Method for classifying monitoring results from analytical sensor system arranged to monitor molecular interactions

Abstract

To provide a simpler and faster, or at least an alternative, method to known methods for classifying monitoring results from an analytical sensor system arranged to monitor molecular interactions.SOLUTION: Disclosed is a method for classifying monitoring results from an analytical sensor system 20 arranged to monitor molecular interactions at a sensing surface, where detection curves representing progress of the molecular interactions with time are produced. The method comprises: a step 100 of acquiring a set of detection curves; a step 101 of fitting a first mathematical model to the set of detection curves; a step 102 of calculating a set of features from the set of detection curves and fitted mathematical model; a step 103 of, based on the calculated set of features, classifying each detection curve into quality classification group; and a step of, based on the classification, determining detection curves to be used in kinetic analysis of the monitored molecular interactions.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] This specification relates to a method for classifying monitoring results from an analytical sensor system arranged to monitor intermolecular interactions, and a computer program and computer program product comprising program code means for performing the method. This document also relates to an analytical system for detecting molecular binding interactions and classifying monitoring results.

Background Art

[0002] Analytical sensor systems arranged to monitor intermolecular interactions, such as those between biomolecules, in real time are often based on label-free biosensors such as optical biosensors. A typical such biosensor system is the BIACORE® instrumentation, which uses surface plasmon resonance (SPR) to detect interactions between molecules in a sample and a molecular structure immobilized on a sensing surface. The sample is passed over the sensor surface, and the progress of binding directly reflects the rate at which the interaction occurs. After injecting the sample, buffer is flowed, during which the detector response reflects the rate of dissociation of the complex on the surface.

[0003] A typical output from a BIACORE® system and similar biosensor systems is a response graph or curve that describes the time-dependent progress of an intermolecular interaction, including an association phase portion and a dissociation phase portion. This response graph or detection curve, usually displayed on a computer screen, is often referred to as a binding curve or “sensorgram”.

[0004] Using a BIACORE® system (and similar sensor systems), it is possible to determine a plurality of interaction parameters for molecules used as ligands and analytes. These parameters include kinetic rate constants relating to the binding (association) and dissociation of intermolecular interactions, as well as interaction affinity.

[0005] In the evaluation of the dynamics of a ligand - analyte system, especially when the evaluation involves a large number of detection curves, it can be cumbersome and difficult even for an experienced user of the sensor system to classify the detection curves as good or bad. Low - quality curves should be excluded from further analysis of the interaction dynamics because they can have an adverse effect on the evaluation. Different users can classify detection curves differently (i.e., as good or bad quality), so the evaluation of the ligand - analyte system can vary and be user - dependent. Low - quality curves can include, for example, an unstable baseline, air spikes, responses below the baseline, etc., and can be caused by, for example, a contaminated flow system or running buffer, the ligand capture approach used, or a detergent concentration that is too low in the running buffer.

[0006] WO2003081425 describes a method and an analysis system for the data processing of a large set of detection curves representing intermolecular interactions on a sensor surface, which assist the user when classifying curves with respect to quality. The quality evaluation of the detection curves involves steps of selecting quality parameters for each detection curve based on at least one quality - related parameter (e.g., baseline slope, air spike, vibration), where each parameter is defined by at least one quality descriptor; calculating for each detection curve value with respect to the descriptor; and calculating for each detection curve a quality classification indicating the quality of the detection curve for all detection curves in the set of curves. Detection curves classified as outliers are selected and a validation procedure is performed to determine whether to include the deviant detection curves in subsequent kinetic analysis.

[0007] To use the method of WO2003081425, an understanding of the algorithms used for data processing is required, and this method also takes a considerable amount of time for fine - tuning (training). SUMMARY OF THE INVENTION PROBLEM TO BE SOLVED BY THE INVENTION

[0008] The object of the present disclosure is to provide a simpler and faster, or at least an alternative method, to known methods for classifying monitoring results from an analytical sensor system arranged to monitor intermolecular interactions. A further object is to provide a computer program and a computer program product comprising product code means for carrying out the method. Further, one object is to provide an analytical system for monitoring molecular binding interactions and classifying the monitoring results.

[0009] The invention is defined by the appended independent patent claims. Non-limiting embodiments arise from the dependent patent claims, the accompanying drawings and the following description.

