Prediction of preclinical pharmacokinetics using a neural ordinary differential equation framework

Abstract

A computer-based method and system for predicting preclinical pharmacokinetics of a therapeutic agent. Dosage regimen data associated with a dosage regimen of a therapy is received. Also received is medical data associated with a subject. Model input data is formed using the dosage regimen data and the medical data. Using a machine learning model and the model input data, a predicted pharmacokinetic (PK) outcome output is generated. The machine learning model may be a neural ordinary differential equation (NODE) system. A recommendation for a clinical trial is formed, using the predicted PK output.

Description

PREDICTION OF PRECLINICAL PHARMACOKINETICS USING A NEURAL ORDINARY DIFFERENTIAL EQUATION FRAMEWORKInventors: I raj Hosseini, Louis Russell JoslynCross-Reference to Related Application

[0001] This application is related to and claims the benefit of the priority date of U.S. Provisional Application No. 63 / 647,614, filed May 14, 2024, and entitled "Prediction of Preclinical Pharmacokinetics Using a Neural Ordinary Differential Equation Framework", which is incorporated herein by reference in its entirety.FIELD

[0002] The present disclosure relates to the prediction of preclinical pharmacokinetics of T- cell dependent bispecific molecules, and more particularly, to Al-based machine learning based prediction of preclinical pharmacokinetics of T-cell dependent bispecific molecules without initial pharmacokinetic data points.BACKGROUND

[0003] Drug development may be a multi-step process, starting with drug discovery and development, followed by preclinical research and clinical research to ensure product safety and efficacy before it can be used commercially. Pharmacokinetics (PK) and pharmacodynamics (PD) may be studied to understand movement of the drugs through the body as well as the body's biological response to drugs, respectively. Specifically, PK may be studied to understand movement of a therapeutic after administration, through four different stages: absorption, distribution, metabolism, and excretion (ADME). PK may be analyzed throughout clinical trials. However, in some cases, PK may be assessed before a drug is administered to clinical trial patients to estimate safe and effective dosages for administration. Predicting these types of preclinical PK may be important.SUMMARY

[0004] In one or more embodiments, a computer-based method of predicting pharmacokinetics of a therapeutic agent in a preclinical setting is provided. Dosage regimen data associated with a dosage regimen of a therapy is received. Medical data associated with a subject is also received. Model input data is used using the dosage regimen data and the medical data. A neural ordinary differential equation (NODE) system generates, using the model input data, a predicted pharmacokinetic (PK) outcome output that may include a recommended action for a clinical trial and the subject using the predicted PK outcome output.

[0005] In one or more embodiments, a system comprises one or more data processors; and a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform one or more of the methods described herein.

[0006] In one or more embodiments, a computer-program product tangibly embodied in a non-transitory machine-readable storage medium is provided, which includes instructions configured to cause one or more data processors to perform one or more of the methods described herein.BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The accompanying drawings, which are incorporated in and constitute a part of this specification, show certain aspects of the subject matter disclosed herein and, together with the description, help explain some of the principles associated with the disclosed implementations. In the drawings,

[0008] FIG. 1 is a block diagram of a preclinical pharmacokinetic prediction system in accordance with one or more example embodiments of the present disclosure.

[0009] FIG. 2 is a flowchart of a process for predicting a pharmacokinetic outcome in accordance with various embodiments of the present disclosure.

[0010] FIG. 3 is a flowchart of a process for predicting preclinical pharmacokinetics of a therapeutic agent in accordance with various embodiments of the present disclosure.

[0011] FIG. 4 is a flowchart illustrating an embodiment of a process for training the model in accordance with various embodiments of the present disclosure.

[0012] FIGS. 5-6 are illustrations of graphs associated with results based on example input data provided to a decoder in accordance with various embodiments of the present disclosure.

[0013] FIG. 7 are illustrations of graphs associated with example results in accordance with various embodiments of the present disclosure.

[0014] FIG. 8 are illustrations of graphs associated with example results in accordance with various embodiments of the present disclosure.

[0015] FIG. 9 is a block diagram of a computing system in accordance with various embodiments.

[0016] It is to be understood that the figures are not necessarily drawn to scale, nor are the objects in the figures necessarily drawn to scale in relationship to one another. The figures are depictions that are intended to bring clarity and understanding to various embodiments of apparatuses, systems, and methods disclosed herein. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts. Moreover, it should be appreciated that the drawings are not intended to limit the scope of the present teachings in any way.DETAILED DESCRIPTIONI. Overview

[0017] The embodiments recognize and take into account that some currently available methods and systems for predicting pharmacokinetic (PK) outcomes may include using models (e.g., machine learning models) that are trained on a single dosing regimen. In some cases, these trained models may be unable to predict PK outcomes for different dosing regiments with the desired level of accuracy. Further, in some cases, training such models may require that one PK cycle of data is known for the patients so that one cycle can be used to predict long term clinical PK. Using these trained models to predict PK outcomes may be more challenging in certain settings than desired and / or may consume more resources than desired. Further, these models may be less useful for predicting PK outcomes where very little early data (e.g., first cycle or preclinical data) is available or known. Further, in some cases, generating the initial PK data for various dosage instructions of various therapeutic agents may be too cumbersome or otherwise infeasible given other clinical trial constraints, such as budget, manpower, time, etc.

[0018] Thus, the embodiments described herein provide methodologies and systems for improved prediction of PK outcomes in a preclinical setting, without requiring initial PK data (e.g., first cycle data) and / or in a manner that allows for application across various dosing regiments (e.g., even unseen dosing regiments). The methodologies and systems disclosed herein relate to automated prediction of preclinical PK outcomes based on algorithms that use a neural ordinary different equations (NODE) framework. The embodiments described herein improve one or more technical fields or technologies, such as for example, the technical field / technology of PK outcome prediction. For example, the embodiments disclosed herein improve the technical field of PK outcome prediction by predicting PK outcomes without the need for initial PK data - unlike conventional systems, allowing for an increased number of dosage instructions and / or therapeutic agents to be assessed for PK outcomes.

