorithms before and/or during a run; (7) verification of droplet integrity (e.g., absence of coalescence); (8) obtaining information on droplet generation and detection frequency (spatially and temporally) using an in-line detector; (9) measuring variations and comparing them to predefined tolerances; (10) processing of raw droplet PCR data to correct for variations and increase test accuracy and performance; (11) incorporating control assays preferably using a single excitation source; and/or (12) quantifying one or more genetic targets by amplifying more than one genetic target in a single droplet.

[0039] Further aspects of the present disclosure are presented in the following sections: (I) definitions, (II) system overview, (III) exemplary instrument controls and calibrators, (IV) exemplary amplification controls, (V), exemplary multi-channel detection, and (VI) exemplary self-normalization of test signals.

[0040] Technical terms used in this disclosure have the meanings that are commonly recognized by those skilled in the art. However, the following terms may have additional meanings, as described below.

[0041] Signal—detectable and/or detected energy and/or information. Any of the signals detected, after detection, may be described as signals and/or data. For example, detected droplet signals may provide test signals and test data, control signals or control data, reference signals and reference data, calibration signals and calibration data, transformed signals and transformed data, or any combination thereof, among others.

[0042] Transform—to change one or more values, and/or the number, of signals of a data set using one or more mathematical and/or logical operations. Transformation of a set of signals may produce a transformed set of the signals by changing values of one or more of the signals and/or by deleting/invalidating any suitable subset of the signals. Signal transformation may include reducing signal variation, deleting/invalidating outlier signals, subtracting a baseline value from signals, reducing the frequency of outliers, reducing the overlap of distributions of positive and negative droplet signals, modifying signals according to a regression line, assigning new values to signals based on comparing signal values to a threshold or range, or any combination thereof, among others.

[0043] Run—an operating period during which a set of droplets, generally droplets of about the same size and including partitions a sample, are tested.

[0044] Oligonucleotide—a nucleic acid of less than about one-hundred nucleotides.

[0045] Exogenous—originating externally. For example, a nucleic acid exogenous to a sample is external to the sample as originally isolated. As another example, a nucleic acid exogenous to an organism or cell is not native to the organism or cell, such as a nucleic acid introduced into the organism or cell by infection or transfection.

[0046] Endogenous—originating internally, such as present in a sample as originally isolated or native to a cell or organism.

[0047] Reporter—a compound or set of compounds that reports the condition of something else, such as the extent of reaction. Exemplary reporters comprise at least one dye, such as a fluorescent dye or an energy transfer pair, and/or at least one oligonucleotide.

[0048] FIG. 3 shows an exemplary system 5740 for performing droplet-based tests of nucleic acid amplification with the aid of controls and/or calibrators. System 5740 may include any combination of a sample/reagent storage/preparation assembly 5742, at least one droplet generator 5744, an amplification assembly, such as a thermal cycler 5746, a detection assembly 5748, and a controller 5750 incorporating a data analyzer 5752 and a feedback and control portion 5754, among others.

[0049] The system may provide at least one flow stream that carries at least one sample and reagents from one or more upstream positions and in a downstream direction to detection assembly 5748. Signals detected from the flow stream, and particularly droplet signals, may be communicated to data analyzer 5752. The data analyzer may analyze the signals to determine one or more test results, control results, calibration results, a quality (e.g., validity, reliability, confidence interval, etc.) of any of the results, or a combination thereof. Any of the results may be communicated to feedback and control portion 5754, which may control and/or adjust control of any of storage/preparation assembly 5742, droplet generator 5744, thermal cycler 5746, detection assembly 5748, and data analyzer 5752, based on the results determined.