Means for Solving the Problems

[0010] According to a first aspect, there is provided a method for classifying monitoring results from an analytical sensor system arranged to monitor intermolecular interactions, the method comprising generating a detection curve representing the progression of the intermolecular interaction over time. The method includes the following steps: obtaining a set of detection curves, the set of detection curves including one or more detection curves representing intermolecular interactions at respective molecular concentrations; fitting a mathematical model to the set of detection curves; calculating a set of features from the set of detection curves and the fitted mathematical model, the calculated set of features including: the association rate constant ka divided by the standard error of ka, the dissociation rate constant kd divided by the standard error of kd, the maximum binding capacity Rmax divided by the standard error of Rmax, the mass transfer limit value tc divided by the standard error of tc, the delayed binding response B divided by Rmax, and the mean square error (MSE) between the detection curve and the fitted mathematical model divided by the squared delayed binding response (B 2 ), and including three or more of the above; and classifying each detection curve into a quality classification group indicating the quality of the detection curve based on the calculated set of features.

[0011] The monitored intermolecular interaction may occur in a substance detection biosensor system with a sensing surface, such as an optical biosensor, where one molecule, the ligand, may be attached / immobilized on the sensor surface and the other molecule, the analyte, passes through the sensor surface and the progress of analyte-ligand binding can be monitored over time. Analytical sensor systems based on other detection principles, such as electrochemical systems, are also possible.

[0012] The set of detection curves can include at least one detection curve. When including two or more detection curves, for example, 2 to 10 detection curves, each detection curve in the set of detection curves represents the intermolecular interaction (between the same ligand and analyte) at different molecular concentrations, i.e., different analyte concentrations.

[0013] The mathematical model that fits the set of detection curves is a model that describes the intermolecular interaction over time. The model may be, for example, a 1:1 binding model. Alternatively, a heterogeneous ligand binding model, a heterogeneous analyte binding model, or a divalent analyte binding model can be used.

[0014] Divide the parameter values of ka, kd, Rmax, and tc by their respective standard errors. The standard error of a parameter is a measure indicating how important the parameter is with respect to the degree of approximation of the fitting mathematical model. Rmax is the calculated Rmax from the fitting. The delayed binding response B is the maximum binding response in the detection curve relative to the baseline in the detection curve. The characteristic B of the delayed binding response is normalized by dividing it by Rmax, and the characteristic MSE of the mean squared error is normalized by dividing it by the squared delayed binding response B 2 by dividing it by the squared delayed binding response B. The characteristics are normalized so as to be applicable to all types of input data.

[0015] The steps of fitting a mathematical model to a set of detection curves, calculating a set of features from the set of detection curves and the fitted mathematical model, and classifying each detection curve into a quality classification group may be performed by an experienced user of the analysis sensor system. Alternatively, the step of quality classification may be performed by an ANN or an expert system trained on a plurality of training sets of detection curves using the respective fitted mathematical models, the calculated set of features, and the quality classification. The parameters used in the training are determined by an experienced user.

[0016] The detection curves are classified into quality classification groups that indicate the quality of the detection curves using this method. Curves of low quality should be excluded from further analysis of the dynamics of the monitored interaction as they may have an adverse effect on the evaluation. Low-quality curves may include, for example, an unstable baseline, air spikes, responses below the baseline, etc., and may be caused by, for example, a contaminated flow system or running buffer, the ligand capture approach used, or a detergent concentration that is too low in the running buffer. What constitutes a sufficiently good quality of the detection curve and whether to include the detection curve in the dynamic analysis of the monitored interaction depends on the purpose of the analysis / experiment.

[0017] The number of different quality classification groups may be two in one example, with the first group including detection curves of sufficiently good quality and the second group including detection curves of low quality. Then, the group including the detection curves of sufficiently good quality can be selected for the dynamic analysis of the monitored intermolecular interaction.

[0018] In another example, the number of quality classification groups can be 100, where group 1 includes detection curves of low quality and group 100 includes detection curves of very good quality. The user can adjust the strictness of the method, for example, set a cut-off at >50, whereby only the detection curves classified into classification groups 51 - 100 can be included in the dynamic analysis of the monitored interactions. Again, whether the quality of the detection curve is good enough and where to place the cut-off depends on the purpose of the analysis / experiment.

[0019] The set of calculated features may be three or more. Using a large number of features in this method can improve the classification of detection curves when the features complement each other and emphasize something that can distinguish what an experienced user considers to be sufficiently good quality from not-so-good quality. If there are too many features that do not contribute to the classification, the training of the neural network and the quality of the classification may decrease. It is also possible to calculate additional features not described above and include them in the step of classifying the detection curves.