[0019] Recognizing and taking into account the importance and utility of a methodology and system that can provide the improvements described above, described herein are various embodiments for Al-based machine learning prediction of preclinical PK outcomes.II. Example System for Predicting Preclinical Pharmacokinetic (PK) Outcome

[0020] Referring now to the figures, FIG. 1 is a block diagram of a pharmacokinetic (PK) prediction system 100 in accordance with various embodiments. Generally, the PK prediction system 100 (also referred to herein as a "system") is used to predict PK outcomes of various dosage levels of a therapeutic agent based on input data 102, which may include therapeutic agent input 104 and medical data 106. The PK prediction system 100 may be used in a hospital setting to predict PK outcomes for subjects before treatment, a preclinical setting to predict PK outcomes for subjects before being selected for clinical trial, a clinical trial setting to predict PK outcomes for subjects before undergoing a clinical trial, or a different type of research setting. For instance, and without limitation, PK prediction system 100 may be used to predict preclinical pharmacokinetics of T-cell dependent bispecific molecules.

[0021] In some embodiments, the PK prediction system 100 includes a computing platform 108, data storage 110, and a display system 112. The PK prediction system 100 also includes a PK predictor 114 that predicts PK outcomes, as determined at a future point in time. For example,the PK prediction system 100 may be used to predict PK outcomes after a selected number of hours, days, weeks, or months (e.g., 0.5 hours, 1 hour, 12 hours, 1 day, 2 days, 4 days, 10 days, 15 days, 10 days, 25 days, etc.). In one or more embodiments, the PK predictor 114 includes a data manager 116 and a model system such as an Ordinary Differential Equations Neural Network ("Neural ODE" or "NODE") 118. Generally, the data manager 116 generates model input data 120 using the input data 102. The model input data 120 may be sent to the neural ODE 118 for processing, and the neural ODE 118 uses the model input data 120 to generate a PK outcome 122. A PK outcome output 124 is generated based on the PK outcome 122. The PK Predictor 114 may generate a set of recommended actions 126 and / or a report 128 using the PK outcome output 124.

[0022] Generally, the input data 102 may be received from a remote system, may be retrieved from a data store, may be retrieved from the data storage 110, may be input by a user, and / or may be obtained in some other manner.

[0023] The therapeutic agent input 104 may include, for example, dosing regimen data for a therapy, which may include identification of one or more therapeutic agents. In some embodiments, the therapy may include T-cell dependent bispecific molecules. The dosing regimen data may include, for example, data (e.g., instructions) regarding a dosage level and a dose time (e.g., dosage frequency and / or dosage schedule). The dosage level may be, for example, the dosage amount per dose, which may remain constant or may change over time (e.g., a different dosage for earlier doses as compared to later doses).

[0024] The medical data 106 may include a body weight, the sex, and / or the age of the subject. In some embodiments, the medical data also includes other types of demographic information.

[0025] The computing platform 108 may take various forms. In one or more embodiments, the computing platform 108 includes a single computer (or computer system) or multiple computers (e.g., one or more processors where a processor may include one or more processors) in communication with each other. In other examples, the computing platform 108 takes the form of a cloud computing platform.

[0026] Generally, the data storage 110 and the display system 112 are each in communication with the computing platform 108. In some examples, the data storage 110, the display system 112, or both may be considered part of or otherwise integrated with the computing platform 108. Thus, in some examples, the computing platform 108, the data storage 110, and the display system 112 may be separate components in communication with each other, but in other examples, some combination of these components may be integrated together. Communication between the different components may be implemented using any number of wired communications links, wireless communications links, optical communications links, or a combination thereof.

[0027] The PK predictor 114, the data manager 116, and / or the neural ODE 118 may be implemented using software, hardware, firmware, or a combination thereof.

[0028] Generally, the data manager 116 generates the model input data 120 using the input data 102. In various embodiments, the data manager 116 may include a machine learning model, such as a neural network (e.g., CNN, ANN, or the like). The data manager 116 may receive the input data 102 as an input and determine the model input data 120 as an output based on the input data 102.

[0029] In various embodiments, the neural ODE 118 generates the PK outcome output 122, as discussed further below herein. The model system such as the neural ODE 118 may include a machine-learning model that may be comprised of any number of or combination of models, algorithms, equations, formulas, etc. In some embodiments, the neural ODE 118 may include one or more encoder-decoder frameworks. In some embodiments, the neural ODE 118 includes a neural ODE framework. The neural ODE 118 generates the PK outcome output 122.

[0030] Now referring to FIG. 2, a block diagram of one embodiment of the neural ODE 118 from FIG. 1 is illustrated in accordance with one or more embodiments of the present disclosure. In one or more embodiments, the neural ODE 118 may use input data, such as input data 102, to generate the PK outcome 122. As illustrated, the neural ODE 118 may include a neural ODE framework having an encoder 118A, a neural ODE 118B, and a decoder 118C. In some embodiments, the encoder 118A is or includes a gated recurrent unit neural network (GRU NN), and the decoder 118C is or comprises a multilayer perception neural network (MLP NN). Theencoder 118A may be configured to handle irregularly sampled data. The decoder 118C may be configured to use pharmacological features to inform predictions. The neural ODE 118B may include a latent neural ODE that, for example, learns dynamic PK. In some embodiments, the neural ODE 118B comprises a solution to ODEs in the latent space. The combination of the encoder, neural ODE, and decoder may provide hyperparameter tunning to determine optimal values. In some embodiments, a neural ODE framework, such as the neural ODE framework of FIG. 2, for preclinical PK may be trained with NHP PK to benchmark the predictive capabilities of a neural ODE framework in preclinical settings. Thus, a pharmacologically-informed neural ODE framework (e.g., for preclinical PK) is created and used to determined PK predictions without initial PK data points.

[0031] In several embodiments, the GRU NN may include a recurrent neural network (RNN) having gating mechanisms. The GRU NN (also referred to herein as a "GRU") may be configured to model sequential data, such pharmacodynamics, biomarker measurements, digital measurements, or other time-series analysis, by allowing information to be selectively remembered or discarded over a period of time. For example, the GRU may be configured to process sequential data one element at a time to update a hidden state based on the current input and the previous hidden state. In various embodiments, the GRU may include a candidate activation vector, which may be updated using a reset gate and an update gate. The reset gate may be configured to determine how much of a past / previous hidden state should be forgotten, and the update gate may determine how much information from the previous hidden state should be retained for a subsequent time step (e.g., update gate may be configured to determine how much of a candidate activation vector, which represents new candidate values that could be added to the memory of the GRU, to incorporate into a new hidden state).

