Reversible model of cardiac function

Abstract

Described herein are systems and methods for predicting changes in cardiac function based on in silico modeling. In various embodiments, the systems and methods are configured to perform inverse mechanistic modeling of cardiac function, enabling effective and efficient prediction of cardiac function and opening the possibility of such approaches being used in many drug discovery and development applications. In one example, in vitro responses of first and second cardiac cell cultures are obtained under baseline and perturbed conditions, respectively. A mechanistic model is then obtained or generated to determine an estimate(s) of the functional response. Upon inputting an input parameter vector, first and second distribution values ​​are obtained from the mechanistic model, from which a first relative distribution shift can be determined, which is indicative of a change in the function of the first cardiac cell culture due to the perturbed conditions.

Description

[Background technology]

[0001] Mechanistic models offer a powerful approach for modeling the dynamic behavior of biological systems, such as cardiac cell cultures. Such models typically require estimating model parameters from empirical data. However, obtaining sufficient data to fit these model parameters is often a challenging and time-consuming task. Moreover, incorporating these mechanistic models into downstream tasks, such as drug development and discovery, is often difficult. Successfully incorporating such models into the drug development and discovery process could offer numerous benefits, including personalizing computational models using patient-specific cells (i.e., patient-derived cells), predicting patient response to therapy, understanding the mechanisms underlying poorly understood cardiac diseases, developing disease models, optimizing treatments, examining drug safety and off-target effects / side effects, and simulating exercise and other significant cardiac loads.

[0002] Therefore, new approaches are needed to model the dynamic behavior of cardiac cell cultures, which can be used as a basis for downstream drug development and discovery tasks. Summary of the Invention [Means for solving the problem]

[0003] According to one aspect of the present disclosure, there is provided a method for predicting changes in cardiac function. The method includes: acquiring, by one or more processors, a first plurality of in vitro responses including measurements of a first functional response of a first cardiac cell culture under reference conditions; and acquiring, by one or more processors, a second plurality of in vitro responses including measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with the first perturbation of the first cardiac cell culture. The method further includes acquiring, by the one or more processors, a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector. The method further includes determining, by the one or more processors, a first distribution of values ​​for the input parameter vector, such that an estimate of a first functional response obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses, and determining, by the one or more processors, a second distribution of values ​​for the input parameter vector, such that an estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates a measurement from the second plurality of in vitro responses. The method further includes determining, by the one or more processors, a first relative distribution shift based on a comparison of the first distribution of values ​​and the second distribution of values, the first relative distribution shift being indicative of a change in function of the first cardiac cell culture due to the perturbation condition.

[0004] According to another aspect of the present disclosure, there is provided a method for inverse mechanistic modeling of cardiac function. The method includes: acquiring, by one or more processors, a first plurality of in vitro responses including measurements of at least one functional response of a first cardiac cell culture under a first set of conditions; and acquiring, by the one or more processors, a mechanistic model of cardiac function, the mechanistic model predicting one or more values ​​of the at least one functional response based on an input parameter vector. The method further includes determining, by the one or more processors, a first value distribution for the input parameter vector, such that an estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the first value distribution approximates a measurement from the first plurality of in vitro responses. The method further includes determining, by the one or more processors, a shifted first value distribution for the input parameter vector, such that an estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the shifted first value distribution satisfies a predetermined functional response criterion. The method further includes determining, by the one or more processors, a first relative distribution shift based on a comparison of the distribution of first values ​​and the shifted distribution of first values, wherein the first relative distribution shift is indicative of a predicted change in cardiac function from the first set of conditions to a set of conditions associated with a predetermined functional response criterion.

[0005] According to yet another aspect of the present disclosure, systems and methods for predicting changes in cardiac function are disclosed. The systems and methods may include acquiring, by one or more processors, a first plurality of in vitro responses from a bioreactor containing a first cardiac cell culture (e.g., of an artificial cardiac tissue), the first plurality of in vitro responses including measurements of a first functional response of the first cardiac cell culture under reference conditions. The systems and methods further include acquiring, by the one or more processors, a second plurality of in vitro responses from the bioreactor including measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with the first perturbation of the first cardiac cell culture. The systems and methods further include acquiring, by the one or more processors, a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector. The systems and methods may further include determining, by one or more processors, a first distribution of values ​​for the input parameter vector, such that an estimate of a first functional response obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses. The systems and methods may further include determining, by one or more processors, a second distribution of values ​​for the input parameter vector, such that an estimate of a first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates a measurement from the second plurality of in vitro responses. The systems and methods may further include determining, by the one or more processors, a first relative distribution shift based on a comparison of the first distribution of values ​​and the second distribution of values, the first relative distribution shift being indicative of a change in function of the first cardiac cell culture due to the perturbation condition.The systems and methods may further include generating one or more outputs that identify or are indicative of one or more of: (a) the target molecule as a potential drug candidate, and / or (b) the potential effect of an unknown compound, based on the change in function of the first cardiac cell culture due to the perturbation conditions.

[0006] In accordance with the above and the disclosure herein, the present disclosure includes applying certain features or aspects with or by using certain machines, e.g., bioreactors. In various embodiments, the bioreactor may include a device configured to grow or manipulate human tissue (e.g., human tissue such as cardiac tissue). Additionally or alternatively, the bioreactor may include a sensor assembly configured to detect one or more functional responses of the tissue within the device.

[0007] Additionally, the present disclosure includes converting or reducing a particular item to a different state or thing, e.g., sensing the functional response of human tissue, such as cardiac tissue, as a waveform, e.g., in a bioreactor, by a sensor assembly, and converting or reducing it to another state or thing, e.g., encoding or converting shifts in distribution values ​​of cardiac cell cultures across different conditions (e.g., baseline and perturbed conditions), thereby encoding functional changes in cardiac cell cultures by direct comparison, linking and categorizing the functional changes. These various shifts and conversions can be exploited for applications such as predicting the effects of unknown compounds and building databases of relative shifts for use in drug discovery and development.

[0008] Furthermore, the present disclosure includes improvements in computer functionality or other technologies, at least in part because the present disclosure discloses systems and methods for implementing in silico (in silico) predictions of the effects (e.g., mechanism of action, toxicity, etc.) of unknown real-world compounds. The systems and methods herein, when implemented on underlying devices, enable the generation of data (e.g., features) defining change(s) in the dynamic behavior of cardiac tissue or its associated functions between different states (e.g., baseline and perturbed states). Such data can be used in drug discovery and development, reducing or eliminating the need for empirical testing and the associated processor and data usage. For example, when deployed on underlying systems, the systems and methods of the present disclosure can perform fewer iterations and use fewer computational resources than prior art related systems and methods. That is, the present disclosure describes improvements in the functionality of the computer itself or “any other technology or technical field” because the generated data provided by the in silico algorithms described herein can cause the underlying computer system to use fewer processing and memory resources compared to prior art systems and methods. At least one reason for this is that in silico data algorithms can generate or determine data of predicted cardiac tissue function (e.g., for various control and test conditions) without requiring a wide range of tests and / or empirical computer simulations using numerous computational cycles and data. Thus, the use of in silico algorithms results in fewer computational cycles, i.e., fewer iterations, and less impact on the underlying computing device, compared to the aforementioned prior art systems and methods. In other words, the systems and methods of the present disclosure are improvements over the prior art.This is at least in part because prior art systems and methods require empirical or trial-and-error approaches that may involve experimentation with real-world human tissue (e.g., cardiac tissue), potentially resulting in or requiring the use of large databases, memory, and processors to arrive at similar real-world results or simulated results that are identical or similar thereto. In contrast, the disclosed systems and methods describe the generation and / or use of bioreactors for growing and testing tissue (e.g., engineered cardiac tissue) to define limited data sets specific to the tissue, which requires less memory and / or processing utilization compared to traditional approaches that use or require large, unknown, and potentially irrelevant data sets. Moreover, the disclosure herein enables prediction of potential treatments for obtaining required changes in functional responses. For example, in some embodiments, this may include generating output(s) based on, for example, a change in function or condition of a cardiac cell culture from a first condition (e.g., a baseline condition) to a second condition (e.g., a perturbed condition). This output may be used for drug discovery or development to more accurately identify or indicate one or more of: (a) target molecules as potential drug candidates, and / or (b) potential effects of unknown compounds, resulting in improvements to the underlying machine or device on which the in silico algorithm is deployed.

[0009] Additionally, the present disclosure relates to improvements over other technologies or technical fields, at least in part because the systems and methods of the present disclosure provide a robust, efficient, and comparable encoding of cardiac tissue behavior that can be used to improve the efficiency and performance of several downstream drug discovery and development tasks. This may be accomplished, for example, by a mechanistic model determined based on or otherwise generated using in vitro responses, including measurements of the functional response(s) of cardiac cell cultures. The mechanistic model can be deployed on an underlying computing device or system, thereby improving accuracy and prediction in performing drug discovery and development tasks, as described herein.

[0010] Furthermore, the present disclosure includes certain features other than what are well-understood, routine, and conventional activities in the art and / or adds unconventional steps that otherwise limit the disclosure to certain useful applications, such as systems and methods for predicting changes in cardiac function based on in silico modeling, which may be used, for example, for effective and efficient output of cardiac function predictions, and which may be used or applied in drug discovery and drug development applications.

[0011] The advantages will become more apparent to those skilled in the art from the following description of preferred embodiments shown and described by way of illustration. As will be realized, the present invention is capable of other and different embodiments, and its details are capable of modification in various respects. Accordingly, the drawings and description are to be regarded as illustrative in nature, and not as restrictive. Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which: [Brief explanation of the drawings]

[0012] [Figure 1] 1 illustrates a general architecture for generating a reversible model of cardiac function, according to one aspect of the present disclosure. [Figure 2]1 illustrates a mechanistic model of cardiac function, according to one embodiment of the present disclosure. [Figure 3] 1 illustrates a Markov Chain Monte Carlo (MCMC) parameter fitting algorithm, according to an embodiment of the present disclosure. [Figure 4] 1 illustrates a hierarchical Bayesian parameter fitting algorithm according to an embodiment of the present disclosure. [Figure 5] 1 illustrates a generative adversarial network (GAN) parameter fitting algorithm, according to an embodiment of the present disclosure. [Figure 6] 1 illustrates inverse mechanism modeling according to one aspect of the present disclosure. [Figure 7] 1 illustrates a method for determining changes in cardiac function according to one aspect of the present disclosure. [Figure 8] 7 illustrates a series of steps performed in conjunction with the method of FIG. 6 according to an embodiment of the present disclosure. [Figure 9] 7 illustrates a series of steps performed in conjunction with the method of FIG. 6 according to an embodiment of the present disclosure. [Figure 10] 1 illustrates a method for inverse mechanism modeling of cardiac function according to one aspect of the present disclosure. [Figure 11] 10 illustrates a series of steps performed in conjunction with the method of FIG. 9 according to an embodiment of the present disclosure. [Figure 12] 1 illustrates a system for acquiring functional response data of cardiac cell cultures, according to one aspect of the present disclosure. [Figure 13] 1 shows a waveform including a functional response of tissue over a period of time, according to an embodiment of the present disclosure. [Figure 14] 1 illustrates an exemplary computing system according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0013] Technical Field The present disclosure relates to the generation of reversible in silico models of cardiac function. In particular, but not limited to, the present disclosure relates to the use of functional response data to generate reversible in silico models of cardiac function. In particular, the functional response data includes, but is not limited to, the in vitro functional response of artificial cardiac tissue.

[0014] The ability to make accurate and biologically plausible predictions of cardiac function from in silico models of cardiac cell cultures is a critical step in identifying target molecules as potential drug candidates and identifying the potential effects of unknown compounds (e.g., mechanism of action, toxicity, etc.) The present disclosure describes systems and methods for inverse mechanistic modeling of cardiac function that enable effective and efficient prediction of cardiac function, thereby opening the door for such approaches to be used in many drug discovery and development applications.

[0015] Figure 1 shows the general architecture for generating a reversible model of cardiac function.

[0016] The general architecture includes a non-reversible phase 102 and a reversible phase 104. The non-reversible phase 102 involves multiple in vitro responses 106 of cardiac cell cultures 108 under one or more conditions, a mechanistic model of cardiac function 110, a parameter fitting algorithm 112, and a distribution of values ​​114. The mechanistic model 110 takes an input parameter vector 116 as input and determines estimated functional response values ​​(e.g., a first set of estimated functional response values ​​118-1) as output. In one embodiment, a database 120 is used to store the distribution of values ​​114 and other such information. The reversible phase 104 involves a shift 122 of the distribution of values ​​114 learned during the non-reversible phase 102 and a set of predicted functional response values ​​(e.g., a first set of predicted functional response values ​​124-1) determined from the mechanistic model 110 based on the shifted distribution of values.