[0020] The mathematical model can be selected from a 1:1 binding model, a heterogeneous ligand binding model, a heterogeneous analyte binding model, and a divalent analyte binding model.

[0021] Based on the classification, the detection curves to be used in the dynamic analysis of the monitored intermolecular interactions can be determined.

[0022] As described above, the determination of which of the classified detection curves should be included in the dynamic analysis of the monitored interactions and where to place the cut-off depends on the purpose of the analysis / experiment.

[0023] Through such dynamic analysis, the association constant, ka, the dissociation constant, kd, and the interaction affinity for the interaction can be obtained.

[0024] The method can further include a step of determining a second mathematical model to be used in the dynamic analysis.

[0025] The second mathematical model used in the dynamic analysis of the detection curve does not necessarily have to be the same mathematical model as the one used in the above-described quality evaluation procedure.

[0026] The second mathematical model can be selected from a 1:1 binding model, a heterogeneous ligand binding model, a heterogeneous analyte binding model, and a bivalent analyte binding model.

[0027] Classifying the detection curves into quality classification groups can be performed by an artificial neural network or an expert system.

[0028] The classification can be carried out by one ANN / expert system or two or more such systems operating in cooperation with each other.

[0029] The artificial neural network can be trained using multiple sets of detection curves representing the progress of different intermolecular interactions over time. The artificial neural network, for each set of detection curves, has a) a set of features calculated from the set of detection curves and a mathematical model fitted to the set of detection curves, where the calculated set of features includes: the association rate constant ka divided by the standard error of ka; the dissociation rate constant kd divided by the standard error of kd; the maximum binding capacity Rmax divided by the standard error of Rmax; the mass transfer limit value tc divided by the standard error of tc; the delayed binding response B divided by Rmax; and the mean squared error MSE between the detection curve and the fitted mathematical model divided by the squared delayed binding response B 2 divided by B; including three or more of these, and b) the classification of each detection curve into a quality classification group.

[0030] The classification of the detection curves into quality classification groups is obtained by visual inspection of the detection curves by experienced users and is input into an artificial neural network (ANN) along with the calculated features. The mathematical model used to fit the data is fixed for the training of the ANN and can be a 1:1 binding model, a heterogeneous ligand model, a heterogeneous analyte model, or a divalent analyte binding model.

[0031] When using such a trained ANN to perform the classification step of the method described above, the ANN is preferably trained on the same set of calculated features that is input into the trained ANN for performing the classification step. However, in some embodiments, in the classification step, it is possible to input the set of calculated features into a trained ANN that includes a smaller number of calculated features than those used in the training of the ANN.

[0032] Determining the second mathematical model to be used in the kinetic analysis can be performed by an artificial neural network or an expert system.

[0033] The artificial neural network can be trained using a plurality of detection curves representing the progress of intermolecular interactions over time, and the artificial neural network is provided with a classification of the detection curves regarding the mathematical model that fits the detection curves.

[0034] The classification of the mathematical models for use in the kinetic analysis of different detection curves can be evaluated by experienced users, and the results are input into the ANN.

[0035] According to a second aspect, there is provided an analysis system for monitoring molecular binding interactions and classifying monitoring results, the analysis system including: a) a sensor device including at least one sensing surface, detecting means for detecting intermolecular interactions on the at least one sensing surface, and means for generating a detection curve representing the progress of the interaction over time; and b) data processing means for classifying each detection curve into a first or second group, the data processing means performing steps b) to d).

[0036] According to a third aspect, there is provided a computer program including program code means for performing the above-described method when the program is executed on a computer.

[0037] According to a fourth aspect, there is provided a computer program product including program code means stored on a computer-readable medium for performing the above-described method when the program is executed on a computer. BRIEF DESCRIPTION OF THE DRAWINGS

[0038]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Best Mode for Carrying Out the Invention

[0039] The present disclosure relates to analytical sensor methods, particularly biosensor-based methods, in which intermolecular interactions are studied and the results are presented in real time as a detection curve, often called a sensorgram, as the interaction progresses.