[0032] In several embodiments, the multilayer perception neural network (also referred to herein as a "MLP") includes a neural network configured to model complex relationships between inputs and outputs. The MLP may include an input layer, one or more hidden layers, and an output layer. The input layer may include nodes, where each node corresponds to an input (e.g., input feature). The one or more hidden layers may be configured to process information received from the input layer and may each include any number of nodes. Theoutput layer may include one or more nodes and generate the prediction (e.g., PK outcome) based on the one or more hidden layers.

[0033] In some embodiments, the model inputs (e.g., input data 102) for the encoder 118A may include time, dose amount, nominal dose, and body weight. In some embodiments, the model inputs (e.g., input data 102) for the decoder 118C include body weight, dose time, and dose amount. In some embodiments, the model inputs for the decoder 118C consist of body weight, dose time, and dose amount. In some embodiments, decoder 118C has less inputs than the encoder 118A. As illustrated in FIG. 2, and in some embodiments, the neural ODE 118 receives the input data 102 without the data manager 116 generating the input data 120 using the input data 102. As such, and in some embodiments, the data manager 116 is omitted from the system 100 and at least a portion of the therapeutic agent input 104 and / or the medical data 106 is directly input into the neural ODE 118.

[0034] During a training stage, the input data 120 may include training input data 130 (illustrated in FIG. 1) that is used to train the neural ODE 118 to generate the PK outcome output 124 with a desired level of accuracy. This training input data 130 may include, for example, training pharmacokinetic data (e.g., early or initial PK data from a first cycle) and / ortraining input data (e.g., training dosing regimen data). Once the neural ODE 118 has been trained, it can be used to generate the PK outcome output 124 for a selected therapy based on input data specific to that selected therapy, even where the dosing regimen is different from the dosing regimen on which the neural ODE 118 was trained.

[0035] PK outcome output 122 may take different forms that relate to a PK outcome (e.g., various PK outcomes relating to absorption, distribution, metabolism, and / or excretion (ADME) of the therapy). In some embodiments, the PK outcome output 122 may include a score, value, metric, and / or assessment that provides a prediction of a PK outcome.

[0036] The PK outcome output 122 may be used to form a recommendation (e.g., set of recommended actions 126) for each dosage instruction of set of dosage instructions. In some embodiments, the PK outcome output 124 may be compared to a predetermined threshold appropriate for the output type, in order to form the recommendation. In some embodiments, the PK outcome output 124 comprises identifying the subject as a subject that is highly likely torespond to the therapy, such as for example when the predicted PK outcome or response is within a predetermined range or outside of a predetermined range.

[0037] In some embodiments, the PK predictor 114 may visually present PK outcome output 118 using the display system 112. In some cases, the PK predictor 114 may additionally, or alternatively, visually present at least a portion of the PK outcome output 124, the set of recommended actions 126, or both using the display system 112. In some embodiments, the PK predictor 114 generates the report 128 that includes the PK outcome 122, the PK outcome output 124 (to the extent that it is different from or does not include the PK outcome 122), the set of recommended actions 126, or a combination thereof. The recommended actions 126 may include, but are not limited to, adjusting and / or updating a therapy, such as adjusting a dosage level, increasing or decreasing a frequency of administration of a dosage, altering a type of therapeutic agent used (e.g., recommending a different therapeutic agent), and / or the like. In some embodiments, recommended action comprises recommending the subject for inclusion into a clinical trial when the subject is identified as highly likely to respond to the therapy. In other embodiments, however, the recommended action comprises recommending the subject for exclusion from a clinical trial when the subject is identified as highly likely to respond to the therapy. In some embodiments, the recommended action comprises a recommendation to administer the dosage regimen of the therapy to the subject in a clinical trial or to administer a dosage regimen that is different from the dosage regiment to the subject in a clinical trial. In some embodiments, the recommendation action is a recommendation to administer the set of dosage instructions during the clinical trial to the subject. Additionally, the recommendation action may be a recommendation to administer the set of dosage instructions during the clinical trial to subjects with a body weight that is similar to the body weight of the subject.

[0038] Generally, the system 100 may improve result accuracy (e.g., result optimization), increase computing efficiency, and / or reduce computing resources used for determining the PK outcome of a subject without providing initial PK data points associated with the subject. The system 100 may additionally provide recommended actions and / or reports based on the PK outcome output.

[0039] While described herein as a preclinical PK system 100, the PK system 100 is not limited to a preclinical setting. Instead, the PK system 100 may be used in a hospital setting to predict PK outcomes for subjects before treatment or a different type of research setting.III. Example Methodologies for Predicting Pharmacokinetic (PK) Outcome

[0040] FIG. 3 is a flowchart of a process 300 for predicting pharmacokinetics of a therapeutic agent, in accordance with various embodiments. In various embodiments, the process 300 is implemented using the pharmacokinetic prediction system 100 described in FIG. 1. The process 300 includes various steps and may be described with continuing reference to FIGS. 1-2. One or more steps that are not expressly illustrated in FIG. 3 may be included before, after, in between, or as part of the steps of the process 300. In some embodiments, the process 300 may begin with step 302.

[0041] The process 300 may optionally include the step 301 of training a model (e.g., the neural ODE 118). Training the model may include any one of the example training processes described herein. The model may be trained to process input data and generate a PK outcome, a PK outcome output, a set of recommended actions, and / or a report. The PK outcome may be, for example, the PK outcome 122 in FIG. 1, the PK outcome output may be, for example, the PK outcome output 124 in FIG. 1, the set of recommended actions may be, for example, the set of recommended actions 126 in FIG. 1, and the report may be, for example, the report 128 in FIG. 1.