[0017] The non-reversible phase 102 attempts to learn a parameterization of the mechanistic model 110 (i.e., a distribution of values ​​114 for an input parameter vector 116). As a result of this parameterization, the mechanistic model 110 generates predicted functional response values ​​of the cardiac cell culture 108 that approximate the empirical functional response values ​​of the cardiac cell culture 108. The parameterization is learned by a parameter fitting algorithm 112 that determines the distribution of values ​​114, such that the mechanistic model 110 generates functional response values ​​that closely match the functional response values ​​of the multiple in vitro responses 106. Once the distribution of values ​​114 is determined, accurate in silico predictions of cardiac function can be obtained by iteratively obtaining predictions from the mechanistic model 110 parameterized according to the distribution of values ​​114. In this way, the mechanistic model 110 parameterized according to the distribution of values ​​114 serves as an in silico model of the cardiac cell culture 108.

[0018] Thus, the distribution of values ​​114 of the input parameter vector 116 provides an efficient characterization of one or more conditions of the cardiac cell culture 108. This link between the distribution of values ​​114 and one or more conditions allows changes in cardiac function between conditions to be encoded by relative shifts in the distribution of values ​​(e.g., a relative shift from a distribution associated with a first condition to a distribution associated with a second condition). Because the distribution shifts are relative, different shifts can be directly compared, allowing functional changes to be linked and classified. As will be described in more detail below, this provides a powerful framework for encoding functional changes in cardiac cell cultures, which can be used for applications such as predicting the effects of unknown compounds and building databases of relative shifts for use in drug discovery and development.

[0019] The reversible phase 104 utilizes a mechanistic model 110 parameterized according to a distribution of values ​​114 (learned during the non-reversible phase 102) to predict the functional response of the cardiac cell culture 108. Because the distribution of values ​​114 provides an accurate in silico model of the cardiac cell culture 108 under one or more conditions, shifts applied to the distribution of values ​​114 provide biologically plausible and meaningful predictions of how changes in one or more conditions may affect the functional response of the cardiac cell culture 108. For example, a shift in a parameter associated with calcium binding may be predicted to increase the contractile force of the cardiac cell culture. Thus, one or more of the distributions of values ​​114 associated with one or more parameters of the input parameter vector 116 may be shifted until the predicted functional response value obtained from the mechanistic model 110 reaches a predetermined criterion or value. The resulting shift in the one or more distributions represents a change in the function of the cardiac cell culture that may be required to achieve a desired change in the functional response.

[0020] As previously discussed, relative shifts in parameter distributions provide effective characterizations, or signatures, of changes in cardiac function, and these relative shifts are used to identify candidate drugs or treatments that are predicted to alter cardiac function.

[0021] Cardiac cell culture 108 is either a two-dimensional (2D) culture (e.g., a monolayer) or a three-dimensional (3D) tissue formed from a 3D culture. As described in more detail below in connection with Figure 12, cardiac cell culture 108 is grown in a bioreactor (e.g., bioreactor 1202 shown in Figure 12) from cells seeded into the wells of the bioreactor.

[0022] Cardiac cell cultures 108 are generally associated with one or more conditions that may be considered measures of the state(s) of the cardiac cell cultures 108. The conditions (states) of cardiac cell cultures may change over time, and therefore the conditions are time-dependent. Thus, the multiple in vitro responses 106 are associated with the condition(s) of the cardiac cell cultures 108 at the time the in vitro responses 106 were obtained or at the time the data from which the in vitro responses 106 were obtained was collected.

[0023] The condition may be a reference condition or a perturbed condition. Generally, a reference condition refers to a condition that provides a baseline comparison to a perturbed condition. In embodiments, a reference condition is a condition associated with a control setting or environment. As such, a reference condition may alternatively be referred to as a control condition or baseline condition. A reference condition may correspond to a cardiac cell culture in a predetermined, natural, or unaltered state (i.e., no drug or agent administration). Alternatively, a reference condition may correspond to a vehicle-treated cardiac cell culture. A perturbed condition refers to a condition in which some perturbation has been applied to the cardiac cell culture. Examples of perturbed conditions include administration of a drug or compound (i.e., a perturbant), a disease state, a different cell line, or a physical perturbation applied to the cardiac cell culture. Thus, a perturbed condition may alternatively be referred to as a treatment condition. In the case of a perturbation involving a drug or compound, the condition is further related to an effect associated with the drug or compound, such as mechanism of action or toxicity. Considering a range of different perturbation conditions, cardiac cell cultures may be associated with more than one perturbation condition (e.g., diseased cardiac cell cultures treated with a particular compound).

[0024] As described in more detail below, the link between one or more conditions of the cardiac cell culture and the input parameter vector 116 of the mechanistic model 110 allows for in silico comparisons and identification of changes or shifts in cardiac function.

[0025] The plurality of in vitro responses 106 acquired during the irreversible phase 102 correspond to empirical measurements of the cardiac cell culture 108 under the associated condition(s). The plurality of in vitro responses 106 provide empirical data regarding the functional response of the cardiac cell culture 108 under the associated condition(s) and are used to fit distributions of input parameters to the mechanistic model 110 during the irreversible phase 102. As such, the plurality of in vitro responses 106 may alternatively be referred to as empirical responses, empirical data, experimental responses, experimental data, in vitro data, or empirical functional response data. The plurality of in vitro responses 106 includes a first set of measurements 106-1 (observations or data points) associated with a first functional response of the cardiac cell culture 108. The plurality of in vitro responses 106 may further include a second set of measurements 106-2 (observations or data points) associated with a second functional response of the cardiac cell culture 108, and / or a third set of measurements 106-3 (observations or data points) associated with a third functional response of the cardiac cell culture 108.

[0026] Here, a functional response is understood to be a biophysical response of a cardiac cell culture, such as contractile force, calcium response or calcium properties (e.g., calcium transients), or an electrical response or property. As such, the plurality of in vitro responses 106 includes measurements or values ​​related to one or more of such functional responses of the cardiac cell culture 108. Accordingly, the functional response measurements within the plurality of in vitro responses 106 include one or more of measurements related to the contractile force of the cardiac cell culture 108, measurements related to the calcium properties of the cardiac cell culture 108, and measurements related to the electrical properties of the cardiac cell culture 108. For example, the first set of measurements 106-1 may include contractile force measurements, while the second set of measurements 106-2 and the third set of measurements 106-3 may include calcium property measurements and electrical property measurements, respectively.

[0027] The plurality of in vitro responses 106 are extracted from one or more functional response waveforms of a cardiac cell culture 108. For example, a first waveform may capture the contractile force of a cardiac cell culture over a 30-second period, while a second waveform may capture calcium transients of the cardiac cell culture over the same period. As described in more detail below in connection with FIGS. 12 and 13, a cardiac cell culture in a bioreactor (such as bioreactor 1202 shown in FIG. 12) may be exposed to electrical stimulation at a predetermined pacing frequency, and the changes over time in the cardiac cell culture's functional response (e.g., contractile force, calcium transients, etc.) may be recorded as a waveform. Alternatively, the waveform may capture the changes over time in the cardiac cell culture's functional response in the absence of stimulation (i.e., the spontaneous response of the cardiac cell culture). In either case, each value in the waveform corresponds to the cardiac cell culture's functional response at a respective time point (e.g., a contractile force of 120 μm at time point t1, a contractile force of 110 μm at time point t2, etc.). The measurements or feature values ​​forming the plurality of in vitro responses 106 are extracted from the waveform using a feature extraction process. In one embodiment, peak analysis is performed on the waveform to extract one or more numerical features from each contraction-relaxation of the cardiac cell culture within the waveform. The numerical features extracted from the waveform include twitch amplitude, contraction time, maximum contraction slope, relaxation time, maximum relaxation slope, and twitch duration. In this manner, the plurality of in vitro responses 106 includes multiple measurements or feature values ​​for each peak within the functional response waveform. Alternatively, the measurements include summary statistics such as mean, median, and standard deviation derived from the waveform feature values. In another embodiment, a spectral transform is applied to the waveform to extract a spectral representation of the functional response. The spectral representation takes the form of a spectral density, which provides a frequency-based characterization of the functional response(s). Examples of spectral transforms include the Fourier transform and the maximum entropy transform.

[0028] In one embodiment, the plurality of in vitro responses 106 are obtained either directly or indirectly from a bioreactor (e.g., bioreactor 1202 shown in FIG. 12 ) containing cardiac cell cultures 108. In the case of direct acquisition, feature extraction processing is performed in the bioreactor, and the plurality of in vitro responses 106 are obtained from the bioreactor. In the case of indirect acquisition, waveforms from which the in vitro responses are extracted are obtained from the bioreactor, and a secondary system (e.g., control unit 1204 shown in FIG. 12 ) performs feature extraction processing and obtains the plurality of in vitro responses 106. Advantageously, utilizing a bioreactor such as that described in connection with FIG. 12 for cultivating and maintaining cardiac cell cultures can efficiently obtain large amounts of functional data under a variety of conditions and across a variety of cardiac cell cultures. This is particularly beneficial for parameter fitting purposes, where greater amounts of data enable improved estimation of the distribution of values ​​114 for the input parameter vector 116.

[0029] The in vitro responses 106 are associated with one or more conditions (i.e., states) of the cardiac cell culture 108 at the time the functional response waveforms from which they are derived were acquired. In this manner, the conditions may be considered labeling information (i.e., metadata) associated with the cardiac cell culture 108. As will be described in more detail below, this information is used to identify potential treatments or effects based on a comparison of distribution shifts.

[0030] The plurality of in vitro responses 106 provides high-fidelity empirical measurements of the behavior of cardiac cell cultures under a set of one or more conditions (i.e., baseline or perturbed conditions). Thus, the plurality of in vitro responses provides an efficient, high-fidelity, and high-density data set from which input parameters of the mechanistic model 110 may be fitted to determine a distribution(s) of values ​​for the input parameter vector 116.

[0031] The mechanistic model 110 is an in silico model of cardiac function (i.e., an in silico model of the function of the cardiac cell culture 108). The mechanistic model 110 takes as input one or more input parameters in the form of an input parameter vector 116 and determines as output estimated functional response value(s) (e.g., a first set of estimated functional response values ​​118-1). The input parameter vector 116, together with the initial conditions of the model, determine the dynamic behavior of the mechanistic model 110. As described in more detail below, the parameter values ​​are unknown a priori and are therefore estimated from the multiple in vitro responses 106 (i.e., the input parameter vector 116 is fitted to the multiple in vitro responses 106). Because the multiple in vitro responses 106 are associated with the condition(s) of the cardiac cell culture 108 at the time the responses were obtained or derived, the distribution of values ​​114 for the input parameter vector 116 is also associated with the condition(s) of the cardiac cell culture 108. Thus, the distribution of values ​​114 provides an in silico parameterization of the conditions of the cardiac cell culture 108 such that predicted functional response values ​​generated by the mechanistic model 110 based on the input parameter vector 116 taking values ​​from the distribution of values ​​114 approximate the empirical functional response values ​​of the cardiac cell culture 108 under that condition(s). That is, once learned, the distribution of values ​​114 for the input parameter vector 116 enables the mechanistic model 110 to estimate the function of the cardiac cell culture 108 under that condition(s).

[0032] At a general level, mechanistic model 110 is a mathematical model parameterized by input parameter vector 116. In one embodiment, mechanistic model 110 includes one or more ordinary differential equations (ODEs). ODEs are well suited to modeling dynamic systems such as the heart or its biological subsystems because they allow the various biochemical interactions occurring within such systems to be converted into one or more rate equations.

[0033] 2 illustrates a mechanistic model 200 of cardiac function according to one embodiment of the present disclosure. In one embodiment, the mechanistic model 200 corresponds to the mechanistic model 110 shown in FIG.

[0034] The mechanistic model 200 includes multiple equations (ODEs) that model the dynamic contractile behavior of cardiac cell cultures or tissues to generate an estimate of contractile force at a given time (t). The equations in the mechanistic model 200 are grouped into functional blocks or groups as shown in FIG. 2. The first functional block 202 estimates the intracellular calcium transient, [Ca](t), that drives contraction. The second functional block 204 estimates the fraction of troponin complexes that are calcium bound, A. The third functional block 206 estimates the fraction of cross-bridge (XB) groups involved, G. XB A fourth function block 208 estimates the total fraction of XB that is in a force-producing state, XB A The fifth function block 210 estimates the proportion of XBs that are connected within the group, XB G A sixth function block 212 estimates the average strain of XB, s. A seventh function block 214 estimates the sarcomere length, SL. An eighth function block 216 estimates the length dependence of the thin filament, LDF. thin The ninth function block 218 estimates the length dependence of the thick filament, LDF thick The tenth function block 220 estimates the active tension, T a An eleventh function block 222 estimates the active force, F A Estimate.