[0040] Biosensors can be based on various detection methods. Typically, such methods include, but are not limited to, substance detection methods such as piezoelectric, optical, thermo-optical, and surface acoustic wave (SAW) device methods, as well as electrochemical methods such as potentiometry, conductivity, amperometry, and capacitance methods. Regarding optical detection methods, representative methods include methods for detecting the surface concentration of a substance, such as reflection-optical methods (including both internal and external reflection methods), angle, wavelength, or phase decomposition, such as ellipsometry and evanescent wave spectroscopy (EWS), the latter including surface plasmon resonance (SPR) spectroscopy, Brewster angle refractometry, critical angle refractometry, frustrated total reflection (FTR), evanescent wave ellipsometry, scattered total internal reflection (STIR), optical waveguide sensors, evanescent wave-based imaging (such as critical angle resolved imaging, Brewster angle resolved imaging, SPR angle resolved imaging, etc.). Furthermore, for example, photometry methods based on evanescent fluorescence (TIRF) and phosphorescence can also be used, as well as using waveguide interferometers.

[0041] One commonly used detection principle is surface plasmon resonance (SPR) spectroscopy. An exemplary type of SPR-based biosensor is sold under the trade name BIACORE® (hereinafter referred to as the "BIACORE device"). These biosensors utilize SPR-based substance sensing technology to provide "real-time" label-free binding interaction analysis between a surface-bound ligand and the analyte of interest.

[0042] The BIACORE device includes a light-emitting diode (LED), a sensor chip including a glass plate covered with a thin gold film, an integrated fluid cartridge that provides a flow of liquid over the sensor chip, and an array of photodetectors. Incident light from the LED is internally totally reflected at the glass / gold interface and detected by the array of photodetectors. At a specific angle of incidence (the "SPR angle"), surface plasmon waves are generated in the gold layer and detected as an intensity loss "or dip" of the reflected light. The phenomenon of SPR associated with the BIACORE device depends on the resonant coupling of monochromatic p-polarized light incident on the thin metal film through the prism and glass plate, and the oscillation of conductive electrons called plasmons in the metal film on the opposite side of the glass plate. These oscillations produce an evanescent field that extends from the surface into the liquid flow by a distance of about one wavelength (~1 μm). When resonance occurs, light energy is lost to the metal film by collective excitation of the electrons, and the intensity of the reflected light decreases at a sharply defined angle of incidence, the SPR angle. The SPR angle depends on the refractive index within the reach of the evanescent field near the metal surface.

[0043] As described above, the SPR angle depends on the refractive index of the medium close to the gold layer. In the BIACORE device, dextran is typically bound to the gold surface, and the analyte-binding ligand is bound to the surface of the dextran layer. The analyte of interest is injected in solution form onto the sensor surface via the fluid cartridge. Since the refractive index in the vicinity of the gold film depends on (i) the refractive index of the solution (which is constant), and (ii) the amount of material bound to the surface, the binding interaction between the bound ligand and the analyte can be monitored as a function of the change in the SPR angle.

[0044] FIG. 1 shows a schematic diagram of an analysis system 20 for detecting molecular binding interactions and classifying monitoring results. The analysis system includes a BIACORE™ device, which includes a sensor device, a chip 1 with a sensing surface, a gold film 2, supports a capture molecule 3 (such as an antibody), and is exposed to a sample flow together with an analyte 4 (such as an antigen) through a flow channel 5. Monochromatic p-polarized light 6 from a light source 7 (LED) is coupled to a glass / metal interface 9 by a prism 8, where the light is totally reflected. The intensity of the reflected light beam 10 is detected by a detection means 11 such as an optical photodetector array.

[0045] A typical output from the BIACORE device is a "sensorgram", which is a plot of the response (measured in "resonance units" or "RU") as a function of time. An increase of 1,000 RU corresponds to an increase in the amount of substance on the sensor surface of about 1 ng / mm². When a sample containing an analyte contacts the sensor surface, the ligand bound to the sensor surface interacts with the analyte in a step called "association". This step is shown on the sensorgram by an increase in RU when the sample is first brought into contact with the sensor surface. Conversely, "dissociation" typically occurs when the sample flow is replaced, for example, by a buffer flow. This step is shown on the sensorgram by a decrease in RU over time as the analyte dissociates from the surface-bound ligand.

[0046] A representative sensorgram of a BIACORE device is shown in Figure 2. Figure 2 shows a sensing surface having an immobilized ligand (e.g., an antibody) that interacts with an analyte in a sample. The y-axis represents the response (here resonance units (RU)), and the x-axis represents time (here seconds). First, buffer is passed over the sensing surface, giving a "baseline response" in the sensorgram. During sample injection, an increase in signal is observed due to the binding (i.e., association) of the analyte to a steady-state condition where the resonance signal reaches a plateau. At the end of sample injection, the sample is replaced with a continuous flow of buffer, and the decrease in signal reflects the dissociation or release of the analyte from the surface. The slope of the association / dissociation curve provides valuable information about the kinetics of the interaction, and the height of the resonance signal represents the surface concentration (i.e., the response resulting from the interaction is related to the change in the concentration of the substance on the surface). The analysis system 20 shown in Figure 1 includes means 12 for generating a detection curve representing the progress of the interaction over time.