[0042] Step 302 includes receiving therapeutic agent input and medical data input of a subject. The therapeutic agent input may include the dosage regimen data for a therapy. The therapy may include at least one therapeutic agent. A therapeutic agent may include T-cell dependent bispecific molecules. The dosage regimen data may include, for example, a time for a dose, a dose amount, a nominal dose amount, and / or other types of dosage-related information. Generally, the therapeutic agent input is associated with a subject and the medical data may include the body weight of the subject. In some embodiments, the therapeutic agent input received in step 302 is the therapeutic agent input 104 in FIG. 1 and the medical data received in step 302 is the medical data 106 in FIG. 1. In some embodiments, the therapeuticagent input and the medical data input is received at the step 302 by the PK prediction system 100, such as by the data manager 116. The therapeutic agent input and the medical data input received at the step 302 may be received from the data storage 110 and / or a user of the PK system 100.

[0043] Step 304 includes forming model input data using the dosage regimen data and the medical data input. The model input data includes at least one of a time, a dose amount, a nominal dose, or a body weight. In some embodiments, the model input data is the input data 120 in FIG. 1. In some embodiments, the data manager 116 in FIG. 1 forms the input data 120 using the input data 102 in FIG. 1. In some embodiments, the model input data does not include initial PK data points associated with the subject.

[0044] Step 306 includes generating, using a neural ordinary differential equation (NODE) system and the model input data, a predicted pharmacokinetic (PK) outcome output. Generally, the predicted PK outcome is associated with the dose and the subject. In some embodiments, the machine learning model is an ordinary differential equations neural network (neural ODE). The PK outcome output may include, for example, a PK concentration for a given time point or plurality of time points within a time period, a predicted time for the therapy (e.g., therapeutic agent) to be absorbed by the body, and / or other PK-related information. In some embodiments, the neural ODE 118 in FIG. 1 generates the PK outcome output, which may be the PK outcome output 124 in FIG. 1. In some embodiments, initial PK data points or an initial PK response associated with the subject are not considered during the step 306. For example, the PK predictor 114 does not use initial PK data points associated with the subject when predicting the PK outcome associated with the subject.

[0045] Step 308 includes generating an output that comprises a recommended action for a clinical trial using the predicted PK outcome output. In some embodiments, the PK prediction system 100 in FIG. 1 generates the output, which may be for example, the PK outcome output 124, the set of recommended actions 126, and / orthe report 128 in FIG. 1. In some embodiments, the recommendation is not associated with a clinical trial.

[0046] In some embodiments, the process 300 predicts PK outcomes for a subject without using initial PK data points associated with the subject, as previously discussed in FIG. 2. Theprocess 300 provides a technical effect of improving accuracy, reducing the overall computing resources, and / or reducing the time needed to predict a PK outcomes for a subject without using initial PK data points associated with the subject.

[0047] In some embodiments, the process 300 provides a technical improvement to the technical field of predicting a specific subject's predicted PK response. In some embodiments, the process 300 includes a new combination of steps that results in the technical improvement over conventional PK response prediction systems.IV. Example Methodologies of Training Portions of the Example PK System

[0048] FIG. 4 is a flowchart illustrating an embodiment of a process 401 fortraining the model in accordance with one or more embodiments. In one or more embodiments, the process 401 may be implemented using the PK predictor 114 described in FIGS. 1 and 2. One or more steps that are not expressly illustrated in FIG. 4 may be included before, after, in between, or as part of the steps of process 401. In some embodiments, process 401 may begin with step 402.

[0049] In some cases, step 402 includes receiving training input data that includes dosing regimes and individual PK profiles across time. In some embodiments, the training input data included the first 7 days of PK concentrations as input for the encoder. In some embodiments, the training input data includes or is the training input data 130 of FIG. 1.

[0050] The process 401 further includes, at step 404, forming model input data using the training input data. In some embodiments, the model input data includes the individual PK profiles across time, a dose amount, a nominal dose, and a body weight. In some embodiments, the data manager 116 in FIG. 1 forms the input data using the training input data 130 in FIG. 1.

[0051] Step 406 of process 401 includes training the model to generate a PK outcome. In some embodiments, the model is the neural ODE 118 and the PK outcome may be, for example, the PK outcome 122 in FIG. 1. In some embodiments, training input data includes dosing regimes for four different dosing levels and associated PK profiles across time. In some embodiments, training the model includes changing the parameters of the encoder and decoder of the neural ODE 118. For example, comparing results for providing 5 parameters to the decoder, providing2 parameters to the decoder, and providing 3 parameters to the decoder were compared to determine that providing 3 parameters to the decoder provided better results.

[0052] FIGS. 5-6 show graphs representing experimentation with providing different input data to the decoder. More specifically, FIGS. 5-6 shows the results of model training and corresponding prediction outcomes for patient PK profiles. The results indicate that having more inputs for the decoder and / or having the least amount of inputs did not yield optimal predictions. Rather, the results indicated that optimization of predictions was achieved using a training process that used hyperparameter tuning, as previously discussed herein.

[0053] FIG. 5 provides a comparison of the results 505 associated with providing 5 parameters to the decoder (e.g., body weight, nominal dose, initial PK concentrations, dose amount, and dose time), results 510 associated with providing 2 parameters to the decoder (e.g., nominal dose and body weight), results 515 associated with providing 3 parameters to the decoder (e.g., body weight, dose amount, and dose time). As illustrated, the prediction results 515 most closely match the actual results compared to the prediction results 505 and 510.

[0054] FIG. 6 provides a comparison of the results 515 associated with providing 3 parameters to the decoder (e.g., body weight, dose amount, and dose time) without tuning and the results 605 associated with providing 3 parameters to the decoder (e.g., body weight, dose amount, and dose time) with tuning. As illustrated, the prediction results 605 more closely match the actual results compared to the prediction results 515.