[0050] Storage/preparation assembly 5742 may contain and/or supply at least one sample 5756, at least one set of test reagents 5758 (also termed target reagents), one or more control reagents 5760, one or more calibration reagents 5762, or any combination thereof. Any of the samples and/or reagents may be stored and/or supplied separately, may be stored and/or supplied as one or more pre-formed mixtures, and/or may be mixed selectably before they are supplied to a downstream region of the system (e.g., droplet generator 5744, thermal cycler 5746, or detection assembly 5748). Furthermore, any of the samples and/or reagents may travel sequentially from storage/preparation assembly 5742 to droplet generator 5744, thermal cycler 5746, and then detection assembly 5748 for detection of droplet signals. Alternatively, any of the samples and/or reagents may reach the detection assembly without travel through the droplet generator, as indicated at 5764, the thermal cycler, or both, as indicated at 5766. Accordingly, any of the samples and/or reagents disclosed herein may be stored and/or supplied in pre-formed droplets. Droplets may, for example, be pre-formed off-line, either locally or remotely. Pre-formed droplets may be intermixed randomly with droplets formed by droplet generator 5744 before reaching detection assembly 5748, or distinct types of droplets may be detected as spatially and/or temporally separated sets of droplets.

[0051] Test reagents 5758 are any reagents used to test for amplification of one or more targets, such as one or more primary targets, in partitions of a sample. Primary targets generally comprise any targets that are of primary interest in a test. Primary targets may be present at an unknown level in a sample, prior to performing tests on the sample. Test reagents 5758 generally include one or more sets of target reagents conferring specificity for amplification of one or more particular nucleic acid targets to be tested in a sample. Thus, the test reagents may include at least one pair (or two or more pairs) of primers capable of priming amplification of at least one (or two or more) nucleic acid target(s). The test reagents also may comprise at least one reporter to facilitate detecting amplification of each test target, a polymerase (e.g., a heat stable polymerase), dNTPs, and/or the like. The test reagents enable detection of test signals from droplets.

[0052] Control reagents 5760 are any reagents used to control for test signal variation (generally, variation other than that produced by differences in amplification) and/or to interpret results obtained with the test reagents (such as a reliability and/or validity of the results). The control reagents permit control signals and/or reference signals to be detected from droplets, either the same or different droplets from the test signals. Control reagents may be mixed with test reagents prior to droplet formation and/or control droplets containing control reagents may be produced separately from the test droplets and introduced independently of the sample.

[0053] The control reagents may provide instrument controls, that is, controls for variation introduced by the system (and/or its environment). Thus, instrument controls may control for variation in droplet volume, droplet detection efficiency, detector drift, and the like. Reference signals may be detected from droplets containing control reagents that function as instrument controls.

[0054] The control reagents also or alternatively may provide amplification controls, that is, controls that test for secondary/control amplification in droplets. The control reagents thus may include reagents used to test for amplification of at least one secondary or control target in droplets. The secondary/control target may be of secondary interest in a test, and/or may be present at a known or expected level in the sample, among others. In any event, the control reagents may include one or more sets of target reagents conferring specificity for amplification of one or more control nucleic acid targets to be tested in droplets. The control reagents may include at least one pair (or two or more pairs) of primers capable of priming amplification of at least one (or two or more) control nucleic acid target(s). The control reagents also may comprise at least one reporter to facilitate detecting amplification of each control target, a polymerase (e.g., a heat stable polymerase), dNTPs, and/or the like, or any suitable combination of these control reagents may be supplied by the test reagents. Control signals may be detected from control reagents that function as amplification controls.

[0055] Calibration reagents 5762 are any reagents used to calibrate system operation and response. Droplets containing a calibration reagent (i.e., calibration droplets) may be introduced into a flow stream of the system, at any position upstream of the detection assembly, for the purpose of calibrating the system (e.g., calibrating flow rates, excitation power, optical alignment, detector voltage, amplifier gain, droplet size, droplet spacing, etc.). Calibration droplets may be introduced into a flow stream of the system before, during, and/or after introduction of test droplets into the flow stream. In some embodiments, the level of a dye within control droplets may be used to calibrate and/or validate detector response, such as by using a pair of dye concentrations providing calibration signals that bracket an intended measuring range and/or that are disposed near upper and lower ends of the measuring range. For example, droplets of known size and containing one or more known dye concentrations may be prepared off-line and introduced into the system, and/or may be generated by the system. In some embodiments, calibration droplets may comprise fluorescent particles such as quantum dots, polymer beads, etc.

[0056] System 5740 may used to perform a method of analyzing one or more samples. The method may include any suitable combination of the steps disclosed herein, performed in any suitable order.