[0047] Detection curves or sensorgrams generated by biosensor systems based on other detection principles have a similar appearance.

[0048] Sometimes, the generated sensorgrams are of unacceptable quality for various reasons and thus must be discarded. Figure 3 shows examples of two acceptable sensorgrams and two unacceptable sensorgrams. The two curves on the left are both within the acceptable range. On the other hand, the upper right curve is too unstable, and the lower right curve is distorted by an air peak (bubbles in the fluid flow). Quality control of the sensorgrams is usually done by the user creating an overlay plot of the curves to be analyzed and visually searching for abnormalities in the curves. Unacceptable curves can be caused, for example, by an unstable baseline, a response below the baseline, etc., and can be caused, for example, by a contaminated flow system or running buffer, the ligand capture approach used, or too low a detergent concentration in the running buffer.

[0049] Current trends in biosensor systems are towards the development of high throughput systems that can generate large sets of sensorgrams in a relatively short time. If the throughput has already increased moderately, it is easily understandable that even an experienced user cannot practically inspect all the sensorgrams at once to evaluate the quality of the sensorgrams. Furthermore, since different users can classify detection curves differently (i.e., as being of sufficiently good or low quality), the evaluation of the ligand-analyte system can vary and be user-dependent.

[0050] FIG. 4 shows a method for classifying monitoring results from an analytical sensor system that generates detection curves representing the progress of intermolecular interactions over time, and FIG. 1 shows data processing means 13 for performing such a task. A set of detection curves including n (e.g., 1 to 8) different detection curves is obtained (100) from the analytical sensor system. The n detection curves represent monitored intermolecular interactions at n different molecular concentrations, i.e., different analyte concentrations.

[0051] A mathematical model is fitted to a set of n detection curves (101). The mathematical model may be a 1:1 binding model. This is the simplest model for dynamic evaluation. The model describes a 1:1 interaction at the surface. The dynamic parameters include ka - the association rate constant for the formation of the ligand - analyte complex; kd - the dissociation rate constant for the ligand - analyte complex; Rmax - the maximum binding capacity; and tc - the mass transfer limit. Usually, these parameters are fitted over a wide range, but they can also be made local or constant. Since the model corresponds to the Langmuir isotherm for the adsorption of substances onto a solid surface, it may also be called the Langmuir model. The movement of the analyte from the bulk solution to the surface is directly proportional to the bulk analyte concentration, and the proportionality constant is the mass transfer coefficient, which is a function of the flow rate, flow cell dimensions, and the diffusion characteristics of the analyte. The unit of this constant is RU·M - 1s - 1. For a globular protein with a molecular weight of about 50,000 daltons, a typical value of the mass transfer coefficient is about 108 RU·M - 1s - 1. If the reported values vary widely in magnitude (e.g., 1012 or 1014), it may indicate that the parameter is not important for the fit (i.e., the observed binding is not limited by mass transfer). The term tc related to the rate of movement of the analyte from the bulk solution to the surface is included in the 1:1 binding model.

[0052] Other mathematical / dynamic models that can be used and fitted to a set of detection curves are the heterogeneous ligand binding model, the heterogeneous analyte binding model, and the divalent analyte binding model. These models are not as commonly used as the 1:1 binding model.

[0053] The heterogeneous ligand model accounts for the presence of two ligand species that bind to the analyte independently of each other. These species may be different molecules or different binding sites on the same ligand molecule. Here, the kinetic parameters are: ka1 - the association rate constant for the formation of the ligand 1 - analyte complex; ka2 - the association rate constant for the formation of the ligand 2 - analyte complex; kd1 - the dissociation rate constant for the ligand 1 - analyte complex; kd2 - the dissociation rate constant for the ligand 2 - analyte complex. The complexity of the mathematical model is limited to two ligand species.