[0055] Other training processes may also be implemented. The training data may include different data such as, for example, training thickness maps, training intensity projection maps, training data that includes training values for various biomarker features, or a combination thereof.V. Example ExperimentsV.A. Example PK Prediction

[0056] In one example experiment, an example PK prediction system, such as the PK prediction system 100, was used to predict preclinical PK across several dose levels spanning three orders of magnitude. In this example, the training dataset was associated with a T-cellengaging bispecific antibody targeting CD20 / CD3, there was a total of 103 non-human primates with individual PK profiles across time, with a total of 1157 data points across animals. On average, there were 11 PK data points per animals and dose groups ranged from 0.001 to 1.0 mg / kg. In this example, the neural ODE was trained on mosunetuzumab PK data in 83 cynomolgus monkeys (~80% of total dataset) and tested individualized model predictions against PK from 20 monkeys (~20% of total dataset). In this example, hyper parameter tuning was performed across epochs, learning rate and L2 regularization to result in a PINODE (pharmacologically-informed Neural ODE). 528 samples were selected following a Latin hyper cube sampling schedule and evaluated for lowest total RMSE in cross validation. The example PINODE does not require the incorporation of early PK time-points to make individualized (weight-based) PK predictions across a dose ranging study. FIG. 7 illustrates example results 700 of the PINODE including residual for all PK predictions across test data and the best individual PK prediction for each dose group. The right-most graph illustrates example results for a nominal dose of 0.1 mg / kg, the second to right-most graph illustrates example results for a nominal dose of 1 mg / kg, and the second from left-most graph illustrates example results for a nominal dose of 0.01 mg / kg. As illustrated, the PINODE predicts preclinical PK across different dose levels and regimes. Additionally and in this example, the ability of the example PINODE to extrapolate and make PK predictions for 'unseen' dose regimens ranging from 0.001 mg / kg to 1 mg / kg was tested. The PINODE was iteratively trained on 3 of the 4 dose groups (e.g., 0.001 mg / kg, 0.01 mg / kg, 0.1 mg / kg, and 1.0 mg / kg) and held out the fourth group as the test dataset. The PINODE was able to make reasonable predictions for unseen doses, with better results when predicting the middle two dose levels or the larger dose levels. FIG. 8 illustrates results 800 for predictions of each of the four groups. In some embodiments, the PINODE is the PK system 100.

[0057] As illustrated in the example experiment, the PK system 100 may be used to generalize across dosing strategies in preclinical settings and potentially be extended to project clinical PK profiles.

[0058] As illustrated, the example PINODE framework accurately reproduces the PK of mosunetuzumab in preclinical settings. The example framework represents a neural ODE thatcan predict individualized PK without having to provide initial PK data points to the encoder, significantly improving its relevance for preclinical applications.VI. Example Computing System

[0059] FIG. 9 is a block diagram illustrating an example of a computing system, in accordance with one or more example embodiments. Computing system 900 may be used to implement computing platform 108 in FIG. 1 and / or any components therein.

[0060] As shown in FIG. 9, and in one or more examples, the computer system 900 can include a bus 902 or other communication mechanism for communicating information, and a processor 904 coupled with bus 902 for processing information. In various embodiments, computer system 900 can also include a memory, which can be a random-access memory (RAM) 906 or other dynamic storage device, coupled to bus 902 for determining instructions to be executed by processor 904. Memory also can be used for storing temporary variables or other intermediate information during execution of instructions to be executed by processor 904. In various embodiments, computer system 900 can further include a read only memory (ROM) 908 or other static storage device coupled to bus 902 for storing static information and instructions for processor 904. A storage device 910, such as a magnetic disk or optical disk, can be provided and coupled to bus 902 for storing information and instructions.

[0061] In various embodiments, computer system 900 can be coupled via bus 902 to a display 912, such as a cathode ray tube (CRT) or liquid crystal display (LCD), for displaying information to a computer user. An input device 914, including alphanumeric and other keys, can be coupled to bus 902 for communicating information and command selections to processor 904. Another type of user input device is a cursor control 916, such as a mouse, a joystick, a trackball, a gesture input device, a gaze-based input device, or cursor direction keys for communicating direction information and command selections to processor 904 and for controlling cursor movement on display 912. This input device 914 typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), that allows the device to specify positions in a plane. However, it should be understood that an input devices 914allowing for three-dimensional (e.g., x, y, and z) cursor movement are also contemplated herein.

[0062] Consistent with certain implementations of the present teachings, results can be provided by computer system 900 in response to processor 904 executing one or more sequences of one or more instructions contained in RAM 906. Such instructions can be read into RAM 906 from another computer-readable medium or computer-readable storage medium, such as storage device 910. Execution of the sequences of instructions contained in RAM 906 can cause processor 904 to perform the processes described herein. Alternatively, hard-wired circuitry can be used in place of or in combination with software instructions to implement the present teachings. Thus, implementations of the present teachings are not limited to any specific combination of hardware circuitry and software.

[0063] According to some example embodiments, the input device 914 can provide input / output operations for a network device. For example, the input / output device 614 can include Ethernet ports or other networking ports to communicate with one or more wired and / or wireless networks (e.g., a local area network (LAN), a wide area network (WAN), the Internet).

[0064] In some example embodiments, the computing system 900 can be used to execute various interactive computer software applications that can be used for organization, analysis and / or storage of data in various formats. Alternatively, the computing system 900 can be used to execute any type of software applications. These applications can be used to perform various functionalities, e.g., planning functionalities (e.g., generating, managing, editing of spreadsheet documents, word processing documents, and / or any other objects, etc.), computing functionalities, communications functionalities, etc. The applications can include various add-in functionalities or can be standalone computing products and / or functionalities. Upon activation within the applications, the functionalities can be used to generate the user interface provided via the input / output device 914. The user interface can be generated and presented to a user by the computing system 900 (e.g., on a computer screen monitor, etc.).

[0065] One or more aspects or features of the subject matter described herein can be realized in digital electronic circuitry, integrated circuitry, specially designed ASICs, field programmable gate arrays (FPGAs) computer hardware, firmware, software, and / or combinations thereof.These various aspects orfeatures can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device. The programmable system or computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[0066] These computer programs, which can also be referred to as programs, software, software applications, applications, components, or code, include machine instructions for a programmable processor, and can be implemented in a high-level procedural and / or object- oriented programming language, and / or in assembly / machine language. As used herein, the term "machine-readable medium" refers to any computer program product, apparatus and / or device, such as for example magnetic discs, optical disks, memory, and Programmable Logic Devices (PLDs), used to provide machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and / or data to a programmable processor. The machine-readable medium can store such machine instructions non-transitorily, such as for example as would a non-transient solid- state memory or a magnetic hard drive or any equivalent storage medium. The machine- readable medium can alternatively or additionally store such machine instructions in a transient manner, such as for example, as would a processor cache or other random access memory associated with one or more physical processor cores.