[0057] Droplets may be obtained. The droplets may be of one type or two or more types. At least a subset, or all, of the droplets may be generated by the system or may be pre-formed off-line. At least a subset of the droplets may include test reagents for testing amplification of a test nucleic acid target. At least a subset of the droplets may include control reagents and/or calibration reagents for testing amplification of a control nucleic acid target. The droplets may contain one or more dyes.

[0058] The droplets may be introduced into a flow stream upstream of a detector. All of the droplets may be introduced into the flow stream at the same position or the droplets, particularly droplets of different types, may be introduced at two or more distinct positions.

[0059] The droplets, in the flow stream, may be subjected to conditions that facilitate amplification. For example, the droplets may be heated and/or may be heated and cooled repeatedly (thermally cycled).

[0060] Signals may be detected from the droplets. The signals may include test signals, control signals, reference signals, calibration signals, or any combination thereof.

[0061] The signals may be analyzed. Analysis may include transforming test signals. Analysis also or alternatively may include comparing test signals and/or transformed test signals to a signal threshold to assign individual droplets as being positive or negative for amplification of a nucleic acid target. A number and/or fraction of target-positive droplets may be determined based on results of the comparison. Analysis further may include estimating a presence of a nucleic acid target in the sample. The estimated presence may be no target in the sample. Estimation of the presence may (or may not) be performed using Poisson statistics.

[0062] FIG. 4 shows selected aspects of system 5740 in an exemplary configuration 5780 for detecting amplification of a nucleic acid target using a first dye and for controlling for system variation during a test using a second dye. In FIG. 4 and in other system configurations presented in succeeding figures of the present disclosure, the terms “droplet generator,” “thermal cycler,” and “detection assembly” are abbreviated “DG,” “TC,” and “DET.”

[0063] Storage/preparation assembly 5742 may supply an amplification mixture to droplet generator 5744. The amplification mixture may incorporate a sample 5756, target reagents 5782 (i.e., test reagents 5758) including a first dye 5784 (dye 1), and a second dye 5786 (dye 2). The second dye and the target reagents may be mixed with one another before introduction into system 5740 or may be mixed within the system. Target reagents 5782 may provide primers for amplification of a nucleic acid target, and the first dye may enable detection of whether amplification occurred. The first and second dyes may be fluorescent dyes that are distinguishable optically. The second dye may be a passive reference or instrument control. In other words, the second dye may provide a detectable signal having an intensity that is at least substantially independent of the extent of amplification, if any, of any nucleic acid target.

[0064] Droplet generator 5744 may form droplets of the amplification mixture. The droplets may travel through thermal cycler 5746, to promote amplification of the nucleic acid target, if any, in each droplet. The droplets then may travel to detection assembly 5748. Assembly 5748 may detect, for each droplet, a test signal from the first dye and a reference signal (also termed a control signal) from the second dye.

[0065] FIG. 5 shows exemplary target reagents 5782 and a control reagent 5760 that may be included in system configuration 5780 of FIG. 4. The target and control reagents may permit detection of test signals in a first detection channel 5788 (“channel 1”) and detection of reference signals in a second detection channel 5790 (“channel 2”). The first and second channels may represent distinct wavelengths and/or at least substantially nonoverlapping wavelength ranges.

[0066] Target reagents may include a reporter, such as a probe 5792, and target-specific forward and reverse primers 5794. Probe 5792 may be an energy transfer probe (e.g., a TAQMAN probe) including a nucleic acid, such as an oligonucleotide 5796, that binds to amplified target, and an energy transfer pair connected to strand 5796. The energy transfer pair may, for example, be formed by first dye 5784 and a quencher 5798.

[0067] Control reagent 5760 may include second dye 5786. The second dye may (or may not) be connected to a nucleic acid, such as an oligonucleotide 5800. Connection to the oligonucleotide may be covalent and/or through a binding interaction. Connection of the second dye to an oligonucleotide or other water-soluble molecule may improve retention of the second dye in the aqueous phase of a droplet and/or may facilitate distribution of the dye throughout the aqueous phase, among others.