[0035] In the first functional block 202 of the mechanistic model 200, the intracellular calcium transient [Ca](t) that drives contraction at time t is estimated as follows:

number

number

[0036] The intracellular calcium transient, [Ca](t), estimated in the first function block 202, is used to estimate the fraction of troponin complexes that are calcium bound, A, in the second function block 204. The second function block 204 estimates the value of A as follows: A=A H *LDF thin (λ)+A L *(1-LDF thin (λ)), (3) During the ceremony,

number

[0037] The estimated fraction of troponin complexes bound to calcium, A, determined in the second functional block 204, is then calculated in the third functional block 206 as the fraction of cross-linked (XB) groups involved in force generation, G XB is used to estimate the proportion of XB groups involved, G XB is estimated as follows:

number

number

number

number

[0038] In a fourth function block 208, the total percentage of XB that is in a force-generating state is calculated as XB A is estimated as follows: XB A =G XB *XB G, (11) In the formula, G XB is the proportion of the XB group involved in force generation estimated in the third function block 206, and XB G is the proportion of XBs that are bound within the group, as estimated in the fifth function block 210.

[0039] The proportion of XBs bonded within a group, XB Gis estimated in a fifth function block 210 as follows:

number

number

number

[0040] In a sixth function block 212, the average XB distortion, s, is estimated as follows:

number

[0041] The sarcomere length, SL, used in the sixth function block 212 is estimated in the seventh function block 214 according to: SL=SL slack *λ, (15)

number

[0042] In an eighth functional block 216, the thin filament length dependence, LDF thin (λ) is estimated as follows: LDF thin (λ)=max{0,min{LDF thinmax , (λ as0 (λ-λ an0 )),(λ as1 (λ-λ an1 ))+LDF thinmax}}. (17) In the above formula, LDF thinmax is the maximum length dependence of the thin filament, λ as0 >λ as1 is the slope of the piecewise linear function, λ an0 <λ an1 is the node that defines the discontinuity.

[0043] In a ninth functional block 218, the length dependence of the thick filament, LDF, is calculated. thick (λ) is estimated as follows: LDF thick (λ)=max{0,min{LDF thickmax , (λ ms0 (λ-λ mn0 )),(λ ms1 (λ-λ mn1 ))+LDF thickmax}}. (18) In the above formula, LDF thickmax is the maximum length dependence of the thick filament, λ ms0 >λ ms1 is the slope of the piecewise linear function, λ mn0 <λ mn1 is the node that defines the discontinuity.

[0044] Next, in a tenth function block 220, the active tension, T a is estimated according to Ta =S a *LDF thick (λ)*XB A *(s+x0). (19) where S a is the scaling factor of the tension generated by the sarcomere, LDF thick (λ) is the thick filament length dependence estimated in the ninth functional block 218, XB A is the total fraction of XB in a force-producing state estimated in the fourth function block 208, s is the average cross-link strain estimated in the sixth function block 212, and x0 is the strain due to the XB power stroke (the step size of the myosin power stroke).

[0045] In an eleventh function block 222, the active tension, T, estimated in the tenth function block 220 is calculated. a is the active force, F, according to A is converted to F A =T a *S T, (20) In the formula, S T is a scaling factor for the force in cell cultures or tissues.

[0046] Thus, the mechanistic model 200 is parameterized according to one or more variable parameters (forming an input parameter vector, such as the input parameter vector 116 shown in FIG. 1 ) and one or more constant parameters. As described in more detail below, the constant parameters of the mechanistic model 200 are known a priori and are therefore fixed, while the distribution of values ​​for each of the variable parameters (i.e., each parameter of the input parameter vector) is determined using empirical data (e.g., the plurality of in vitro responses 106 shown in FIG. 1 ).

[0047] In one embodiment, the variable parameters, i.e., the parameters forming the input parameter vector, are the amplitude of the calcium transient, Ca amp ∈[0.5,2] μM, diastolic calcium level Ca diast∈[0.05,0.4]μM, the transient decay time constant τ2∈[20,140]ms, and the calcium binding rate constant k to the troponin site onT ∈[10,70]μM -1 s -1 , the dissociation rate constant k to the high affinity site offHT ∈[5,70]s -1 , the dissociation rate constant k to the low affinity site offLT ∈[20,1400]s -1 , Hill coefficient n A ∈[3,15]au, half-saturation constant A 50 ∈[0.4,1.2]au, and parameters

number

number

[0048] Referring again to FIG. 1 , a parameter fitting algorithm 112 is used to determine a distribution of values ​​114 for an input parameter vector 116 of the mechanistic model 110 (as described above in connection with FIG. 2 ) based on the plurality of in vitro responses 106. For each parameter in the input parameter vector 116 (e.g., calcium transient amplitude, diastolic calcium level, etc.), the parameter fitting algorithm 112 determines a distribution of values ​​for that parameter. The distribution of values ​​114 may alternatively be referred to as a distribution, parameter distribution, distribution set, or parameter distribution set. The parameter fitting algorithm 112 determines the distribution of values ​​114 such that estimated functional response values ​​obtained from the mechanistic model 110 based on input values ​​obtained from the distribution of values ​​114 approximate measured values ​​from the plurality of in vitro responses 106 (e.g., such that the first set of estimated functional response values ​​118-1 approximates the first set of functional response values ​​106-1). That is, the estimated or predicted distribution of functional responses approximates or matches the empirical distribution determined from the plurality of in vitro responses 106.

[0049] In one embodiment, the parameter fitting algorithm 112 includes a Markov Chain Monte Carlo (MCMC) algorithm, as shown in FIG.

[0050] 3 illustrates an input parameter vector 302, a mechanistic model of cardiac function 304, a plurality of in vitro responses 306, and a distribution of values ​​308 for the input parameter vector 302. In one embodiment, the input parameter vector 302, the mechanistic model 304, the plurality of in vitro responses 306, and the distribution of values ​​308 correspond to the input parameter vector 116, the mechanistic model 110, the plurality of in vitro responses 106, and the distribution of values ​​114 shown in FIG.

[0051] The input parameter vector 302 includes a first parameter 310 and may further include a second parameter 312 and a third parameter 314. The mechanistic model 304 determines a first set of estimated functional response values ​​304-1 associated with a first functional response based on the input parameter vector 302. For example, the mechanistic model 304 may determine an estimated contractile force value based on the input parameter vector 302. In one embodiment, the mechanistic model 304 further determines a second set of estimated functional response values ​​304-2 associated with a second functional response and / or a third set of estimated functional response values ​​304-3 associated with a third functional response. The plurality of in vitro responses 306 includes a first set of measurements 306-1 associated with the first functional response. In one embodiment, the plurality of in vitro responses 306 further includes a second set of measurements 306-2 associated with the second functional response and / or a third set of measurements 306-3 associated with the third functional response.

[0052] The Markov Chain Monte Carlo (MCMC) algorithm is a Bayesian approach to estimating the entire posterior distribution of the parameters of a model or process. At a general level, an MCMC algorithm draws a sample set of input parameters from a prior distribution of the input parameters and evaluates the likelihood of a functional response output from a mechanistic model given the sample set of input parameters. It then uses an accept / reject strategy based on Bayes' rule to decide whether to retain the sample set of parameters and update the prior distribution, or to discard them. The MCMC algorithm then draws a new set of parameters and repeats the process until a termination criterion is met (e.g., the number of retained parameters reaches a threshold). Once the termination criterion is met, the retained parameters approximate the posterior distribution of the parameters of the mechanistic model, and repeated sampling generates sample values ​​for the input parameter vector under the target posterior distribution.

[0053] MCMC algorithms can be either unconstrained or constrained. Unconstrained MCMC algorithms treat the parameters in the input parameter vector θ as random variables rather than unknown constants. Starting with a specific prior density q(θ), the goal is to update the prior density through data X to obtain a posterior density q(θ|X), which is proportional to L(θ,y,μ,σ)q(θ) (Bayes' theorem), where L(θ,y,μ,σ) is the likelihood of the mechanistic model.

number

[0054] In one embodiment, the MCMC algorithm comprises the Metropolis-Hastings algorithm. The posterior density q(θ|X) is the density of the next state θ k+1 is the current state θ k The Markov chain (θ k ) is obtained as the equilibrium distribution density π(θ). In the Metropolis-Hastings algorithm, the Markov chain is implemented using an accept-reject sampling method: given the current state θ of the chain, generate a new candidate state θ' according to the proposal density π(θ'|θ) and accept that candidate with the following probability:

number

number

[0055] After reaching a termination criterion, continued implementation of the MCMC algorithm generates (random) sample values ​​for the input parameter vector θ under the target posterior distribution q(θ|X). The input parameters of the mechanistic model may be obtained from the fitted distribution by taking the posterior mean (also known as Bayesian estimation) or the posterior median, both of which are empirically determined based on the posterior samples. Beneficially, MCMC can be used to obtain accurate point estimates of the input parameter vector and to determine the entire posterior distribution.

[0056] Constrained MCMC algorithms take a similar approach to constrained MCMC algorithms, but split the input parameter vector 302 into shared and variable parameters. The distribution of the shared parameters is shared across treatment groups or conditions, while the variable parameters are separate and can vary for each treatment group or condition. The split of the input parameter vector 302 into shared and variable parameters may be determined by multiple in vitro responses 306 or may be manually identified. In FIG. 3 , a first parameter 310 and a third parameter 314 are shared parameters that are learned across treatment groups or conditions, while a second parameter 312 is a variable parameter that is learned individually for each treatment group or condition. As an example of a constrained MCMC algorithm, consider a situation involving the in vitro response of a cardiac cell culture under a reference condition (i.e., a control) and the in vitro response of a cardiac cell culture under a perturbed condition. When applying a constrained MCMC algorithm to learn the parameter-by-parameter distributions of the input parameter vector, the reference condition response and the perturbation condition response are used together to determine the distributions of the shared parameters (e.g., the first parameter 310 and the third parameter 314), while the distributions of the variable parameters are learned independently (e.g., the reference second parameter 312-1 is learned from the reference condition response, and the perturbation second parameter 312-2 is learned from the perturbation response). Thus, the distributions of the values ​​of the reference condition and the perturbation condition will have the same distributions for the shared parameters and different distributions for the variable parameters.

[0057] Beneficially, a constrained approach to MCMC algorithms allows for the variation of parameters associated with a treatment or condition while increasing the available data on common parameters, thereby improving estimates of the input parameter vector 302 and leading to improvements in downstream drug discovery and development tasks that utilize the input parameter vector 302.

[0058] In one embodiment, the parameter fitting algorithm 112 of FIG. 1 includes a hierarchical Bayesian algorithm as shown in FIG.

[0059] 4 illustrates an input parameter vector 402, a mechanistic model of cardiac function 404, a plurality of in vitro responses 406, and a distribution of values ​​408 for the input parameter vector 402. In one embodiment, the input parameter vector 402, the mechanistic model 404, the plurality of in vitro responses 406, and the distribution of values ​​408 correspond to the input parameter vector 116, the mechanistic model 110, the plurality of in vitro responses 106, and the distribution of values ​​114 shown in FIG. 1. FIG. 4 also illustrates a set of prior distributions 410 associated with the input parameter vector 402.

[0060] The input parameter vector 402 includes a first parameter 412, a second parameter 414, and optionally a third parameter 416. The first parameter 412 and the third parameter 416 are shared or common parameters, while the second parameter 414 is a variable parameter. The set of prior distributions 410 includes a first prior distribution 418 associated with the first parameter 412, a second prior distribution 420 associated with the second parameter 414, and optionally a third prior distribution 422 associated with the third parameter 416.

[0061] Similar to the MCMC approach described in connection with FIG. 3 above, the mechanistic model 404 determines a first set of estimated functional response values ​​404-1 associated with a first functional response based on the input parameter vector 402. In one embodiment, the mechanistic model 404 further determines a second set of estimated functional response values ​​404-2 associated with a second functional response and / or a third set of estimated functional response values ​​404-3 associated with a third functional response. The plurality of in vitro responses 406 includes a first set of measurements 406-1 associated with the first functional response. In one embodiment, the plurality of in vitro responses 406 further includes a second set of measurements 406-2 associated with the second functional response and / or a third set of measurements 406-3 associated with the third functional response.

[0062] In general, a hierarchical Bayesian approach uses different levels to approximate the complete posterior distribution of the parameters in the input parameter vector 402 across multiple groups. Here, the groupings are associated with in vitro data used to fit the parameters (e.g., the plurality of in vitro responses 406) of the mechanistic model 110. For example, the in vitro data may include data acquired from one or more cardiac cell cultures in a control group and data acquired from one or more cardiac cell cultures in a treatment group. Alternatively, the in vitro data may include data collected over different time periods (e.g., over different days). In one embodiment, the groupings are based on one or more features or patterns in the plurality of in vitro responses 406. In the example shown in FIG. 4 , the hierarchical Bayesian approach is applied to two groups such that the distribution of the variable parameter (i.e., the second parameter 414) is learned for each group. Thus, the input parameter vector 402 includes a second parameter 414-1 associated with the first group and a second parameter 414-2 associated with the second group. Both the second parameter associated with the first group 414-1 and the second parameter associated with the second group 414-2 have an associated prior distribution 420-1.