[0054] The heterogeneous analyte model is mainly intended for situations where two analytes of different sizes are deliberately mixed. The model describes this competitive state and returns two sets of rate constants, one set for each reaction. The kinetic parameters are: ka1 - the association rate constant for the formation of the analyte 1 - ligand complex; ka2 - the association rate constant for the formation of the analyte 2 - ligand complex; kd1 - the dissociation rate constant for the complex analyte 1 - ligand; and kd2 - the dissociation rate constant for the complex analyte 2 - ligand. The heterogeneous analyte model is useful for indirectly determining the kinetics of the smaller analyte due to competition with the larger analyte. The response contributions from both analytes were considered, but the high molecular weight analyte was the cause of the major component of the observed sensorgram. Concentrations and molecular weights are required for both analytes. If the absolute molecular weight is unknown, relative values can be entered without affecting the fitting results. The model cannot evaluate the interaction if the ratio and relative size of the analytes are unknown.

[0055] The divalent analyte binding model describes the binding of a divalent analyte to an immobilized ligand, where one analyte molecule can bind to one or two ligand molecules. It is assumed that the two analyte sites are equivalent. The model may be relevant, in particular, to studies using signaling molecules that bind to immobilized cell surface receptors (where receptor dimerization is common) and intact antibodies that bind to immobilized antigens. As a result of the binding of one analyte molecule to two ligand sites, the overall binding is enhanced compared to 1:1 binding. This effect is often referred to as avidity. The kinetic parameters are: ka1 - the association rate constant for the formation of the analyte-ligand site 1 complex; ka2 - the association rate constant for the formation of the analyte-ligand site 2-ligand site 2 complex; kd1 - the dissociation rate constant for the analyte-ligand site 1 complex; and kd2 - the dissociation rate constant for the analyte-ligand site 1-ligand site 2 complex. Since the binding at the second site does not change the surface substance, no response occurs. For this reason, the association rate constant for the second interaction is reported in units of RU-1s-1 and can only be obtained in M-1s-1 if the conversion factor between RU and M is available. Similarly, the values of the overall affinity or avidity constants are not reported. In general, it is preferred to immobilize the divalent reactant and thereby avoid the complex situations caused by the combined affinity resulting from multivalent binding (avidity). In some cases, the avidity effect can be reduced by using very low ligand levels and high analyte concentrations. The low ligand levels provide sparsely distributed ligands where the probability of two ligand molecules being within the reach of a single analyte is low. The high analyte concentration competes with the second-site binding and is thus favorable for the formation of 1:1 complexes.

[0056] After fitting the mathematical model to a set of n detection curves, a set of features from the detection curves and the fitted mathematical model is calculated (102). The set of features can include three or more of the following: - The association rate constant ka divided by the standard error of ka, - The dissociation rate constant kd divided by the standard error of kd, - The ratio of the maximum binding capacity Rmax to its standard error, - The ratio of the mass transfer limit value tc to its standard error, - The ratio of the delayed binding response B to Rmax, and - The mean square error (MSE) between the detection curve and the fitted mathematical model divided by the square of the delayed binding response (B 2 ).

[0057] The delayed binding response B is the maximum binding response in the detection curve relative to the baseline in the detection curve. See the description of Figure 6.

[0058] The features calculated and used within the set of features, and the number of features, may depend on the mathematical model used.

[0059] Based on the set of calculated features, each detection curve is classified into a quality classification group indicating the quality of the detection curve103, where it is shown as four different classification groups here. The number of different quality classification groups may vary. As described above, curves of low quality may have an adverse effect on the evaluation and should be excluded from the dynamic analysis of the monitored interaction. What constitutes a sufficiently good quality of the detection curve, and whether to include the detection curve in the dynamic analysis of the monitored interaction, depends on the purpose of the analysis / experiment.

[0060] Then, the classification group / groups containing detection curves of sufficiently good quality can be selected for the dynamic analysis104 of the monitored intermolecular interaction. In the example of Figure 4, the detection curves classified into the first and second classification groups are considered to have a sufficiently good quality for the dynamic analysis of the monitored intermolecular interaction, while the detection curves classified into the third and second groups are of lower quality and are not included in the dynamic analysis. The user can adjust the strictness of the method and set a cutoff, for example, such that only the detection curves of the first group or the detection curves of groups 1 - 3 are used in the dynamic analysis.

[0061] The mathematical model can determine parameters such as ka and kd for the interaction by fitting it to a detection curve classified as having a sufficiently good quality for the dynamic analysis of the interaction. This mathematical model may be the same as the mathematical model used to classify the curve, or it may be a different model. The model can be selected from any of the binding models described above.