[0068] FIG. 6 shows a flowchart illustrating of an exemplary approach to correcting for system variation using system configuration 5780 (FIG. 4), and, optionally, the reagents illustrated in FIG. 5. Test signals (i.e., target signals) and reference signals may be detected from the same droplets. For example, test signals may be detected in a first channel and reference signals may be detected in a second channel. Graphs illustrating coincident detection of test signals and reference signals are shown at 5810, 5812, respectively.

[0069] Test signal variation may introduce errors in data processing. For example, graph 5810 shows substantial variation in the intensity of the test signals detected. As a result, some of the test signals may be erroneously classified as positives or negatives. In the present illustration, two false positives are marked. However, variation of the test signals may be mirrored by variation of the reference signals detected from the same droplets. Accordingly, the test signals may be transformed based on the reference signals, indicated at 5814, to correct for variation in the test signals, as shown in a graph 5816, which plots the transformed test signals. The test signals may be transformed by any suitable operation or set of operation involving the reference signals. For example, the test signals may be transformed through dividing test signals by reference signals, such as dividing each test signal by its corresponding reference signal, which may be described as normalizing the test signals. Alternatively, the test signals may be transformed based on the reference signals by, for example, baseline subtraction, distance from the regression line, or the like. A transformation may compensate for variations in the test channel. This compensation or correction may make the test signals (i.e., negative test signals and/or positive test signals) more uniform in value and/or more Gaussian. The transformation also or alternatively may reduce the frequency of outliers and/or the overlap of the distributions of positive and negative signals.

[0070] FIG. 7 shows selected aspects of system 5740 in an exemplary configuration 5830 for (a) detecting amplification of a nucleic acid target in a set of droplets and (b) system calibration and/or correction for system variation in another set of droplets. Configuration 5830 is similar to configuration 5780 of FIG. 4, except that target reagents 5782 and control reagent 5760 are not in the same droplets. Accordingly, the target reagents and the control reagent may be supplied to respective distinct droplet generators of the system, indicated at 5832, may be supplied to the sample droplet generator at different times, or the control reagent may be supplied in pre-formed droplets that do not pass through the droplet generator, indicated at 5834, 5836. Since the target reagents and the control reagent are not in the same droplets in this configuration, the control reagent may include the same dye as the target reagent (i.e., first dye 5784) or may include a distinct dye (such as second dye 5786).

[0071] FIG. 8 shows an exemplary graph 5850 of fluorescence signals that may be detected over time from a flow stream of system configuration 5830 (FIG. 7) during system calibration, indicated at 5852, and sample testing, indicated at 5854. Calibration and sample testing may be performed without or with mixing of calibration and test droplets.

[0072] Calibration and sample testing may be performed serially, without mixing of droplet types, using the same dye (and/or detection of the same wavelength(s)). By keeping calibration and test droplets separate, the distributions of test and calibration signal intensities may overlap. For example, calibration droplets and test droplets may be separated temporally in the flow stream, such that each type of droplet is identifiable based on its time of arrival at the detection assembly. The time of arrival may be calculated based on the relative time of introduction of each droplet type into the flow stream and the velocity of the flow stream. Thus, the calibration and test droplets may not (or may) be distinguishable based on signal intensity, but may be distinguishable temporally. In particular, the test and calibration droplets may be separated by a temporal (and spatial) gap 5856, which may identify a transition between droplet types. The use of temporal gaps also may permit introduction of a set of calibration droplets within a set of test droplets (i.e., within a test run), with a gap preceding and following the set of calibration droplets, to provide identification of each transition to a different droplet type. Stated differently, calibration may be performed during sample testing, by inserting calibration droplets into a train of test droplets, such that the train of test droplets is divided into two or more discrete groups.

[0073] Calibration droplets may include two or more types of droplet, which may be introduced separately or intermixed. For example, FIG. 8 shows a set of stronger calibration signals 5858 followed by a set of weaker calibration signals 5860 produced by distinct types of calibration droplets. Stronger and weaker calibration signals 5858, 5860 may correspond generally in intensity to respective positive test signals 5862 and negative test signals 5864. In other embodiments, only one type or three or more types of calibration droplet may be used, and may be configured respectively to provide one or three or more intensities of calibration signals.