[0063] The hierarchical Bayesian approach assumes a shared population-level prior distribution for both groups and then varies the posterior distribution for each group. Consider, as an example, an input parameter vector θ consisting of a set of shared parameters a and one or more variable parameters b, such as θ = [a, b]. For two groups, the hierarchical Bayesian approach fits an input parameter vector θ = [a, b1, b2], where b1 corresponds to one or more variable parameters of the first group and b2 corresponds to one or more variable parameters of the second group. For the in vitro (empirical) data associated with the first group, the fitted model parameters are [a, b1], while for the in vitro (empirical) data associated with the second group, the model parameters are [a, b2]. The parameters of the input parameter vector θ are fitted using a Markov chain Monte Carlo (MCMC) approach, as described above in connection with Figure 3. However, unlike the approaches described above, a separate Markov chain is associated with each of the parameter groups (i.e., a separate Markov chain can be associated with each parameter or parameter group a, b1, and b2), allowing different posterior distributions to be learned for the variable parameters across both groups while maintaining the same posterior distribution for the common parameters.

[0064] In a hierarchical Bayesian approach, differences in the posterior distributions can be examined after fitting the parameters by adjusting the prior distributions (i.e., adjusting the prior distributions in the set of prior distributions 410). Furthermore, comparing the posterior distributions of the parameters for each of the groups can allow for the identification of significant differences between the groups.

[0065] In one embodiment, the parameter fitting algorithm 112 of FIG. 1 includes a generative adversarial network (GAN) 500, as shown in FIG.

[0066] GAN 500 includes a generator network 502 and a first classifier network 504. Figure 5 further shows an estimated input parameter vector 506, a mechanistic model of cardiac function 508, a plurality of in vitro responses 510, and random variables 512. In one embodiment, GAN 500 further comprises a second classifier network 514 and a simulated input parameter vector 516.

[0067] At a general level, the GAN 500 is trained to transform normally distributed random variables into an input parameter vector such that the estimated or predicted functional response values, determined by the mechanistic model 508 based on the input parameter vector, approximate a plurality of in vitro responses 510.

[0068] In one implementation, the generator network 502 includes eight hidden layers, each containing 160 nodes, and the first classifier network 504 has eight hidden layers, each containing 80 nodes. Both networks use a ReLU activation function, and the first classifier network 504 utilizes dropout with a dropout rate of 0.01. The second classifier network 514 includes an architecture identical to the first classifier network 504.

[0069] In one embodiment, GAN 500 is trained using the ADAM solver with an initial learning rate of 1e-3 and early stopping based on validation loss. A training dataset with 10,000 training examples and 2,000 validation examples is used.

[0070] In one embodiment, the input parameter vector is divided into shared and variable parameters (in a manner similar to that described above in connection with the constrained MCMC approach of FIG. 3 and the hierarchical Bayesian approach of FIG. 4). In such an embodiment, multiple generator networks are used, one generator network for each parameter group. For example, a first generator network is used for the shared parameters, a second generator network is used for the first variable parameters under a first condition, and a third generator network is used for the first variable parameters under a second condition.

[0071] 1 , the output of the non-reversible phase 102 is a distribution 114 of values ​​of the input parameter vector 116 of the mechanistic model 110 as determined by the parameter fitting algorithm 112. The distribution 114 of values ​​may be output to another process or system (e.g., to the reversible phase 104 or to a database 120 for storage therein). Additionally or alternatively, the distribution 114 of values ​​is saved to a persistent storage medium, such as non-volatile memory, a non-transitory medium, or the like. The distribution 114 of values ​​is output or saved in combination with conditions associated with the cardiac cell culture 108.

[0072] According to one aspect of the present invention, a first distribution of values ​​(e.g., distribution of values ​​114) obtained during the non-reversible phase 102 described above is compared to a second distribution of values ​​obtained during the non-reversible phase 102. The first distribution of values ​​is associated with a cardiac cell culture under a first condition, and the second distribution of values ​​is associated with a cardiac cell culture under a second, different condition. In one embodiment, the first condition is a reference or control condition, and the second condition is a perturbed condition associated with a perturbation. As described above, examples of perturbations include drug treatments, disease states or models, different cell lines, etc. Thus, by comparing the first distribution of values ​​with the second distribution of values, a relative distribution shift of the parameters of the mechanistic model 110 from the first condition to the second condition is identified. This relative distribution shift thus provides an indication of a change in the function of the cardiac cell culture due to a change from the first condition to the second condition (e.g., a change due to a perturbation associated with the perturbed condition). The relative distribution shift may be used as a feature or signature of the change from a first condition to a second condition. Because the parameters of the mechanistic model 110 encode the dynamic behavior of an in silico model of cardiac function, the relative shift in the distribution of values ​​for the input parameter vector serves as a surrogate for the change in cardiac function due to the change in conditions.

[0073] In one embodiment, the relative distributions are stored in a database 120 along with condition information. Database 120 is generated using the methodology described above from cardiac cell cultures under a variety of different treatments, disease models, cell lines, and other perturbations. Each entry in the database corresponds to a relative distribution shift associated with a relative shift in cardiac function from a first state to a second state. For example, a shift in cardiac function from a control condition to a condition involving a particular compound, or from a disease condition to a condition involving the administration of a drug for the disease. This database can then be searched to identify similar relative shifts, thereby identifying potential therapeutic candidates (as described in more detail below).

[0074] In some embodiments, one or more outputs may be generated based on a change in function or condition of the cardiac cell culture from a first condition (e.g., a baseline condition) to a second condition (e.g., a perturbed condition). The one or more outputs may be used to identify or be indicative of one or more of: (a) a target molecule as a potential drug candidate, and / or (b) a potential effect of an unknown compound.

[0075] For example, the relative distribution shift also provides a mechanism for in silico prediction of effects (e.g., toxicity, mechanism of action, etc.) associated with unknown compounds or drugs. In one example, a first relative distribution shift is generated from a distribution associated with a cardiac cell culture under control conditions and a distribution associated with a cardiac cell culture under perturbed conditions associated with a dose of the unknown compound. Searching a database of relative distribution shifts (e.g., database 120) using the first relative distribution shift identifies a second relative distribution shift associated with a shift from the control condition to the perturbed condition associated with the first drug. If the first drug has a known mechanism of action and / or toxicity, similar relative distribution shifts indicate that the unknown compound is likely to share the same effect. This is because the relative distribution shift encodes changes in cardiac function due to changes in conditions. Thus, because the unknown compound and the first drug cause the same or similar changes in cardiac function, it can be inferred that the unknown compound and the first drug are likely to share the same or similar effects (e.g., mechanism of action and / or toxicity). The irreversible phase 102 therefore provides a framework for quickly and efficiently identifying the potential effects of unknown compounds or drugs, and can reduce the need to conduct costly and time-consuming in vitro testing to identify such effects.

[0076] In the reversible phase 104, the distribution of values ​​114 is used in conjunction with the mechanistic model 110 to determine an estimated functional response value. Shifts 122 (e.g., relative shifts) made to the distribution of values ​​114 result in a change in the estimated functional response value. By matching the shifts 122 with previously calculated shifts stored in association with specific conditions and effects, the reversible phase 104 enables rapid and effective in silico discovery of potential drugs or treatments that may result in a desired change in functional response.

[0077] FIG. 6 illustrates inverse mechanism modeling, as performed in the reversibility phase 104 of FIG. 1, according to one embodiment of the present disclosure.

[0078] FIG. 6 illustrates a distribution 602 of values ​​of an input parameter vector of a mechanistic model 604 of cardiac function. The distribution 602 of values ​​includes a first distribution 606 for a first parameter of the input parameter vector and may further include a second distribution 608 for a second parameter of the input parameter vector and / or a third distribution 610 for a third parameter of the input parameter vector. Also illustrated in FIG. 6 are a relative shift 612, a shifted first distribution 614, a first set of predicted functional response values ​​616, and a second set of predicted functional response values ​​618. FIG. 6 also illustrates a database 620 and a first medication 622. In one embodiment, the distribution 602 of values, the mechanistic model 604, the first set of predicted functional response values ​​616, and the database 620 correspond to the distribution 114 of values, the mechanistic model 110, the first set of predicted functional response values ​​124-1, and the database 120 shown in FIG. 1 .

[0079] Generally, the reversible phase illustrated in FIG. 6 provides an in silico mechanism for predicting functional changes in cardiac cell cultures. The distribution of values ​​602 for the input parameter vector of the mechanistic model 604 corresponds to the distribution of values ​​learned during the non-reversible phase (as described above in connection with FIGS. 3-5). The distribution of values ​​602 is associated with cardiac cell cultures under associated conditions, such as reference conditions (i.e., the distribution was learned from cardiac cell cultures under control conditions) or perturbed conditions (e.g., the distribution was learned from cardiac cell cultures under a particular disease state). In this manner, the distribution of values ​​602 provides a realistic distribution of the input parameter vector for the in silico model of cardiac cell cultures under associated conditions. A shift applied to the distribution of values ​​602, for example, applying a shift to one or more of the distributions associated with one or more of the parameters of the input parameter vector, can result in a change in the predicted functional response value generated by the mechanistic model 604. 6 , a relative shift 612 applied to a first parameter of an input parameter vector results in a change in the predicted functional response values ​​from a first set of predicted functional response values ​​616 to a second set of predicted functional response values ​​618. The difference between the first set of predicted functional response values ​​616 and the second set of predicted functional response values ​​618 corresponds to a predicted functional change in the cardiac cell culture resulting from the relative shift 612.

[0080] By exploring the space of shifts applied to the distribution of values ​​602, a relative distribution shift that results in a desired or targeted change in the predicted functional response can be identified. That is, one or more of the distributions for the input parameter vector can be iteratively and incrementally adjusted until one or more of the estimated functional response values ​​meets a predetermined criterion. For example, one distribution in the distribution of values ​​602 associated with calcium binding can be shifted until the predicted contractile force value exceeds a predetermined threshold. The relative distribution shift (e.g., relative shift 612) can then be used to search database 620 to identify a matching relative distribution shift in database 620. This matching relative distribution shift has an identical or similar shift to that of the query relative distribution shift. Because the matching relative distribution shift is associated with a change from a first condition to a second condition in the cardiac cell culture (e.g., from a control condition to a treatment condition involving a drug), a first drug 622, or treatment, associated with the second condition is identified as a candidate drug. The candidate agents can be applied to cardiac cell cultures with the goal of producing desired changes in predicted functional responses.

[0081] For example, the distribution of values ​​602 may have been learned from the artificial cardiac tissue under perturbation conditions associated with a disease state of hypertrophic cardiomyopathy. Thus, the first set of predicted functional response values ​​616 would correspond to first functional response values ​​corresponding to the contractile force when the artificial cardiac tissue is in this disease state. In this example, because one of the characteristics of hypertrophic cardiomyopathy is a decrease in myocardial contractile force, it is desirable to increase the contractile force of the artificial cardiac tissue. Thus, one or more shifts are applied to the distribution of values ​​602, or to particular distributions within the distribution of values ​​associated with particular parameters, to identify a shift that results in an increase in contractile force (i.e., the second set of predicted functional response values ​​618 is an improvement over the first set of predicted functional response values ​​616). Once a suitable shift (e.g., relative shift 612) is identified, it may be considered to represent the physiological change in the artificial cardiac tissue necessary to produce a functional change in contractile force (e.g., a decrease in calcium concentration). Because the shift is relative, i.e., it represents the relative distributional change of the input parameter vector from a disease state to a state associated with a change in contractile force, the shift can be used to search for treatments (e.g., compounds or drugs) that may affect such functional changes. The relative shift 612 can be used to search a database 620 to identify similar shifts associated with drugs (or treatments), which may then be identified as candidate drugs for increasing the contractile force of the artificial cardiac tissue.

[0082] Beneficially, the relative nature of distribution shifts in input parameter vectors allows for direct comparison of such shifts, thereby enabling their use in identifying and classifying potential treatments and effects. Moreover, by learning the distribution of values ​​for input parameter vectors from large amounts of high-fidelity data acquired from bioreactors (as described in connection with FIG. 12), the distributions can be shifted in biologically plausible and meaningful ways, generating accurate estimates of the change(s) in the functional response. The reversible phase shown in FIG. 6 thus provides a mechanism for rapidly and efficiently identifying candidate drugs or treatments.