[0062] Step 103 of classifying the detection curves into different quality classification groups can be performed by an artificial neural network (ANN) or an expert system. To do this, the ANN needs to be trained on data from a plurality of sets 200 of detection curves representing the progress of different intermolecular interactions over time. The data input into the ANN for training includes, for each set 202 of detection curves, the features calculated from the set of detection curves and the mathematical model 201 fitted to the set of detection curves. The mathematical model used to fit the data can be fixed for the training of the ANN. The set of features can include three or more of the following: - The association rate constant ka divided by the standard error of ka, - The dissociation rate constant kd divided by the standard error of kd, - The maximum binding capacity Rmax divided by the standard error of Rmax, - The mass transfer limit value tc divided by the standard error of tc, - The delayed binding response B divided by Rmax, and - The mean squared error (MSE) between the detection curve and the fitted mathematical model divided by the squared delayed binding response (B 2 ).

[0063] The ANN is also provided with the classification of each detection curve into a quality classification group indicating the quality of the detection curve 202. Such classification of the detection curve can be determined by visual inspection of the detection curve by an experienced user and input into the ANN model. The training of the ANN model is shown in FIG. 5.

[0064] Next, such a trained ANN model returns the quality classification of the unknown detection curve to the quality classification group.

[0065] The ANN or expert system may also be trained to assist the user of the analysis sensor system in determining the mathematical model to be used in the dynamic analysis of the detection curve. The artificial neural network can be trained using a plurality of detection curves representing intermolecular interactions on the sensor surface, and the artificial neural network is provided with the classification of the detection curves determined by an experienced user regarding which mathematical model fits the detection curve appropriately. The trained ANN can predict with high accuracy which mathematical model to use for a set of data.

[0066] The trained ANN mimics how to classify the detection curve into the quality classification group of an experienced user using both the detection curve data features and the associated fitting parameters. The training may be standard ANN training including splitting the entire data set into a training set and a validation set, in which case 80% of the data is used for training and 20% of the data is used for validation. The ANN should be trained on the training data and not on the validation set. Using one hidden layer and several hidden nodes, an ANN that is a generalized detection curve classifier can be obtained. If two hidden layers and too many hidden nodes are allowed, the ANN is very good at adapting to the training data but poor at classifying / predicting the validation set or other non-training data.

[0067] In one example, the trained ANN model included one hidden layer, 15 hidden nodes, a cross-entropy loss function, and a learning rate of 0.1, and was built using Microsoft Azure Machine Learning Studio in the cloud. For training the ANN model, for each detection curve, feature data was calculated, normalized with Biacore dynamic applications if necessary, classification was added, and it was sent to the cloud. The data could be added to the entire training dataset, and older data of the same origin could be removed from the training set. Different algorithms may be used to handle duplicates, etc. There may be several ANN models, one for each research group or company in the cloud, or there may be a shared ANN shared among some users within the community. The output from training is an ANN model that can be used for quality classification of detection curves.

[0068] Figure 7 shows an example of how the kinetic screen machine learning application processes data when training and predicting.

[0069] In this example, three datasets with six detection curves each are used. The curves marked with circles in each set are the detection curves, and some have different concentrations. The curves with solid black lines are the curves fitted using a mathematical model, in this case, a 1:1 binding Langmuir model. There is one fitted curve for each detection curve. An expert in the application can label the datasets as follows: R = rejected series, and A = accepted series based on data quality.

[0070] For each set, features are calculated using both the detection curves, the mathematical model curves, and the results from the fitted mathematical model.

[0071] [Table 1]

[0072] Classifying each detection curve into quality classification groups indicating the quality of the detection curve is our method of training a machine learning model, in this case an artificial neural network (ANN), using all (or at least three) of the quality label data and the feature data set. The ANN uses a cross-entropy loss function and a learning rate of 0.1, with one hidden layer and 15 hidden nodes. Each column is normalized using a Gaussian normalization function, and the ANN is trained and evaluated using Microsoft Azure Machine Learning Studio in the cloud. And the purpose of the ANN is to accurately classify quality labels using only features. Once trained, when a new set of data without quality labels is analyzed, the ANN can predict whether such a data set should be automatically accepted or rejected based only on features, which can save the user's time and also enable the operation of the system by non-expert users.

[0073] By continuously training and developing these models, experienced users can easily correct the classification results and retrain the models in the cloud. Furthermore, by using a trained ANN as described above to classify detection curves, the classification process can be sped up, especially when there are a large number of detection curves during analysis, and the evaluation results are less user-dependent and non-experienced users can be assisted in the evaluation of monitored intermolecular interactions.