[0074] Calibration and sample testing alternatively may be performed with calibration and test droplets randomly intermixed and thus not distinguishable temporally. Intermixed calibration and test droplets may be distinguishable by incorporating distinguishable dyes into the respective droplet types and, optionally, by detection of the distinguishable dyes at respective distinct wavelengths. Alternatively, or in addition, calibration droplets and test droplets may be distinguishable according to signal intensity detected at the same wavelength(s) and optionally from the same dye. In particular, calibration droplets may be designed to have one or more signal intensities outside the signal range of test droplets (i.e., the signal range provided by the collective distribution of signal intensities from negative and positive test droplets (e.g., see FIG. 2)). Thus, calibration droplets may be identified based on their calibration signals having signal intensities above and/or below the signal range of test droplets.

[0075] FIG. 9 shows a flowchart 5880 of an exemplary approach to correcting for signal variation during an amplification test using system configuration 5830 of FIG. 7. The approach illustrated in FIG. 9 distinguishes types of droplet signals, namely, test droplet signals 5882 and reference droplet signals 5884, based on differences in signal intensity detected in the same detection channel, as described above for calibration droplets. In particular, test droplets may produce a range 5886 of signal intensities, and reference signals 5884 may have intensities below (or above) the range. Accordingly, the distinct types of droplets may be interspersed randomly in the flow stream.

[0076] The reference droplets may be formed with the same amount (or two or more discrete amounts) of dye. Accordingly, without signal variation generated by the system, the reference droplets should produce reference signals of the same intensity. Variation in reference signal intensity may be mirrored by corresponding changes in the intensity of test signals. For example, in graph 5888, the intensity of reference signals 5884 and negative test signals 5890 show a gradual increase with respect to time. As a result, test signals from amplification-negative droplets may produce false positives 5892.

[0077] Variation in test signals 5882 may be reduced by transforming the test signals, indicated at 5894, based on reference signals 5884, to produce normalized test signals 5896 presented in graph 5898. Transformation may, for example, be performed by transforming each test signal based on one or more reference signals temporally proximate to the test signal, a weighted average of reference signals temporally proximate to the test signal, a sliding window of averaged reference signals that overlaps the test signal, or the like. Transformation before comparing test signals to a threshold may reduce the incidence of false positives, as shown here, the incidence of false negatives, or both.

[0078] FIG. 10 show selected aspects of system 5740 of FIG. 3, with the system in an exemplary configuration 5910 for testing amplification of at least a pair of nucleic acid targets in the same droplets. System configuration 5910 may form an amplification mixture, which is supplied to droplet generator 5744. The amplification mixture may incorporate a sample 5756, test amplification reagents 5858, control amplification reagents 5912, and at least one control template 5914. Any combination of the sample, test reagents, control reagents, and control template may be mixed with one other before introduction into system 5740, or may be mixed within the system. Test reagents 5758 and control reagents 5912 may provide primers for respective amplification of at least one test target and at least one control target.

[0079] Amplification of the test and control targets may, for example, be detected via a first dye and a second dye, respectively, which may be included in respective first and second reporters (e.g., first and second probes). Signals from the first and second dyes may be detected in distinct (e.g., at least substantially nonoverlapping) first and second channels (i.e., a test channel and a control channel) as test signals and control signals, respectively.

[0080] Control template 5914 may comprise exogenous molecules of the control target. In contrast, the sample may be tested for a presence of endogenous molecules of the test target. The control template 5914 may be present in any suitable amount to provide any suitable average number of control template molecules per droplet, to generate a desired fraction of droplets positive for the control template. For example, the number of template molecules provided by template 5914 may be substantially less than an average of one per droplet, such as an average of about 0.1, 0.05, 0.02, or 0.01 molecule per droplet. Accordingly, the number/concentration of control template molecules may be selected such that the frequency of amplification of both test and control targets in the same droplet is low, which may minimize competition that may be caused by amplification of both test and control targets. For example, the control template may be present in no more than about one in five droplets.