[0083] In some embodiments, further validation of the candidate agent (or treatment) is performed to validate the in silico prediction and determine possible additional effects associated with the candidate agent. For example, if the in silico prediction indicates that the candidate agent results in an increase in contractile force, additional validation would help identify whether the contractile force increased and whether other effects (e.g., an increase / decrease in calcium transients) were also observed. Given a first plurality of in vitro responses including functional response measurements of a first cardiac cell culture under a first set of conditions, a second plurality of in vitro responses is obtained. The second plurality of in vitro responses includes functional response measurements of the second cardiac cell culture under a second set of conditions associated with perturbation of the second cardiac cell culture involving the candidate agent. The second cardiac cell culture can be either the same as the first cardiac cell culture or a different cardiac cell culture. For example, the first and second cardiac cell cultures may be identical if the first cardiac cell culture (used to obtain the in vitro response that resulted in the candidate drug's identified distribution shift) is under a disease condition, and administering the candidate drug to the first cardiac cell culture serves to test the efficacy of the candidate drug. The second plurality of in vitro responses provides an empirical representation of the cardiac cell culture's response to the candidate drug. This response is compared to one or more criteria to determine an effect associated with the candidate drug. For example, if the candidate drug is predicted to increase contractile force by a certain amount, F1, the empirical contractile force of the cardiac cell culture captured in the second plurality of in vitro responses is compared to F1 to determine whether this increase was achieved. Thus, the one or more criteria may include a threshold criterion that is met if the empirical contractile force matches or exceeds the predicted contractile force. In this case, the effect associated with the candidate drug would be that the candidate drug caused the predicted change in contractile force.

[0084] In one embodiment, the relative change in predicted functional response and the empirical functional response are compared to determine efficacy. The relative change in predicted or desired functional response is determined by comparing a first set of predicted functional response values ​​616 with a second set of predicted functional response values ​​618. The relative change in empirical functional response is obtained by comparing in vitro responses obtained under a first condition (e.g., control, disease state, etc.) with a second plurality of in vitro responses obtained under conditions involving the candidate agent.

[0085] In one embodiment, to determine the effect, the relative distribution shift is compared to the relative shift 612. The relative distribution shift is obtained in a manner similar to that described in connection with the irreversible phase 102 above and represents a shift from a first condition (e.g., control, disease state, etc.) to a second condition involving the candidate agent. A comparison of the predicted shift (i.e., relative shift 612) with the empirical shift is then used to validate the predicted shift and determine whether the predicted effect was observed.

[0086] We now describe a method that can be used in combination with the architecture described above to perform inverse modeling of cardiac function.

[0087] FIG. 7 illustrates a method 700 for determining changes in cardiac function according to one embodiment of the present disclosure.

[0088] Method 700 includes steps 702 of acquiring a first plurality of in vitro responses, 704 of acquiring a second plurality of in vitro responses, 706 of acquiring a mechanistic model, 708 of determining a distribution of first values, 710 of determining a distribution of second values, and 712 of determining a first relative distribution shift. Method 700 further includes optional step 714 of outputting the first relative distribution shift.

[0089] In the acquiring step 702, a first plurality of in vitro responses of a first cardiac cell culture (e.g., a plurality of in vitro responses 106 of cardiac cell culture 108 shown in FIG. 1 ) is acquired. The first plurality of in vitro responses includes measurements of a first functional response of the first cardiac cell culture under reference conditions (e.g., a set of first measurements 106-1 shown in FIG. 1 ). Here, functional response is understood to be a biophysical response of the cardiac cell culture, such as the contractile force of the cardiac cell culture, a calcium response or characteristic of the cardiac cell culture, or an electrical response or characteristic of the cardiac cell culture. Thus, the first plurality of in vitro responses includes measurements or values ​​related to one or more of such functional responses of the first cardiac cell culture.

[0090] As described in further detail above, the first cardiac cell culture is either a 2D culture (e.g., a monolayer) or a 3D tissue formed from the 3D culture. The first plurality of in vitro responses corresponds to empirical measurements of the cardiac cell culture under reference condition(s). The reference condition may be a control condition (e.g., the cardiac cell culture is vehicle-treated before the first plurality of in vitro responses is obtained), or the reference condition may be a condition that serves as a reference for the condition(s) associated with the second plurality of in vitro responses (e.g., the reference condition corresponds to the cardiac cell culture being in a disease state in which no drug or treatment has been administered).

[0091] The first functional response measure includes one or more waveform feature values, such as a twitch amplitude value, a contraction time value, a maximum contraction slope value, a relaxation time value, a maximum relaxation slope value, and a twitch duration value. Alternatively, the one or more waveform feature values ​​include spectral feature values. The one or more waveform feature values ​​are determined from a first functional response waveform of the first cardiac cell culture acquired under reference conditions over a first period of time.

[0092] In the acquiring step 704, a second plurality of in vitro responses is acquired. The second plurality of in vitro responses includes measurements of a first functional response of the first cardiac cell culture under perturbation conditions. The perturbation conditions are associated with a first perturbation of the first cardiac cell culture. The first perturbation of the first cardiac cell culture includes one or more of a drug treatment, a disease model, or a different cell line.

[0093] The second plurality of in vitro responses includes measurements or values ​​related to one or more functional responses of the first cardiac cell culture. For example, if the first plurality of in vitro responses includes contractile force-related measurements, the second plurality of in vitro responses also includes contractile force-related measurements. The first functional response measurements include one or more waveform feature values, such as a twitch amplitude value, a contraction time value, a maximum contraction slope value, a relaxation time value, a maximum relaxation slope value, or a twitch duration value. Alternatively, the one or more waveform feature values ​​include spectral feature values. The one or more waveform feature values ​​are determined from a second functional response waveform of the first cardiac cell culture acquired over a second period under perturbed conditions.

[0094] In one embodiment, the first plurality of in vitro responses and the second plurality of in vitro responses are obtained from a bioreactor (e.g., bioreactor 1202 shown in FIG. 12 ) containing the first cardiac cell culture. Alternatively, the first plurality of in vitro responses and the second plurality of in vitro responses are obtained or derived from data obtained from a bioreactor containing the first cardiac cell culture. Advantageously, utilizing a bioreactor such as that described in connection with FIG. 12 to cultivate and maintain cardiac cell cultures can efficiently obtain large amounts of functional data under a variety of conditions and across a variety of cardiac cell cultures. This is particularly beneficial for parameter fitting purposes, where greater amounts of data allow for improved estimation of the distribution of values ​​for the input parameter vector.

[0095] In the obtaining step 706, a mechanistic model of cardiac function (e.g., the mechanistic model 110 shown in FIG. 1) is obtained (e.g., the mechanistic model 110 shown in FIG. 1). The mechanistic model determines one or more estimates of a first functional response based on an input parameter vector (e.g., a first estimated functional response value 118-1 determined from the mechanistic model 110 based on the input parameter vector 116 shown in FIG. 1). In one embodiment, the mechanistic model also determines one or more estimates of a second functional response and / or one or more estimates of a third functional response based on the input parameter vector.

[0096] A mechanistic model is an in silico model of cardiac function, where an input parameter vector, together with the initial conditions of the model, determines the dynamic behavior of the mechanistic model (i.e., the dynamic behavior of the in silico cardiac model). The input parameter vector includes multiple input parameters, each of which takes an input value. For example, the input parameter vector may include a parameter associated with calcium binding, a parameter associated with the gating properties of cardiac ion channels, etc. The mechanistic model is a mathematical model parameterized by the input parameter vector, and in one embodiment, the mechanistic model includes one or more ordinary differential equations (ODEs). Further details are provided in connection with FIG. 2 above.

[0097] The distributions of the input parameters are unknown a priori and are learned from empirical (in vitro) functional response data. Because the functional response data (i.e., multiple in vitro responses) capture different conditions in cardiac cell cultures, different distributions for the input parameter vector are learned for each different condition.

[0098] In a determining step 708, a first value distribution (e.g., value distribution 114 shown in FIG. 1 ) for the input parameter vector is determined such that an estimate of a first functional response obtained from the mechanistic model based on input values ​​obtained from the first value distribution approximates a measured value from the first plurality of in vitro responses.

[0099] In a determining step 710, a distribution of second values ​​for the input parameter vector is determined such that an estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of second values ​​approximates a measured value from the second plurality of in vitro responses.

[0100] The first and second value distributions include a distribution for each of the parameters of the input parameter vector. For example, if the input parameter vector includes three parameters, then both the first and second value distributions will include three distributions, one associated with a respective parameter of the input parameter vector. The first and second value distributions are determined using a learning algorithm. As described in detail above in connection with Figures 3-5, the learning algorithm may be one of a constrained or unconstrained Markov Chain Monte Carlo (MCMC) algorithm, a hierarchical Bayesian algorithm, or a generative adversarial network (GAN).

[0101] In a determining step 712, a first relative distribution shift is determined based on a comparison of the distribution of the first values ​​to the distribution of the second values, the first relative distribution shift being indicative of a change in function of the first cardiac cell culture due to the perturbation conditions.

[0102] The first relative distribution shift is used as a feature or signature of the change from baseline conditions to perturbed conditions. Because the parameters of the mechanistic model encode the dynamic behavior of the in silico model of cardiac function, the first relative shift in the distribution of values ​​for the input parameter vector serves as an efficient surrogate for the change in cardiac function due to the changed conditions.

[0103] In some embodiments, one or more outputs may be generated based on a change in function or condition of the cardiac cell culture from baseline conditions to perturbed conditions, and the one or more outputs may be used to identify or be indicative of one or more of: (a) target molecules as potential drug candidates, and / or (b) potential effects of unknown compounds.

[0104] In an optional outputting step 714, the first relative distribution shift is output. In one embodiment, outputting the first relative distribution shift includes storing or saving the first relative distribution shift along with associated condition information in a database or other structured or unstructured repository (e.g., database 120 shown in FIG. 1 ). This allows for databases to be built from a wide variety of compounds and conditions. Alternatively, outputting the first relative distribution shift includes saving the first relative distribution shift to a persistent storage medium, such as non-volatile memory, a non-transitory medium, or the like. Additionally or alternatively, outputting the first relative distribution shift includes transmitting or displaying the first relative distribution shift for review by a user.

[0105] FIG. 8 illustrates a first series of steps 800 performed in conjunction with the method 700 of FIG. 7 according to an embodiment of the present disclosure.

[0106] In one embodiment, the first set of steps 800 is performed after the steps of method 700 have been performed. In particular, the first set of steps 800 is performed after the determining step 712 of method 700 has been performed. The first set of steps 800 includes comparing 802 the first relative distribution shift to a plurality of distribution shifts and determining 804 a first effect. The first set of steps 800 further includes the optional step 806 of outputting the first effect.

[0107] In general, the first series of steps 800 shown in Figure 8 provides a method for identifying the potential effect of a compound or treatment whose effect is unknown. For example, the first series of steps 800 can be used to identify the potential mechanism of action of cardiac cell cultures treated with a drug whose mechanism of action is unknown.

[0108] In a comparing step 802, the first relative distribution shift is compared to a plurality of relative distribution shifts associated with a plurality of perturbations, each of which has a known effect. For example, the first relative distribution shift is compared to distribution shifts stored in a database, such as database 120 shown in FIG. 1. Because the distribution shifts are relative, direct comparisons between various shifts are possible.

[0109] In determining step 804, a first effect associated with the first perturbation of the first cardiac cell culture is determined based on comparing 802. For example, a most similar relative distribution shift among the plurality of relative distribution shifts is identified, and an effect associated with this relative distribution shift is determined. Here, the most similar relative distribution shift may be a relative distribution shift having the highest similarity score to the first relative distribution shift, or a relative distribution shift having a similarity score to the first relative distribution shift above a predetermined threshold. The first effect may include one or more of a mechanism of action, toxicity, etc.

[0110] In an optional step of outputting 806, the first effect is output. Outputting the first effect may include storing the first effect in a persistent storage medium, such as non-volatile memory, a non-transitory medium, etc. Additionally or alternatively, outputting the first effect may include transmitting or displaying the first effect for review by a user.

[0111] FIG. 9 illustrates a second series of steps 900 performed in conjunction with the method 700 of FIG. 7 according to an embodiment of the present disclosure.

[0112] In one embodiment, the second set of steps 900 is performed after steps of method 700 have been performed. In particular, the second set of steps 900 is performed after determining step 712 of method 700 has been performed. The second set of steps 900 includes obtaining 902 a second relative distribution shift, comparing 904 the first relative distribution shift to the second relative distribution shift, and validating 906 the distribution of the first values ​​or the distribution of the second values.

[0113] Generally, a second series of steps 900 shown in FIG. 9 provides a method for verifying a distribution shift.

[0114] In obtaining step 902, a second relative distribution shift is obtained. The second relative distribution shift is associated with a second cardiac cell culture and a perturbation condition. The second relative distribution shift is obtained in the same manner as described in connection with the first relative distribution shift in FIG. 7 , i.e., the in vitro responses of the cardiac cell culture under both reference and perturbation conditions are obtained. A distribution of values ​​for the input parameter vector of the mechanistic model is determined from each set of in vitro responses, and a second relative distribution shift is determined from the two distributions of values.