[0067] To provide for interaction with a user, one or more aspects or features of the subject matter described herein can be implemented on a computer having a display device, such as for example a cathode ray tube (CRT) or a liquid crystal display (LCD) or a light emitting diode (LED) monitor for displaying information to the user and a keyboard and a pointing device, such as for example a mouse or a trackball, by which the user may provide input to the computer. Otherkinds of devices can be used to provide for interaction with a user as well. For example, feedback provided to the user can be any form of sensory feedback, such as for example visual feedback, auditory feedback, or tactile feedback; and input from the user may be received in any form, including acoustic, speech, or tactile input. Other possible input devices include touch screens or other touch-sensitive devices such as single or multi-point resistive or capacitive track pads, voice recognition hardware and software, optical scanners, optical pointers, digital image capture devices and associated interpretation software, and the like.

[0068] The term "computer-readable medium" (e.g., data store, data storage, storage device, data storage device, etc.) or "computer-readable storage medium" as used herein refers to any media that participates in providing instructions to processor 904 for execution. Such a medium can take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Examples of non-volatile media can include, but are not limited to, optical, solid state, magnetic disks, such as storage device 910. Examples of volatile media can include, but are not limited to, dynamic memory, such as RAM 906. Examples of transmission media can include, but are not limited to, coaxial cables, copper wire, and fiber optics, including the wires that comprise bus 902.

[0069] Common forms of computer-readable media include, for example, a floppy disk, a flexible disk, hard disk, magnetic tape, or any other magnetic medium, a CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, a RAM, PROM, and EPROM, a FLASH-EPROM, any other memory chip or cartridge, or any other tangible medium from which a computer can read.

[0070] In addition to computer readable medium, instructions or data can be provided as signals on transmission media included in a communications apparatus or system to provide sequences of one or more instructions to processor 904 of computer system 900 for execution. For example, a communication apparatus may include a transceiver having signals indicative of instructions and data. The instructions and data are configured to cause one or more processors to implement the functions outlined in the disclosure herein. Representative examples of data communications transmission connections can include, but are not limited to, telephone modemconnections, wide area networks (WAN), local area networks (LAN), infrared data connections, NFC connections, optical communications connections, etc.

[0071] It should be appreciated that the methodologies described herein, flow charts, diagrams, and accompanying disclosure can be implemented using computer system 1100 as a standalone device or on a distributed network of shared computer processing resources such as a cloud computing network.

[0072] The methodologies described herein may be implemented by various means depending upon the application. For example, these methodologies may be implemented in hardware, firmware, software, or any combination thereof. For a hardware implementation, the processing unit may be implemented within one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described herein, or a combination thereof.

[0073] In various embodiments, the methods of the present teachings may be implemented as firmware and / or a software program and applications written in conventional programming languages such as C, C++, Python, etc. If implemented as firmware and / or software, the embodiments described herein can be implemented on a non-transitory computer-readable medium in which a program is stored for causing a computer to perform the methods described above. It should be understood that the various engines described herein can be provided on a computer system, such as computer system 900, whereby processor 904 would execute the analyses and determinations provided by these engines, subject to instructions provided by any one of, or a combination of, the memory components RAM 906, ROM, 908, or storage device 910 and user input provided via input device 914.VII. Example Definitions and Context

[0074] The disclosure is not limited to these example embodiments and applications or to the manner in which the example embodiments and applications operate or are described herein.Moreover, the figures may show simplified or partial views, and the dimensions of elements in the figures may be exaggerated or otherwise not in proportion.

[0075] Unless otherwise defined, scientific and technical terms used in connection with the present teachings described herein shall have the meanings that are commonly understood by those of ordinary skill in the art. Further, unless otherwise required by context, singular terms shall include pluralities and plural terms shall include the singular. Generally, nomenclatures utilized in connection with, and techniques of, chemistry, biochemistry, molecular biology, pharmacology, and toxicology are described herein are those well-known and commonly used in the art.

[0076] As the terms "on," "attached to," "connected to," "coupled to," or similar words are used herein, one element (e.g., a component, a material, a layer, a substrate, etc.) can be "on," "attached to," "connected to," or "coupled to" another element regardless of whether the one element is directly on, attached to, connected to, or coupled to the other element or there are one or more intervening elements between the one element and the other element. In addition, where reference is made to a list of elements (e.g., elements a, b, c), such reference is intended to include any one of the listed elements by itself, any combination of less than all of the listed elements, and / or a combination of all of the listed elements. Section divisions in the specification are for ease of review only and do not limit any combination of elements discussed.

[0077] The term "subject" may refer to a subject of a clinical trial, a subject of study, a person undergoing treatment, a person undergoing anti-cancer therapies, a person being monitored for remission or recovery, a person undergoing a preventative health analysis (e.g., due to their medical history), or any other person or patient of interest. In various cases, "subject" and "patient" may be used interchangeably herein.

[0078] As used herein, "substantially" means sufficient to work for the intended purpose. The term "substantially" thus allows for minor, insignificant variations from an absolute or perfect state, dimension, measurement, result, or the like such as would be expected by a person of ordinary skill in the field but that do not appreciably affect overall performance. When used with respect to numerical values or parameters or characteristics that can be expressed as numerical values, "substantially" means within ten percent.

[0079] The term "ones" means more than one.

[0080] As used herein, the term "plurality" may be 2, 3, 4, 5, 6, 7, 8, 9, 10, or more.

[0081] As used herein, the term "set of" means one or more. For example, a set of items includes one or more items.

[0082] As used herein, the phrase "at least one of," when used with a list of items, means different combinations of one or more of the listed items may be used and only one of the items in the list may be needed. The item may be a particular object, thing, step, operation, process, or category. In other words, "at least one of" means any combination of items or number of items may be used from the list, but not all of the items in the list may be required. For example, without limitation, "at least one of item A, item B, or item C" means item A; item A and item B; item B; item A, item B, and item C; item B and item C; or item A and C. In some cases, "at least one of item A, item B, or item C" means, but is not limited to, two of item A, one of item B, and ten of item C; four of item B and seven of item C; or some other suitable combination.