[0081] The frequency of amplification of the control target may be determined by performing an analysis with the system. In some embodiments, this frequency may be compared with one or more previously determined frequencies of amplification for the control target and/or may be compared with an expected value for the frequency provided by a manufacturer. In any event, a control value may be determined, with the control value corresponding to a number and/or fraction of the droplets that are amplification-positive for the control nucleic acid target.

[0082] Control signals acquired in the control channel may be used to measure and/or verify the quantitative accuracy of a run and/or the measurement precision of the system during two or more runs. The control signals also or alternatively may be used to interpret a test result, such as the quality of test data measured from a sample, for example, to verify the quantitative accuracy of the test data and/or to determine the validity and/or reliability of the test data. The test result may be interpreted based on control value determined. For example, the test result may be determined as being invalid if the control value is less than a threshold value. Furthermore, data acquired from the control channel, such as signals from amplification-negative control droplets, may provide reference signals, as described above in relation to FIG. 6. In other words, test signals may be transformed using control signals that functions as reference signals, to normalize the test signals.

[0083] FIG. 11 shows selected aspects of system 5740 of FIG. 3, with the system in another exemplary configuration 5920 for testing amplification of at least a pair of nucleic acid targets in the same droplets. System configuration 5920 differs from configuration 5910 of FIG. 10 by including a different set of control amplification reagents 5922 (or a second set of test amplification reagents) and by the absence of an exogenous control template. Control reagents 5922 may amplify a control target that is known or expected to be present in sample 5756, and/or that has a known or expected representation with respect to a bulk nucleic acid population present in the sample (e.g., total DNA, total genomic DNA, genomic DNA from a particular species of organism, total RNA, total mRNA, etc.). In contrast, target reagents 5758 may amplify a test target that has an unknown presence in the sample and/or an unknown presence in with respect to the bulk nucleic acid population. In any event, amplification of the control target may be used to determine the quality of test data measured from a sample, such as to verify the quantitative accuracy of the test data and/or to determine the reliability of the test data. Furthermore, an amount of control target determined to be present in the sample may provide a standard against which an amount of test target determined to be present in the sample can be compared and/or normalized. In some embodiments, a control target is selected that is rare in the sample, such as a target representing a particular gene mutation. By selecting a rare control target, amplification of the control target can indicate the limit of detection of a test target and/or whether amplification of a low-abundance test target can occur. In some embodiments, the control target may be replaced by a second test target with an unknown presence in the sample (before testing).

[0084] FIG. 12 shows exemplary test target reagents 5758 and control target reagents 5912 (or 5922) that may be included in system configuration 5910 (or 5920) of FIG. 10 (or 11), to permit detection of amplification signals in a different detection channel (i.e., channels 1 and 2, respectively) for each nucleic acid target. Test target reagents for channel 1 are described above in relation to FIG. 5. Control target reagents 5912 (or 5922) may be similar in general structure to the test target reagents, but different with respect to the nucleic acid sequences of the primers and probes, to provide test target and control target specificity, respectively. Also, the test and control probes may include distinct dyes 5784, 5786 and/or distinct energy transfer partners 5798, 5930 (e.g., distinct quenchers suitable for the respective dyes). In other embodiments, at least one of the probes may be replaced by a reporter including an intercalating dye, such as SYBR® Green.

[0085] FIGS. 13 and 14 show representative portions of exemplary data that may be obtained using system configuration 5910 or 5920 and the reagents of FIG. 12. The figures show exemplary graphs 5940-5946 of fluorescence signals that may be detected over time from a flow stream of the system using different detection channels, namely, a test channel (channel 1) that detects test data and a control channel (channel 2) that detects control data. In FIG. 13, graph 5940 of the test data contains no positive droplet signals. In contrast, graph 5942 of the control data identifies positive droplet signals, such as a positive signal 5948, at a frequency of about one in ten. Thus, the control data demonstrates that amplification in the droplets is not inhibited substantially and suggests that the lack of positive signals from the test data is due to an absence or undetectable level of the test target in the sample. Accordingly, the control data supports and helps to validate the negative result in the test data. In contrast, control graph 5946 of FIG. 14 shows no amplification of the control target (a substantially larger data set may be analyzed to demonstrate that the control result holds). The control data of graph 5946 thus indicates that amplification of the test target also is inhibited (or the sample is defective, such as too dilute (configuration 5920)), and that the negative test result is not valid.