[0115] The first relative distribution shift is compared to the second relative distribution shift in a comparing step 904. Because both distribution shifts are relative, the first relative distribution shift can be directly compared to the second relative distribution shift.

[0116] In a validating step 906, the first distribution set or the second distribution set is validated based on the comparison 904. For example, if the comparison results in two relative distribution shifts that are identical or similar, this indicates that the first and second relative distribution shifts accurately represent changes in cardiac function in silico between baseline and perturbed conditions.

[0117] FIG. 10 illustrates a method 1000 for inverse mechanism modeling of cardiac function according to one aspect of the present disclosure.

[0118] Method 1000 includes step 1002 of acquiring a first plurality of in vitro responses, step 1004 of acquiring a mechanistic model, step 1006 of determining a distribution of first values, step 1008 of determining a distribution of shifted values, and step 1010 of determining a first relative distribution shift. Method 1000 further includes optional step 1012 of outputting the first relative distribution shift.

[0119] In the acquiring step 1002, a first plurality of in vitro responses is acquired. The first plurality of in vitro responses includes measurements of at least one functional response of the first cardiac cell culture under a first set of conditions. Examples of functional responses include contractile force, calcium transients, and electrical responses.

[0120] The first plurality of in vitro responses corresponds to empirical measurements of a first cardiac cell culture (e.g., a 2D culture or a 3D tissue formed from the 3D culture) under a first set of conditions (e.g., reference conditions). The at least one functional response measurement includes one or more waveform feature values, such as a twitch amplitude value, a contraction time value, a maximum contraction slope value, a relaxation time value, a maximum relaxation slope value, and a twitch duration value. Alternatively, the one or more waveform feature values ​​include spectral feature values. The one or more waveform feature values ​​are determined from a first functional response waveform of the first cardiac cell culture acquired over a first period of time under the first set of conditions.

[0121] In one embodiment, the first plurality of in vitro responses are obtained from a bioreactor containing the first cardiac cell culture (e.g., bioreactor 1202 described in connection with FIG. 12 ). Alternatively, the first plurality of in vitro responses are obtained or derived from data obtained from a bioreactor containing the first cardiac cell culture.

[0122] A mechanistic model of cardiac function is obtained in an obtaining step 1004. The mechanistic model predicts one or more values ​​of at least one functional response based on an input parameter vector.

[0123] A mechanistic model is an in silico model of cardiac function, where an input parameter vector, together with the initial conditions of the model, determines the dynamic behavior of the mechanistic model (i.e., the dynamic behavior of the in silico cardiac model). The input parameter vector includes multiple input parameters, each of which takes an input value. A mechanistic model is a mathematical model parameterized by the input parameter vector, and in one embodiment, the mechanistic model includes one or more ordinary differential equations (ODEs). Further details are provided in the description associated with FIG. 2 above.

[0124] In a determining step 1006, a distribution of first values ​​for the input parameter vector is determined such that an estimate of at least one functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of first values ​​approximates a measured value from the first plurality of in vitro responses, i.e., the predicted distribution of functional response values ​​approximates or matches the distribution of empirical functional response values ​​from the first plurality of in vitro responses.

[0125] The distribution of first values ​​is determined using a learning algorithm, which, as detailed above in connection with Figures 3-5, is one of a Markov Chain Monte Carlo (MCMC) algorithm, a hierarchical Bayesian algorithm, or a generative adversarial network (GAN), which may be constrained or unconstrained.

[0126] In determining step 1008, a shifted first value distribution for the input parameter vector is determined such that an estimated value of at least one functional response obtained from the mechanistic model based on input values ​​obtained from the shifted first value distribution satisfies a predetermined functional response criterion. For example, the value distribution is shifted until the at least one functional response obtained from the mechanistic model satisfies a threshold (i.e., is greater than or less than the threshold). As a further example, the value distribution is shifted until the at least one functional response obtained from the mechanistic model falls within a predetermined value range. Here, shifting the first value distribution may include applying a shift operation to a distribution associated with one parameter of the input parameter vector (e.g., shifting a distribution associated with a calcium binding parameter). Alternatively, shifting the value distribution may include applying multiple shift operations to multiple distributions associated with multiple parameters of the input parameter vector. In one embodiment, the first relative distribution shift is determined based on a shift of one or more prior distributions.

[0127] A first relative distribution shift is determined based on a comparison of the distribution of first values ​​and the shifted distribution of first values ​​in determining step 1010. The first relative distribution shift is indicative of a predicted change in cardiac function from the first set of conditions to the set of conditions associated with a predetermined functional response criterion.

[0128] Advantageously, method 1000 provides a fast and efficient mechanism for searching through the space of distribution shifts of input parameter vectors to identify shifts that result in desired or targeted changes to the functional response(s) of cardiac cell cultures, which may help reduce the cost and time associated with identifying targeted therapies or agents during drug discovery and development.

[0129] In an optional outputting step 1012, the first relative distribution shift is output. Outputting the first relative distribution shift may include storing the first relative distribution shift in a persistent storage medium, such as a non-volatile memory, a non-transitory medium, etc. Additionally or alternatively, outputting the first relative distribution shift may include transmitting or displaying the first relative distribution shift for review by a user.

[0130] FIG. 11 illustrates a first series of steps 1100 that are performed in some embodiments in conjunction with the method 1000 of FIG.

[0131] In an embodiment, the first series of steps 1100 is performed after the steps of method 1000 have been performed. In particular, the first series of steps 1100 is performed after the determining step 1010 of method 1000 has been performed. The first series of steps 1100 includes comparing 1102 the first relative distribution shift to a plurality of distribution shifts, identifying 1104 a first agent, obtaining 1106 a second plurality of in vitro responses, comparing 1108 the second plurality of in vitro responses to predetermined functional response criteria, and identifying 1110 an effect. The first series of steps 1100 further includes the optional step 1112 of outputting the effect.

[0132] In a comparing step 1102, the first relative distribution shift is compared to a plurality of relative distribution shifts. The plurality of relative distribution shifts are associated with a corresponding plurality of drugs. In one embodiment, the plurality of relative distribution shifts are stored in a persistent repository or database (e.g., database 120 shown in FIG. 1). The plurality of relative distribution shifts are determined from cardiac cell cultures across a range of conditions and treatments using the non-reversible approach described in connection with FIG. 7.

[0133] In identifying step 1104, a first drug is identified from the corresponding plurality of drugs based on comparison 1102. For example, the most similar relative distribution shift among the plurality of relative distribution shifts is identified, and the drug associated with this relative distribution shift (i.e., the drug associated with the perturbation condition from which the relative distribution shift was generated) is determined. Here, the most similar relative distribution shift may be the relative distribution shift with the highest similarity score to the first relative distribution shift, or the relative distribution shift with a similarity score to the first relative distribution shift that exceeds a predetermined threshold.

[0134] In the acquiring step 1106, a second plurality of in vitro responses is acquired. The second plurality of in vitro responses includes measurements of at least one functional response of the second cardiac cell culture under a second set of conditions. The second set of conditions is associated with a perturbation of the second cardiac cell culture involving a first agent. In one embodiment, the first cardiac cell culture and the second cardiac cell culture are the same.

[0135] The second plurality of in vitro responses corresponds to empirical measurements of the second cardiac cell culture under a second set of conditions (e.g., perturbed conditions). The at least one functional response measurement includes one or more waveform feature values, such as a twitch amplitude value, a contraction time value, a maximum contraction slope value, a relaxation time value, a maximum relaxation slope value, and a twitch duration value. Alternatively, the one or more waveform feature values ​​include spectral feature values. The one or more waveform feature values ​​are determined from a second functional response waveform of the second cardiac cell culture acquired over a second period of time under the second set of conditions.

[0136] The second plurality of in vitro responses is obtained from a bioreactor containing the second cardiac cell culture (e.g., bioreactor 1202 described in connection with FIG. 12 ). Alternatively, the second plurality of in vitro responses is obtained or derived from data obtained from a bioreactor containing the second cardiac cell culture.

[0137] The second plurality of in vitro responses is compared to a predetermined functional response criterion in a comparing step 1108. In one embodiment, the at least one functional response includes a first functional response and a second functional response, and the predetermined functional response criterion is associated with the second functional response.

[0138] In identifying step 1110, an effect associated with the first agent is identified based on comparing 1108. The effect is associated with a change in functional response (e.g., a change in the first functional response, a change in the second functional response, etc.), particularly with a change in the predicted functional response determined as a result of the shift performed in determining step 1008 of FIG. 10 and a change in the empirical functional response within the second plurality of in vitro responses.

[0139] In an optional outputting step 1112, the effect is output. In one embodiment, outputting the effect includes storing or saving the effect in a database or other structured or unstructured storage location. Alternatively, outputting the effect includes saving the effect to a persistent storage medium, such as non-volatile memory, a non-transitory medium, etc. Additionally or alternatively, outputting the effect includes transmitting or displaying the effect for review by a user.

[0140] FIG. 12 illustrates a system 1200 for acquiring functional response data of cardiac cell cultures according to one embodiment of the present disclosure.

[0141] System 1200 includes a bioreactor 1202 and a control unit 1204. Bioreactor 1202 includes an apparatus 1206 for growing cardiac cell cultures, a sensor assembly 1208, and an interface 1210. Control unit 1204 includes a feature extractor 1212. Interface 1210 communicatively couples bioreactor 1202 and control unit 1204 such that data can be exchanged between bioreactor 1202 and control unit 1204. In one embodiment, control unit 1204 is configured to implement the above-described methods illustrated in FIGS.

[0142] As shown in enlarged portion 1206-1 of device 1206, device 1206, i.e., substrate, includes one or more wells, such as well 1214, one or more cell culture wells, such as cell culture well 1216, a pair of electrodes including first electrode 1218-1 and second electrode 1218-2, and a pair of elements including first element 1220-1 and second element 1220-2. Well 1214 is disposed within cell culture well 1216 and has a bottom, a first end 1222-1, and a second end 1222-2 on device 1206. Well 1214 is configured to grow cardiac cell culture 1224 from cells seeded therein. Culture medium may be added to cell culture well 1216 to grow and / or maintain cardiac cell culture 1224. In one embodiment, the cardiac cell culture is a 2D culture (e.g., a monolayer). In an alternative embodiment, the cardiac cell culture is a 3D tissue formed from a 3D cell culture (ie, an artificial tissue).

[0143] The pair of electrodes is separated by a gap within which the well 1214 is disposed. The pair of electrodes is configured to apply electrical stimulation to a cell culture in one or more wells of the device 1206 (e.g., cardiac cell culture 1124 in well 1214, shown in enlarged portion 1206-1). While the cell culture matures within the device 1206, the pair of electrodes stimulates the cell culture according to a multi-week electrical stimulation protocol. After the cell culture matures, the pair of electrodes may be configured to stimulate the cell culture (e.g., cardiac cell culture 1224) at a set frequency, i.e., a pacing frequency. In one embodiment, the frequency at which the cell culture is stimulated, i.e., the pacing frequency, is set by the control unit 1204. In this manner, the control unit 1204 may be configured to send instructions 1226 to the bioreactor 1202 to cause the bioreactor 1202 to stimulate the cardiac cell culture(s) within the device 1206 at the set pacing frequency.

[0144] First element 1220-1 and second element 1220-2 are positioned across well 1214 such that a gap is provided between the bottom of well 1214 and the pair of elements. First element 1220-1 and second element 1220-2 are configured to (a) allow attachment of cardiac cell culture 1224 formed therebetween, thereby suspending cardiac cell culture 1224 above the bottom of well 1214, and (b) deform in response to contractile forces exerted by cardiac cell culture 1224 on the pair of elements, thereby simulating the specific physiological environment of cardiac cell culture 1224 and / or allowing measurement of contractile forces (e.g., by sensor assembly 1208). For example, the pair of electrodes may apply electrical stimulation to cardiac cell culture 1224 at a frequency of 0.1 Hz. The cardiac cell culture 1224 will contract in response to this electrical stimulation, causing deformation of at least one of the first element 1220-1 and the second element 1220-2. By measuring the deformation of the pair of elements, the functional response of the cardiac cell culture 1224 when stimulated at 0.1 Hz can be recorded.

[0145] Sensor assembly 1208 is configured to detect one or more functional responses of cardiac cell cultures in device 1206 (e.g., one or more functional responses of cardiac cell culture 1224). In one embodiment, sensor assembly 1208 includes an optical sensor. The optical sensor is configured to detect deformation of first element 1220-1 and / or second element 1220-2 (e.g., resulting from contractile forces exerted on first element 1220-1 and / or second element 1220-2 by cardiac cell culture 1224). By detecting the deformation of first element 1220-1 and / or second element 1220-2, one or more functional responses, such as a displacement, i.e., contractile displacement, of cardiac cell culture 1224, or a contractile force of cardiac cell culture 1224, can be determined. Additionally or alternatively, the optical sensor of sensor assembly 1208 is configured to detect fluorescence intensity of cardiac cell culture 1224. By detecting the fluorescence intensity of the cardiac cell culture 1224, one or more functional responses can be determined, such as a transient calcium response of the cardiac cell culture 1224 or a change in membrane potential of the cardiac cell culture 1224. Additionally or alternatively, the optical sensor of the sensor assembly 1208 is configured to detect a change in a dimension of the cardiac cell culture 1224 over a time frame. By detecting a change in a dimension of the cardiac cell culture 1224 over a time frame, one or more functional responses can be determined, such as a displacement, i.e., contractile displacement, of the cardiac cell culture 1224, or a contractile force of the cardiac cell culture 1224.