[0083] As used herein, a "model" can refer to a system, process, relationship, or set of rules, which is instantiated, stored, and / or executed within a computing environment. The model may include, without limitation, algorithms, parameters, data structures, or executable instructions configured to perform one or more functions when processed by one or more computing devices.

[0084] As used herein, "machine learning" may be the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something in the world. Machine learning uses algorithms that can learn from data without relying on rules-based programming.

[0085] As used herein, an "artificial neural network" or "neural network" (NN) may refer to computational models that mimic an interconnected group of artificial nodes or neurons that processes information based on a connectionist approach to computation. Neural networks, which may also be referred to as neural nets, can employ one or more layers of nonlinear units to predict an output for a received input. Some neural networks include one or more hidden layers in addition to an output layer. The output of each hidden layer is used as input to the next layer in the network, e.g., the next hidden layer or the output layer. Each layer of the network generates an output from a received input in accordance with current values of a respective setof parameters. In the various embodiments, a reference to a "neural network" may be a reference to one or more neural networks.

[0086] A neural network may compute digital data in two ways: when it is being trained it is in training mode and when it puts what it has learned into practice it is in inference (or prediction) mode. Neural networks learn through a feedback process (e.g., backpropagation) which allows the network to adjust the weight factors (modifying its behavior) of the individual nodes in the intermediate hidden layers so that the output matches the outputs of the training data. In other words, a neural network learns by being fed training data (learning examples) and eventually learns how to reach the correct output, even when it is presented with a new range or set of inputs. A neural network may include, for example, without limitation, at least one of a Feedforward Neural Network (FNN), a Recurrent Neural Network (RNN), a Modular Neural Network (MNN), a Convolutional Neural Network (CNN), a Residual Neural Network (ResNet), an Ordinary Differential Equations Neural Networks (neural ODE), or another type of neural network.VIII. Recitation of Example Embodiments

[0087] Embodiment 1: A computer-based method of predicting pharmacokinetics of a therapeutic agent in a preclinical setting, the method comprising: receiving, by one or more processors, dosage regimen data associated with a dosage regimen of a therapy; receiving, by the one or more processors, medical data associated with a subject; forming model input data using the dosage regimen data and the medical data; generating, using a neural ordinary differential equation (NODE) system and the model input data, a predicted pharmacokinetic (PK) outcome output; and generating, by the one or more processors, an output that comprises a recommended action for a clinical trial and the subject using the predicted PK outcome output.

[0088] The method of embodiment 1, wherein the model input data comprises a body weight of the subject, a dose amount, a dose time, and a nominal dose; wherein the NODE system comprises an encoder that uses inputs comprising the dose amount, the nominal dose, the body weight, and the dose time; and wherein the NODE system comprises a decoder that uses inputs comprising the dose amount, the body weight, and the dose time.

[0089] The method of any one of embodiments 1-2, wherein the inputs for the decoder consist of the dose amount, the body weight, and the dose time.

[0090] The method of any one of embodiments 1-3, wherein an initial PK response of the subject is omitted from the model input data.

[0091] The method of any one of embodiments 1-4, wherein the therapy comprises at least one therapeutic agent that comprises T-cell dependent bispecific molecules.

[0092] The method of any one of embodiments 1-5, wherein the PK outcome output is a PK concentration.

[0093] The method of any one of embodiments 1-6, wherein the NODE system comprises an encoder, a neural ODE, and a decoder; and wherein the number of inputs for the encoder is larger than the number of inputs for the decoder.

[0094] The method of any one of embodiments 1-7, wherein the NODE system was trained using dosage regimen data associated with a first dosing regimen; and wherein the model input data is associated with a second dosing regimen that is different than the first dosing regimen.

[0095] The method of any one of embodiments 1-8, wherein the NODE system was trained using dosage regimen data associated with a first dosing regimen and a first therapy; and wherein the model input data is associated with a second dosing regimen that is different than the first dosing regimen and / or different than the first therapy.

[0096] The method of any one of embodiments 1-9, wherein the encoder comprises a gated recurrent unit neural network (GRU NN); wherein the neural ODE comprises a solution to ODEs in the latent space; and wherein the decoder comprises a multilayer perception neural network (MLP NN).

[0097] The method of any one of embodiments 1-10, wherein the NODE system was trained using training data comprising full PK responses for subjects; and wherein the trained NODE system provides the predicted PK outcome output for a subject without receiving an initial PK response for the subject.

[0098] The method of any one of embodiments 1-11, wherein generating, using the NODE system and the model input data, the predicted PK outcome output comprises the NODE system identifying the subject as a subject that is highly likely to respond to the therapy.

[0099] The method of any one of embodiments 1-12, wherein the subject that is highly likely to respond to the therapy is a subject with a predicted PK response within a predetermined range.

[0100] The method of any one of embodiments 1-13, wherein the subject that is highly likely to respond to the therapy is a subject with a predicted PK response outside of a predetermined range.

[0101] The method of any one of embodiments 1-14, wherein the recommended action for the clinical trial and the subject usingthe predicted PK outcome output comprises recommending the subject for inclusion into a clinical trial when the subject is identified as highly likely to respond to the therapy.

[0102] The method of any one of embodiments 1-14, wherein the recommended action for the clinical trial and the subject usingthe predicted PK outcome output comprises recommending the subject for exclusion from a clinical trial when the subject is identified as highly likely to respond to the therapy.

[0103] The method of any one of embodiments 1-16, wherein the recommended action for the clinical trial and the subject using the predicted PK outcome output is to administer the dosage regimen of the therapy to the subject in a clinical trial.

[0104] The method of any one of embodiments 1-17, wherein the recommended action for the clinical trial and the subject using the predicted PK outcome output is to administer a dosage regimen that is different from the dosage regimen to the subject in a clinical trial.

[0105] The method of any one of embodiments 1-18, wherein the dosage regimen of the therapy comprises a set of dosage instructions associated with the therapy; and wherein the recommended action for the clinical trial and the subject comprises a recommendation to administer the set of dosage instructions during the clinical trial to the subject.

[0106] The method of any one of embodiments 1-19, wherein the recommended action for the clinical trial and the subject further comprises a recommendation to administer the set of dosage instructions during the clinical trial to subjects with a body weight that is similar to the body weight of the subject.