[0086] FIG. 15 shows selected aspects of system of FIG. 3, with the system in an exemplary configuration 5960 for testing amplification of a pair of nucleic acid targets in respective different (i.e., nonoverlapping) sets of droplets. Configuration 5960 may be similar to that of configuration 5910, except that control reagents 5912 and control template 5914 are not mixed with sample 5756 and test target reagents 5758. Instead, droplets containing the control reagents and the control template may be formed separately in the system, indicated at 5962, or may be supplied as pre-formed droplets that are introduced into the flow stream downstream of droplet generator 5744, indicated at 5964.

[0087] FIG. 16 shows a pair of exemplary graphs 5980, 5982 of fluorescence signals that may be detected over time from a flow stream of system configuration 5960 of FIG. 15 using different detection channels. Graph 5980 plots fluorescence signals detected from a first channel, which detects amplification, if any, of a test target. Graph 5982 plots fluorescence signals detected from a second channel, which detects amplification, if any, of a control target. Successful amplification of the control target, as shown here, may, for example, verify and/or measure aspects of the system, such as operation of the thermal cycler and/or the detection assembly, the quality of the reagents, fraction of amplification-positive droplets, or any combination thereof, among others.

[0088] In configuration 5960, the test and control reagents are disposed separately in distinct droplets, so droplet signals in the first and second channels are not coincident, that is, they are not detected at the same time. In other embodiments, the control target may, instead, be a second test target and the control template may, instead, be another sample (or the same sample). Thus, the use of at least two detection channels permits droplets for distinct amplification tests to be interspersed in the flow stream.

[0089] FIG. 17 shows a pair of graphs 5990, 5992 illustrating exemplary absorption and emission spectra of fluorescent dyes that may be used in the system of FIG. 3. The dyes are arbitrarily labeled dye 1 and dye 2, respectively. However, either dye may be used to detect test signals or control signals in the various system configurations disclosed herein. Moreover, while illustrated here for two distinguishable dyes, the system may be used for detection and analysis with three, four, or more distinguishable dyes.

[0090] Each graph plots the intensity of absorption (“AB”), indicated at 5994, 5996, and emission (“EM”), indicated at 5998, 6000, for the corresponding dye. The dyes may have substantially overlapping absorption spectra, such that the same wavelength of light may be utilized to excite both dyes. In contrast, the dyes may exhibit Stokes shifts (i.e., the difference (in wavelength or frequency units) between the maxima of the absorption and emission spectra) of different magnitudes. For example, dye 1 may exhibit a smaller Stokes shift and dye 2 a larger Stokes shift, or vice versa. Accordingly, the emission spectra of the dyes may be substantially shifted with respect to one another. As a result, emission from the two dyes may be detected at least substantially independently of one another in different detection channels, such as a detection channel that detects light of a first wavelength or wavelength range (e.g., λ1) and another detection channel that detects light of a second wavelength or wavelength range (e.g., λ2).

[0091] FIG. 18 is a schematic diagram illustrating exemplary use of the fluorescent dyes of FIG. 17 in an exemplary embodiment 6010 of system 5740 of FIG. 3. Droplets 6012 containing dyes 1 and 2, either in the same droplets or different sets of droplets, may be carried in a flow stream 6014 in a channel 6016. Flow stream 6014 may pass through a detection area 6018 established by an embodiment 6020 of detection assembly 5748.

[0092] Detection assembly 6020 may include a light source 6022 for exciting the fluorescent dyes in the droplets and at least one detector 6024 for detecting light emitted from the droplets. Light source 6022 may, for example, include an LED or laser that emits at least substantially a single wavelength of excitation light. Alternatively, or in addition, the light source may include at least one excitation optical filter that excludes other wavelengths of light emanating from the light source. Detector 6024 may be equipped with detection optics 6026, 6028 (e.g., beamsplitters, emission optical filters, separate detectors) that permit emitted light from the dyes to be detected separately.