[0146] The optical sensor of the sensor assembly 1208 is configured to acquire multiple image-based representations of one or more functional responses of the tissue (e.g., cardiac cell culture 1224) over a time frame or period. For example, the optical sensor may be configured to capture an image, or frame, of the tissue every n seconds, where n is related to a predetermined rate at which images or frames are captured. For example, if n=1, one image of the tissue is captured per second. The value of n may be selected to be any suitable value, such as n={1 / 60, 1 / 50, 1 / 30, 1 / 24, 1 / 12, 1 / 4, 1 / 2, 1, 2}. In one embodiment, the frame rate is determined according to the frequency at which the tissue within the device is stimulated. Thus, a series of images or frames of the tissue over a time frame will capture one or more functional responses of the tissue over that time frame.

[0147] In one embodiment, bioreactor 1202 is configured to convert a series of images capturing one or more functional responses of tissue over a time frame into a waveform representation. For example, sensor assembly 1208, interface 1210, or another component of bioreactor 1202 may process images in the series to extract features related to the functional response of the tissue at the time associated with the images. The features related to the functional response of the tissue extracted from the series of images may then be combined to form a waveform containing one or more functional responses of the tissue over time. A waveform, such as waveform 1228, may then be output from bioreactor 1202. In another embodiment, raw image or frame data is output from bioreactor 1202, and a separate unit, e.g., control unit 1204 or an image analysis unit (not shown), processes this data to determine time-series functional response data.

[0148] The time-series functional response data (e.g., waveform 1228) generated by bioreactor 1202 provides a high-fidelity, information-dense representation of the functional response of the artificial tissue over a predetermined period of time (e.g., contractile force generated by a first tissue in response to a pacing frequency of 1 Hz for 30 seconds). To obtain such waveform data from bioreactor 1202 for cardiac cell culture 1224, a command can be sent to bioreactor 1202 (e.g., from control unit 1204) to begin stimulating cardiac cell culture 1224 at a pacing frequency. Alternatively, to observe the spontaneous response of cardiac cell culture 1224, no electrical stimulation is applied. Sensor assembly 1208 then captures a series of images of cardiac cell culture 1224 over a predetermined period of time. Crucially, the series of images captures the response (e.g., deformation) of first element 1220-1 and / or second element 1220-2 as a result of the contractile response of cardiac cell culture 1224. The series of images is then processed to extract the response of the first element 1220-1 and / or the second element 1220-2 over a predetermined period of time and convert the element's response over time into a time series of functional responses over time. For example, the displacement of the first element 1220-1 and / or the second element 1220-2 may be used to determine the force of the contractile response of the cardiac cell culture 1124. The time series (e.g., waveform 1228) is then output from the bioreactor 1202 for further processing and / or analysis. A feature extractor 1212 in the control unit 1204 extracts one or more waveform features (i.e., functional response values ​​or features) from the waveform 1228 and outputs a plurality of measurements 1230. The features extracted by the feature extractor 1212 include, for each peak in the waveform 1228, the twitch amplitude, contraction time, maximum contraction slope, relaxation time, maximum relaxation slope, and twitch duration. The feature extractor 1212 is also configured to extract spectral features from the waveform 1228 using a spectral transform algorithm, such as a Fourier transform.

[0149] The tissue (e.g., cardiac cell culture 1224) within bioreactor 1202 may be periodically administered a drug or compound, and the functional response of the tissue after administration may be recorded in the form of a waveform. Alternatively, the functional response of the tissue in the absence of an external administration regimen may be recorded periodically. In either case, changes in the functional response of the tissue are captured in subsequent waveforms, as shown in FIG.

[0150] FIG. 13 shows a waveform 1302 containing the functional response of tissue over a period of time.

[0151] Waveform 1302 is acquired from a bioreactor, such as bioreactor 1202 shown in FIG. 12, as described above. Waveform 1302 provides a high-fidelity, information-dense representation of the functional response of tissue over a time frame from t1 to t2. Waveform 1302 includes multiple peaks corresponding to the contractile response of tissue over a time frame (e.g., 30 seconds, 40 seconds, etc.). A magnified view of the peaks within portion 1304 of waveform 1302 is shown in magnified portion 1304-1.

[0152] Typically, multiple features are extracted from each peak of waveform 1302 (e.g., using feature extractor 1212 shown in FIG. 12 ) to characterize waveform 1302 and thereby enable further processing or analysis of waveform 1302. As shown in zoomed-in portion 1304-1, features extracted from a single peak include peak (or twitch) amplitude 1306, time to peak amplitude 1308 (or contraction time), time to peak decline 1310 (or relaxation time), duration 1312 (or twitch duration), maximum rate of development 1314 (or maximum contraction slope), maximum rate of decline 1316 (or maximum relaxation slope), and passive tension 1318.

[0153] Figure 14 illustrates an exemplary computing system for performing the methods of the present disclosure. Specifically, Figure 14 illustrates a block diagram of one embodiment of a computing system according to an exemplary embodiment of the present disclosure.

[0154] The computing system 1400 can be configured to perform any of the operations disclosed herein, for example, any of the operations discussed with reference to the functional modules described in connection with FIG. 1A . The computing system includes one or more computing device(s) 1402. The computing device(s) 1402 of the computing system 1400 include one or more processors 1404 and memory 1406. The one or more processors 1404 can be any general-purpose processor(s) configured to execute a set of instructions. For example, the one or more processors 1404 can be one or more general-purpose processors, one or more field programmable gate arrays (FPGAs), and / or one or more application-specific integrated circuits (ASICs). In one embodiment, the one or more processors 1404 include a single processor. Alternatively, the one or more processors 1404 include multiple processors operatively connected. The one or more processors 1404 are communicatively coupled to the memory 1406 via an address bus 1408, a control bus 1410, and a data bus 1412. The memory 1406 can be random access memory (RAM), read only memory (ROM), a persistent storage medium device such as a hard drive, and / or erasable programmable read only memory (EPROM), etc. The computing device(s) 1402 further includes an input / output (I / O) interface 1414 communicatively coupled to the address bus 1408, the control bus 1410, and the data bus 1412.

[0155] Memory 1406 may store information accessible to one or more processors 1404. For example, memory 1406 (e.g., one or more non-transitory computer-readable storage media, memory devices) may include computer-readable instructions (not shown) that may be executed by one or more processors 1404. The computer-readable instructions may be software written in any suitable programming language or may be implemented in hardware. Additionally or alternatively, the computer-readable instructions may execute in logically and / or virtually separate threads on one or more processors 1404. For example, memory 1406 may store instructions (not shown) that, when executed by one or more processors 1404, cause the one or more processors 1404 to perform operations, such as any of the operations and functions that computing system 1400 is configured to perform, as described herein. Additionally or alternatively, memory 1406 may store data (not shown) that may be retrieved, received, accessed, written, manipulated, created, and / or stored. The data may include, for example, the data and / or information described herein in connection with Figures 1-13. In some implementations, the computing device(s) 1402 may retrieve data from and / or store data in one or more memory device(s) located remotely from the computing system 1400.

[0156] The computing environment 1400 further comprises a storage unit 1416, a network interface 1418, an input controller 1420, and an output controller 1422. The storage unit 1416, the network interface 1418, the input controller 1420, and the output controller 1422 are communicatively coupled to the computing device(s) 1402 via the I / O interface 1414.

[0157] The storage unit 1416 is a computer-readable medium, preferably a non-transitory computer-readable medium, that contains one or more programs that, when executed by the one or more processors 1404, contain instructions that cause the computing environment 1400 to perform the method steps of the present disclosure. Alternatively, the storage unit 1416 is a transient computer-readable medium. The storage unit 1416 can be a persistent storage device, such as a hard drive, cloud storage, or any other suitable storage device.

[0158] The network interface 1418 may be a Wi-Fi module, a network interface card, a Bluetooth module, and / or any other suitable wired or wireless communication device. In one embodiment, the network interface 1418 is configured to connect to a network, such as a local area network (LAN) or a wide area network (WAN), the Internet, or an intranet.

Claims

1. 1. A method for predicting changes in cardiac function, comprising: acquiring, by one or more processors, a first plurality of in vitro responses comprising measurements of a first functional response of the first cardiac cell culture under reference conditions; acquiring, by the one or more processors, a second plurality of in vitro responses comprising measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with a first perturbation of the first cardiac cell culture; obtaining, by the one or more processors, a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector; determining, by the one or more processors, a distribution of first values ​​for the input parameter vector, such that an estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of first values ​​approximates a measurement from the first plurality of in vitro responses; determining, by the one or more processors, a second distribution of values ​​for the input parameter vector, such that an estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates a measurement from the second plurality of in vitro responses; determining, by the one or more processors, a first relative distribution shift based on a comparison of the distribution of the first values ​​and the distribution of the second values, wherein the first relative distribution shift is indicative of a change in function of the first cardiac cell culture due to the perturbation condition.

2. The method of claim 1 , further comprising outputting, by the one or more processors, the first relative population shift.

3. 10. The method of claim 1, wherein the first perturbation of the first cardiac cell culture comprises one or more of a drug treatment, a disease model, or a different cell line.

4. comparing, by the one or more processors, the first relative population shift to a plurality of relative population shifts associated with a plurality of perturbations, each of the plurality of perturbations having a known effect; 10. The method of claim 1, further comprising determining, by the one or more processors, a first effect associated with the first perturbation of the first cardiac cell culture based on the comparison.

5. The method of claim 4 , further comprising outputting, by the one or more processors, the first effect.

6. The method of claim 4 , wherein the first effect comprises a mechanism of action.

7. The method of claim 4 , wherein the first effect comprises toxicity.

8. obtaining, by the one or more processors, a second relative distribution shift associated with a second cardiac cell culture and the perturbation condition; comparing, by the one or more processors, the first relative distribution shift and the second relative distribution shift; The method of claim 1 , further comprising: validating, by the one or more processors, the first distribution set or the second distribution set based on the comparison.

9. The method of claim 1 , wherein the first functional response measure comprises one or more waveform feature values.

10. 10. The method of claim 9, wherein the one or more waveform feature values ​​comprise one or more of a twitch amplitude value, a contraction time value, a maximum contraction slope value, a relaxation time value, a maximum relaxation slope value, and a twitch duration value.

11. The method of claim 9 , wherein the one or more waveform feature values ​​include spectral feature values.

12. 10. The method of claim 9, wherein the one or more waveform feature values ​​for the first plurality of in vitro responses are determined from a first functional response waveform of the first cardiac cell culture under reference conditions acquired over a first period of time.

13. 10. The method of claim 9, wherein the one or more waveform feature values ​​for the second plurality of in vitro responses are determined from a second functional response waveform of the first cardiac cell culture under the perturbed conditions over a second period of time.

14. 10. The method of claim 1, wherein the first plurality of in vitro responses and the second plurality of in vitro responses are obtained from a bioreactor containing the first cardiac cell culture.

15. The method of claim 1 , wherein the first cardiac cell culture comprises a two-dimensional cardiac cell culture.

16. The method of claim 1 , wherein the first cardiac cell culture comprises three-dimensional cardiac tissue.

17. The method of claim 1 , wherein the mechanistic model comprises a mathematical model.

18. The method of claim 17 , wherein the mathematical model comprises one or more ordinary differential equations.

19. The method of claim 1 , wherein the input parameter vector includes a plurality of input parameters, each of which takes on an input value.

20. 20. The method of claim 19, wherein the distribution of first values ​​comprises a distribution for each of the plurality of input parameters.

21. 20. The method of claim 19, wherein the distribution of second values ​​comprises a distribution for each of the plurality of input parameters.

22. The method of claim 1 , wherein the distribution of the first values ​​for the input parameter vector is determined using a first learning algorithm.

23. 23. The method of claim 22, wherein the distribution of the second values ​​for the input parameter vector is determined using the first learning algorithm.

24. 23. The method of claim 22, wherein the first learning algorithm comprises a Markov Chain Monte Carlo (MCMC) algorithm.

25. 25. The method of claim 24, wherein the MCMC is constrained according to one or more groupings of parameters in the input parameter vector.

26. The method of claim 22 , wherein the first learning algorithm comprises a generative adversarial network.

27. 23. The method of claim 22, wherein the first learning algorithm comprises a hierarchical Bayesian algorithm or model.