[0074] The foregoing description includes multiple specificities, but these are not intended to limit the scope of the concepts described herein, but rather should be construed as merely providing an illustration of some exemplary embodiments of the described concepts. The scope of the concepts described herein fully encompasses other embodiments that may be apparent to those skilled in the art, and thus it should be understood that the scope of the concepts described herein is not limited. References to a single element are not intended to mean "only one" unless explicitly stated as such, but rather are intended to mean "one or more." All structural and functional equivalents of the elements of the above-described embodiments known to those skilled in the art are expressly incorporated herein and are intended to be included herein.

Claims

1. A method for classifying monitoring results from an analytical sensor system arranged to monitor intermolecular interactions, wherein a detection curve representing the progression of said intermolecular interaction over time is generated, said method comprising: a) obtaining a set of detection curves (100), said set of detection curves comprising one or more detection curves representing intermolecular interactions at respective molecular concentrations; b) fitting a mathematical model to said set of detection curves (101); c) calculating a set of features from said set of detection curves and the fitted mathematical model (102), said calculated set of features comprising: the association rate constant ka divided by the standard error of ka; the dissociation rate constant kd divided by the standard error of kd; the maximum binding capacity Rmax divided by the standard error of Rmax; the mass transfer limit value tc divided by the standard error of tc; the delayed binding response B divided by Rmax; and The mean squared error (MSE) between the detected curve and the fitted mathematical model, divided by the squared delay coupling response (B 2 ), including three or more of the above; d) classifying each detection curve into a quality classification group indicative of the quality of said detection curve based on said calculated set of features (103). A method as described above.

2. The method according to claim 1, wherein said mathematical model is selected from a 1:1 binding model, a heterogeneous ligand binding model, a heterogeneous analyte binding model, and a bivalent analyte binding model.

3. The method according to claim 1 or 2, wherein said intermolecular interaction is monitored at a sensing surface.

4. The method according to any one of claims 1 to 3, further comprising determining (104) a detection curve to be used in a kinetic analysis of said monitored intermolecular interaction based on said classification.

5. The method according to claim 4, further comprising determining a second mathematical model to be used in said kinetic analysis.

6. The method according to claim 5, wherein said second mathematical model is selected from a 1:1 binding model, a heterogeneous ligand binding model, a heterogeneous analyte binding model, and a bivalent analyte binding model.

7. The method according to any one of claims 1 to 6, wherein the step of classifying the detection curve into a quality classification group (103) is performed by an artificial neural network or an expert system.

8. The artificial neural network is trained using a plurality of sets (200) of detection curves representing the progression of intermolecular interactions over time, and for each set of detection curves, the artificial neural network a) a set of features (202) calculated from the set of detection curves and a mathematical model fitted to the set of detection curves, wherein the calculated set of features is: - the association rate constant ka divided by the standard error of ka, - the dissociation rate constant kd divided by the standard error of kd, - the maximum binding capacity Rmax divided by the standard error of Rmax, - the mass transfer limit value tc divided by the standard error of tc, - the delayed binding response B divided by Rmax, and - The mean squared error (MSE) between the detected curve and the fitted mathematical model, divided by the squared delay coupling response (B 2 ), a set containing three or more of: b) the classification (202) of each detection curve into a quality classification group, The method according to claim 7, wherein is provided.

9. The method according to claim 5 or 6, wherein the step of determining the second mathematical model used for the dynamic analysis is performed by an artificial neural network or an expert system.

10. The method according to claim 9, wherein the artificial neural network is trained using a plurality of sets of detection curves representing the progression of intermolecular interactions over time, and the artificial neural network is provided with the classification of the detection curves as to which mathematical model fits the detection curves.

11. An analysis system (20) for detecting molecular binding interactions and classifying monitoring results, comprising: a) a sensor device (1) comprising at least one sensing surface (2), detection means (11) for detecting intermolecular interactions at the at least one sensing surface (2), and means (12) for generating a detection curve representing the progression of the interaction over time; and b) data processing means (13) for classifying each detection curve into a quality classification group, the data processing means performing steps b) to d) according to any one of claims 1 to 10.

12. A computer program comprising program code means for performing the method according to any one of claims 1 to 10 when the program is executed on a computer.

13. A computer program product comprising program code means stored on a computer-readable medium for executing the method according to any one of claims 1 to 10 when the program is executed on a computer.

Images