[0093] Exemplary fluorescent dyes that may detected using system 6010 include a fluorescein derivative, such as carboxyfluorescein (FAM), and a PULSAR 650 dye (a derivative of Ru(bpy)₃). FAM has a relatively small Stokes shift, while PULSAR 650 dye has a very large Stokes shift. Both FAM and PULSAR 650 dye may be excited with light of approximately 460-480 nm. FAM emits light with a maximum of about 520 nm (and not substantially at 650 nm), while PULSAR 650 dye emits light with a maximum of about 650 nm (and not substantially at 520 nm). Carboxyfluorescein may be paired in a probe with, for example, BLACK HOLE Quencher™1 dye, and PULSAR 650 dye may be paired in a probe with, for example, BLACK HOLE Quencher™2 dye.

[0094] Test signals may be normalized using methods different from those described above in relation to FIGS. 6 and 9. In particular, the methods illustrated in FIGS. 6 and 9 involve transformation of test data with reference data detected (a) in a different detection channel (FIG. 6) or detected (b) in different droplets (FIG. 9). This section describes methods that transform test data using aspects of itself rather than another data set.

[0095] FIG. 19 shows a flowchart 6040 illustrating an exemplary method of correcting for system fluctuations during a test. The method involves processing a set of droplet test signals, shown in a first graph 6042, to produce a transformed set of test signals, shown in a second graph 6044. Negative test signals 6046 and positive test signals 6048 each should have respective constant values over time if there is no system variation. However, system variation, such as the negative drift over time illustrated in graph 6042, may produce false negatives, such as a false negative signal 6050, and/or false positives. Transformation of the test signals may be performed to correct for system variation before the test signals are used to estimate a presence of a test target in sample being tested. In particular, individual test signals may be transformed differently using the test data, accordingly to the temporal position of each test signal. For example, each test signal may be transformed using temporally proximate test data, such as normalization of each test signal with respect to a sliding window that averages a subset of the test signals including or adjacent the test signal. The subset of the test signals used may be provisionally negative, positive, or negative plus positive test signals, any of which may be re-assigned as negative/positive after transformation. For example, graph 6044 shows re-assignment of false negative signal 6050 as positive after transformation.

[0096] FIG. 20 shows a flowchart 6060 illustrating an exemplary method of transforming droplet signals based on the width of respective signal peaks providing the droplet signals. The flowchart involves graphs 6062, 6064, which represent test data before and after transformation, respectively.

[0097] Graph 6062 presents test data in which the width and height of each droplet peak is shown. (Here, each droplet peak is presented as a square wave to simplify the presentation. However, in other embodiments, each droplet peak may be detected as having any suitable shape, such as a wave with sloped leading and trailing sides.) The width of a droplet fluorescence peak may be used to determine the size and volume of each droplet, if droplet signals are detected in a flow stream with known flow rate, generally within a channel of fixed geometry. Knowing the volume of sample that is tested for amplification in droplets may be required for accurately determining the concentration/number of target molecules in the sample. If droplets of uniform size are desired, peak width may be used to identify droplets of sizes that are outside the desired range. For example, in FIG. 20, peaks 6066, 6068 having widths outside a predefined range are excluded from the data set. The droplet signals also may be transformed based on width, to provide transformed test data (i.e., graph 6064), that has been corrected for volume variation and/or variation in peak width.

[0098] The disclosure set forth above may encompass multiple distinct inventions with independent utility. Although each of these inventions has been disclosed in its preferred form(s), the specific embodiments thereof as disclosed and illustrated herein are not to be considered in a limiting sense, because numerous variations are possible. The subject matter of the inventions includes all novel and nonobvious combinations and subcombinations of the various elements, features, functions, and/or properties disclosed herein. The following claims particularly point out certain combinations and subcombinations regarded as novel and nonobvious. Inventions embodied in other combinations and subcombinations of features, functions, elements, and/or properties may be claimed in applications claiming priority from this or a related application. Such claims, whether directed to a different invention or to the same invention, and whether broader, narrower, equal, or different in scope to the original claims, also are regarded as included within the subject matter of the inventions of the present disclosure.