28. 28. The method of claim 27, wherein the hierarchical Bayesian model comprises a plurality of prior distributions associated with one or more groupings of parameters in the input parameter vector.

29. 10. The method of claim 1, wherein the first plurality of in vitro responses and the second plurality of in vitro responses further comprise a measure of a second functional response of the first cardiac cell culture.

30. 30. The method of claim 29, wherein the mechanistic model determines one or more estimates of the second functional response based on the input parameter vector.

31. The method of claim 1 , wherein the first functional response comprises contractile force of the first cardiac cell culture.

32. 2. The method of claim 1, wherein the first functional response comprises a calcium transient response of the first cardiac cell culture.

33. The method of claim 1 , wherein the first functional response comprises an electrical response of the first cardiac cell culture.

34. A non-transitory machine-readable medium storing instructions that, when executed by one or more processors of a device, cause the one or more processors of the device to: obtaining a first plurality of in vitro responses comprising measurements of a first functional response of the first cardiac cell culture under reference conditions; acquiring a second plurality of in vitro responses comprising measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with a first perturbation of the first cardiac cell culture; obtaining a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector; determining a first distribution of values ​​for the input parameter vector, wherein the first functional response estimate obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining a second distribution of values ​​for the input parameter vector, wherein the determination is performed such that the estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates measurements from the second plurality of in vitro responses; determining a first relative distribution shift based on a comparison of the distribution of first values ​​to the distribution of second values, wherein the first relative distribution shift is indicative of a change in function of the first cardiac cell culture due to the perturbation condition.

35. one or more processors; a memory for storing instructions, When the instructions are executed by the one or more processors, the instructions cause the one or more processors to: obtaining a first plurality of in vitro responses comprising measurements of a first functional response of the first cardiac cell culture under reference conditions; acquiring a second plurality of in vitro responses comprising measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with a first perturbation of the first cardiac cell culture; obtaining a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector; determining a first distribution of values ​​for the input parameter vector, wherein the first functional response estimate obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining a second distribution of values ​​for the input parameter vector, wherein the determination is performed such that the estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates measurements from the second plurality of in vitro responses; determining a first relative distribution shift based on a comparison of the distribution of the first values ​​to the distribution of the second values, wherein the first relative distribution shift is indicative of a change in function of the first cardiac cell culture due to the perturbation condition.

36. 1. A method for inverse mechanism modeling of cardiac function, comprising: acquiring, by one or more processors, a first plurality of in vitro responses comprising measurements of at least one functional response of the first cardiac cell culture under a first set of conditions; obtaining, by the one or more processors, a mechanistic model of cardiac function, the mechanistic model predicting values ​​of one or more of the at least one functional response based on an input parameter vector; determining, by the one or more processors, a distribution of first values ​​for the input parameter vector, such that an estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining, by the one or more processors, a distribution of shifted first values ​​for the input parameter vector, such that an estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of shifted first values ​​satisfies a predetermined functional response criterion; determining, by the one or more processors, a first relative distribution shift based on a comparison of the distribution of first values ​​and the shifted distribution of first values, wherein the first relative distribution shift is an indicator of a predicted change in cardiac function from the first set of conditions to a set of conditions associated with the predetermined functional response criterion.

37. 37. The method of claim 36, further comprising outputting, by the one or more processors, the first relative population shift.

38. comparing, by the one or more processors, the first relative distribution shift to a plurality of relative distribution shifts, the plurality of relative distribution shifts being associated with a corresponding plurality of medications; 37. The method of claim 36, further comprising identifying, by the one or more processors, a first medication from the corresponding plurality of medications based on the comparison.

39. 39. The method of claim 38, further comprising acquiring, by the one or more processors, a second plurality of in vitro responses comprising measures of the at least one functional response of the second cardiac cell culture under a second set of conditions, the second set of conditions being associated with perturbation of the second cardiac cell culture involving the first agent.

40. 40. The method of claim 39, wherein the first cardiac cell culture and the second cardiac cell culture are identical.

41. comparing, by the one or more processors, the second plurality of in vitro responses to the predetermined functional response criteria; 40. The method of claim 39, further comprising: identifying, by the one or more processors, an effect associated with the first medication based on the comparison.

42. 42. The method of claim 41, further comprising outputting, by the one or more processors, the effect.

43. 42. The method of claim 41, wherein the at least one functional response comprises a first functional response and a second functional response.

44. 44. The method of claim 43, wherein the predetermined functional response criterion is associated with the second functional response.

45. 45. The method of claim 44, wherein the effect is associated with a change in the second functional response.

46. 37. The method of claim 36, wherein the at least one first functional response measure comprises one or more waveform feature values.

47. 47. The method of claim 46, wherein the one or more waveform feature values ​​comprise one or more of a twitch amplitude value, a contraction time value, a maximum contraction slope value, a relaxation time value, a maximum relaxation slope value, and a twitch duration value.

48. 47. The method of claim 46, wherein the one or more waveform feature values ​​comprise spectral feature values.

49. 47. The method of claim 46, wherein the one or more waveform feature values ​​for the first plurality of in vitro responses are determined from a first functional response waveform of the first cardiac cell culture under the first set of conditions acquired over a first period of time.

50. 37. The method of claim 36, wherein the first plurality of in vitro responses is obtained from a bioreactor containing the first cardiac cell culture.

51. 37. The method of claim 36, wherein the first cardiac cell culture comprises a two-dimensional cardiac cell culture.

52. 37. The method of claim 36, wherein the first cardiac cell culture comprises three-dimensional cardiac tissue.

53. The method of claim 36 , wherein the mechanistic model comprises a mathematical model.

54. 54. The method of claim 53, wherein the mathematical model comprises one or more ordinary differential equations.

55. 37. The method of claim 36, wherein the input parameter vector comprises a plurality of input parameters, each of which takes on an input value.

56. 56. The method of claim 55, wherein the distribution of first values ​​comprises a distribution for each of the plurality of input parameters.

57. 37. The method of claim 36, wherein the distribution of the first values ​​for the input parameter vector is determined using a first learning algorithm.

58. 58. The method of claim 57, wherein the first learning algorithm comprises a Markov Chain Monte Carlo (MCMC) algorithm.

59. 59. The method of claim 58, wherein the MCMC is constrained according to one or more groupings of parameters in the input parameter vector.

60. 58. The method of claim 57, wherein the first learning algorithm comprises a generative adversarial network.

61. 58. The method of claim 57, wherein the first learning algorithm comprises a hierarchical Bayesian algorithm.

62. 62. The method of claim 61 , wherein the hierarchical Bayesian model comprises one or more prior distributions associated with one or more groupings of parameters in the input parameter vector.

63. 63. The method of claim 62, wherein the distribution of shifted first values ​​for the input parameter vector is determined by shifting one or more of the one or more prior distributions.

64. 37. The method of claim 36, wherein the at least one functional response comprises contractile force of the first cardiac cell culture.

65. 37. The method of claim 36, wherein the at least one functional response comprises a calcium transient response of the first cardiac cell culture.

66. 37. The method of claim 36, wherein the at least one functional response comprises an electrical response of the first cardiac cell culture.

67. A non-transitory machine-readable medium storing instructions that, when executed by one or more processors of a device, cause the one or more processors of the device to: obtaining a first plurality of in vitro responses comprising measurements of at least one functional response of the first cardiac cell culture under a first set of conditions; obtaining a mechanistic model of cardiac function, the mechanistic model predicting one or more values ​​of the at least one functional response based on an input parameter vector; determining a first distribution of values ​​for the input parameter vector, wherein the determination is performed such that an estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining a distribution of shifted first values ​​for the input parameter vector, wherein the determination is performed such that the estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of shifted first values ​​satisfies a predetermined functional response criterion; determining a first relative distribution shift based on a comparison of the distribution of first values ​​to the shifted distribution of first values, the first relative distribution shift being an indicator of a predicted change in cardiac function from the first set of conditions to a set of conditions associated with the predetermined functional response criterion.

68. one or more processors; a memory for storing instructions, When the instructions are executed by the one or more processors, the instructions cause the one or more processors to: obtaining a first plurality of in vitro responses comprising measurements of at least one functional response of the first cardiac cell culture under a first set of conditions; obtaining a mechanistic model of cardiac function, the mechanistic model predicting one or more values ​​of the at least one functional response based on an input parameter vector; determining a first distribution of values ​​for the input parameter vector, wherein the determination is performed such that an estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining a distribution of shifted first values ​​for the input parameter vector, wherein the determination is performed such that the estimate of the at least one functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of shifted first values ​​satisfies a predetermined functional response criterion; determining a first relative distribution shift based on a comparison of the distribution of the first values ​​with the shifted distribution of the first values, the first relative distribution shift being an indicator of a predicted change in cardiac function from the first set of conditions to a set of conditions associated with the predetermined functional response criterion.

69. 1. A method for predicting changes in cardiac function, comprising: acquiring, by one or more processors, a first plurality of in vitro responses from a bioreactor containing a first cardiac cell culture, the first plurality of in vitro responses comprising measurements of a first functional response of the first cardiac cell culture under reference conditions; acquiring, by the one or more processors, a second plurality of in vitro responses from the bioreactor, the second plurality of in vitro responses comprising measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with a first perturbation of the first cardiac cell culture; obtaining, by the one or more processors, a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector; determining, by the one or more processors, a distribution of first values ​​for the input parameter vector, such that an estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the distribution of first values ​​approximates a measurement from the first plurality of in vitro responses; determining, by the one or more processors, a second distribution of values ​​for the input parameter vector, such that an estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates a measurement from the second plurality of in vitro responses; determining, by the one or more processors, a first relative distribution shift based on a comparison of the distribution of the first values ​​and the distribution of the second values, wherein the first relative distribution shift is indicative of a change in function of the first cardiac cell culture due to the perturbation condition; and generating one or more outputs that identify or are indicative of one or more of: (a) a target molecule as a potential drug candidate; and / or (b) a potential effect of an unknown compound based on the change in function of the first cardiac cell culture due to the perturbation conditions.

70. 1. A system configured to predict a change in cardiac function, the system comprising: one or more bioreactors configured to grow cardiac tissue therein; a control unit communicatively coupled to the bioreactor and including one or more processors configured to execute instructions, the instructions including: obtaining a first plurality of in vitro responses from the bioreactor containing a first cardiac cell culture of the cardiac tissue, the first plurality of in vitro responses comprising measurements of a first functional response of the first cardiac cell culture under reference conditions; acquiring a second plurality of in vitro responses from the bioreactor, the second plurality of in vitro responses comprising measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with a first perturbation of the first cardiac cell culture; obtaining a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector; determining a first distribution of values ​​for the input parameter vector, wherein the first functional response estimate obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining a second distribution of values ​​for the input parameter vector, wherein the determination is performed such that the estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates measurements from the second plurality of in vitro responses; determining a first relative distribution shift based on a comparison of the distribution of the first values ​​to the distribution of the second values, wherein the first relative distribution shift is indicative of a change in function of the first cardiac cell culture due to the perturbation condition; The system generates one or more outputs that identify or are indicative of one or more of: (a) a target molecule as a potential drug candidate; and / or (b) a potential effect of an unknown compound based on the change in function of the first cardiac cell culture due to the perturbation conditions.

71. 1. A non-transitory computer-readable medium storing instructions for predicting changes in cardiac function, the instructions, when executed by one or more processors of a device, causing the one or more processors of the device to: obtaining a first plurality of in vitro responses from a bioreactor containing a first cardiac cell culture, the first plurality of in vitro responses comprising measurements of a first functional response of the first cardiac cell culture under reference conditions; acquiring a second plurality of in vitro responses from the bioreactor, the second plurality of in vitro responses comprising measurements of the first functional response of the first cardiac cell culture under perturbation conditions, the perturbation conditions being associated with a first perturbation of the first cardiac cell culture; obtaining a mechanistic model of cardiac function, the mechanistic model determining one or more estimates of the first functional response based on an input parameter vector; determining a first distribution of values ​​for the input parameter vector, wherein the first functional response estimate obtained from the mechanistic model based on input values ​​obtained from the first distribution of values ​​approximates a measurement from the first plurality of in vitro responses; determining a second distribution of values ​​for the input parameter vector, wherein the determination is performed such that the estimate of the first functional response obtained from the mechanistic model based on input values ​​obtained from the second distribution of values ​​approximates measurements from the second plurality of in vitro responses; determining a first relative distribution shift based on a comparison of the distribution of the first values ​​to the distribution of the second values, wherein the first relative distribution shift is indicative of a change in function of the first cardiac cell culture due to the perturbation condition; A non-transitory computer-readable medium that generates one or more outputs that identify or are indicative of one or more of: (a) a target molecule as a potential drug candidate; and / or (b) a potential effect of an unknown compound based on the change in function of the first cardiac cell culture due to the perturbation conditions.

Images