[0107] A system comprising: one or more data processors; and a non -transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform the method of any one of embodiments 1-20.

[0108] A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform the method of any one of embodiments 1-20.IX. Additional Considerations

[0109] Any headers and / or subheaders between sections and subsections of this document are included solely for the purpose of improving readability and do not imply that features cannot be combined across sections and subsection. Accordingly, sections and subsections do not describe separate embodiments.

[0110] While the present teachings are described in conjunction with various embodiments, it is not intended that the present teachings be limited to such embodiments. On the contrary, the present teachings encompass various alternatives, modifications, and equivalents, as will be appreciated by those of skill in the art. The present description provides preferred example embodiments, and is not intended to limit the scope, applicability, or configuration of the disclosure. Rather, the present description of the preferred example embodiments will provide those skilled in the art with an enabling description for implementing various embodiments. It is understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope as set forth in the appended claims. Thus, such modifications and variations are considered to be within the scope set forth in the appended claims. Further, the terms and expressions which have been employed are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed.

[0111] In describing the various embodiments, the specification may have presented a method and / or process as a particular sequence of steps. However, to the extent that themethod or process does not rely on the particular order of steps set forth herein, the method or process should not be limited to the particular sequence of steps described, and one skilled in the art can readily appreciate that the sequences may be varied and still remain within the spirit and scope of the various embodiments.

[0112] Some embodiments of the present disclosure include a system including one or more data processors. In some embodiments, the system includes a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform part or all of one or more methods and / or part or all of one or more processes disclosed herein. Some embodiments of the present disclosure include a computer-program product tangibly embodied in a non-transitory machine- readable storage medium, including instructions configured to cause one or more data processors to perform part or all of one or more methods and / or part or all of one or more processes disclosed herein.

[0113] Specific details are given in the present description to provide an understanding of the embodiments. However, it is understood that the embodiments may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.

Claims

CLAIMSWhat is claimed is:

1. A computer-based method of predicting pharmacokinetics of a therapeutic agent in a preclinical setting, the method comprising: receiving, by one or more processors, dosage regimen data associated with a dosage regimen of a therapy; receiving, by the one or more processors, medical data associated with a subject; forming model input data using the dosage regimen data and the medical data; generating, using a neural ordinary differential equation (NODE) system and the model input data, a predicted pharmacokinetic (PK) outcome output; and generating, by the one or more processors, an output that comprises a recommended action for a clinical trial and the subject using the predicted PK outcome output.

2. The method of claim 1, wherein the model input data comprises a body weight of the subject, a dose amount, a dose time, and a nominal dose; wherein the NODE system comprises an encoder that uses inputs comprising the dose amount, the nominal dose, the body weight, and the dose time; and wherein the NODE system comprises a decoder that uses inputs comprising the dose amount, the body weight, and the dose time.

3. The method of any one of claims 1-2, wherein the inputs for the decoder consist of the dose amount, the body weight, and the dose time.

4. The method of any one of claims 1-3, wherein an initial PK response of the subject is omitted from the model input data.

5. The method of any one of claims 1-4, wherein the therapy comprises at least one therapeutic agent that comprises T-cell dependent bispecific molecules.

6. The method of any one of claims 1-5, wherein the PK outcome output is a PK concentration.

7. The method of any one of claims 1-6, wherein the NODE system comprises an encoder, a neural ODE, and a decoder; and wherein the number of inputs for the encoder is larger than the number of inputs for the decoder.

8. The method of any one of claims 1-7, wherein the NODE system was trained using dosage regimen data associated with a first dosing regimen; and wherein the model input data is associated with a second dosing regimen that is different than the first dosing regimen.

9. The method of any one of claims 1-8, wherein the NODE system was trained using dosage regimen data associated with a first dosing regimen and a first therapy; and wherein the model input data is associated with a second dosing regimen that is different than the first dosing regimen and / or different than the first therapy.

10. The method of any one of claims 1-9, wherein the encoder comprises a gated recurrent unit neural network (GRU NN); wherein the neural ODE comprises a solution to ODEs in latent space; and wherein the decoder comprises a multilayer perception neural network (MLP NN).

11. The method of any one of claims 1-10, wherein the NODE system was trained using training data comprising full PK responses for subjects; and wherein the trained NODE system provides the predicted PK outcome output for a subject without receiving an initial PK response for the subject.

12. The method of any one of claims 1-11, wherein generating, using the NODE system and the model input data, the predicted PK outcome output comprises the NODE system identifying the subject as a subject that is highly likely to respond to the therapy.

13. The method of any one of claims 1-12, wherein the subject that is highly likely to respond to the therapy is a subject with a predicted PK response within a predetermined range.

14. The method of any one of claims 1-13, wherein the subject that is highly likely to respond to the therapy is a subject with a predicted PK response outside of a predetermined range.

15. The method of any one of claims 1-14, wherein the recommended action for the clinical trial and the subject using the predicted PK outcome output comprises recommending the subject for inclusion into a clinical trial when the subject is identified as highly likely to respond to the therapy.

16. The method of any one of claims 1-14, wherein the recommended action for the clinical trial and the subject using the predicted PK outcome output comprises recommending the subject for exclusion from a clinical trial when the subject is identified as highly likely to respond to the therapy.

17. The method of any one of claims 1-16, wherein the recommended action for the clinical trial and the subject using the predicted PK outcome output is to administer the dosage regimen of the therapy to the subject in a clinical trial.

18. The method of any one of claims 1-17, wherein the recommended action for the clinical trial and the subject using the predicted PK outcome output is to administer a dosage regimen that is different from the dosage regimen to the subject in a clinical trial.

19. The method of any one of claims 1-18, wherein the dosage regimen of the therapy comprises a set of dosage instructions associated with the therapy; and wherein the recommended action for the clinical trial and the subject comprises a recommendation to administer the set of dosage instructions during the clinical trial to the subject.

20. The method of any one of claims 1-19, wherein the recommended action for the clinical trial and the subject further comprises a recommendation to administer the set of dosage instructions during the clinical trial to subjects with a body weight that is similar to the body weight of the subject.

21. A system comprising: one or more data processors; and a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform the method of any one of claims 1-20.

22. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform the method of any one of claims 1-20.

Images