Systems and methods for computational drug discovery

Abstract

Systems, methods, and means for computational drug discovery are described. Computational experiments can be created and conducted based on ligand data sets and receptors against which the ligands are to be evaluated. Computational experiments can include pose calculations and free energy calculations from which the mapping and alchemical properties of the ligands can be determined. Various forms of feedback can be provided regarding the status, progress, and results of the computational experiments, and the user can add, modify, or rerun experiments based on the feedback to further explore the affinity and / or selectivity of the ligands.

Description

Technical Field

[0001] The present application relates to systems and methods for computational drug discovery. Background Art

[0002] The traditional drug discovery process relies almost entirely on developing and running assays and performing chemical syntheses in a "wet" laboratory. With advances in computer technology, computational drug discovery techniques have been increasingly adopted in recent years to create "dry" environments where high-throughput screening, compound optimization, and / or binding affinity prediction, etc. can be performed based on computation and / or simulation. However, despite these developments, there is still a lack of turnkey computational drug discovery solutions that incorporate project management, molecular simulation, and data analysis features. Summary of the invention

[0003] Disclosed herein are systems, methods, and means associated with computational drug discovery. According to an embodiment of the present disclosure, a computational drug discovery system or platform can obtain a data set, which can include information related to multiple ligands. The computational drug discovery system can determine the computational experiment to be performed using the data set. The computational experiment can include free energy calculations and posture calculations, wherein the free energy calculations can be designed to determine each binding free energy between a receptor and multiple ligands, and wherein the posture calculations can be designed to determine each mode of binding of multiple ligands to the receptor that results in the binding free energy determined by the free energy calculations. The computational drug discovery system can perform computational experiments by performing at least free energy calculations and posture calculations, and provide feedback related to the computational experiment. Feedback can include a three-dimensional representation of each mode of binding of one or more ligands to the receptor among multiple ligands, i.e., a 3D representation. Feedback can also include an indication of each binding free energy between a receptor and one or more ligands.

[0004] In an example, a data set and a receptor associated with a computational experiment may be received via user input. In an example, after initiating an original run of a computational experiment, a computational drug discovery system may receive an indication of a change to a pose calculation or a free energy calculation (e.g., a change to one or more parameters of a pose calculation or a free energy calculation). In response, the computational drug discovery system may initiate an additional run of the computational experiment based on the indicated change while maintaining the original run of the computational experiment. The computational drug discovery system may provide a graphical representation of the status of the computational experiment, wherein the graphical representation may depict the original run and the additional run of the computational experiment as two bifurcated branches.

[0005] In an example, the free energy calculations described herein may include relative binding free energy calculations, i.e., RBFE calculations, associated with at least a subset of a plurality of ligands, and the feedback described herein may also include a graphical representation of a subset of a plurality of ligands. Such a graphical representation may include nodes and edges, wherein each node may represent a corresponding ligand, and each edge may represent a corresponding RBFE associated with two ligands connected by an edge. In an example, in response to receiving user input indicating a selection of an edge in the graphical representation, the computational drug discovery system may provide a 3D representation of the two ligands connected by the selected edge, wherein the 3D representation may indicate corresponding atoms of the two ligands mapped during the RBFE calculation. In an example, in response to determining that the user has selected corresponding atoms of a plurality of ligands, the computational drug discovery system may establish a mapping between corresponding atoms, and perform an RBFE calculation based on the established mapping.

[0006] In an example, as part of a computational experiment, a computational drug discovery system can simulate each interaction between a receptor and each ligand in a plurality of ligands, and the feedback described herein can also include an indication of each interaction between a receptor and each ligand in a plurality of ligands. For example, feedback can include a comparison between the following two modes: a mode of interaction between a receptor and a first ligand; and a mode of interaction between a receptor and a second ligand. Feedback also includes a heat map that compares each mode of interaction between a receptor and a plurality of ligands. In an example, feedback can also include a molecular dynamics trajectory movie, i.e., an MD trajectory movie, depicting the interaction between a receptor and a ligand. Such a trajectory movie can include multiple frames, wherein each frame can correspond to a step in the simulation and describe the interaction between a receptor and a ligand during the step. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 is a simplified block diagram illustrating an example of a computational drug discovery system.

[0008] FIG. 2A to FIG. 2O Example user interfaces associated with creating, executing, monitoring, and / or managing computational experiments are illustrated.

[0009] Figure 3 An example of feedback associated with pose calculation is illustrated.

[0010] Figure 4 An example of feedback associated with force calculation is illustrated.

[0011] Figure 5 An example of feedback associated with a network map is illustrated.

[0012] FIG. 6A to FIG. 6C An example illustrating feedback associated with free energy calculations.

[0013] FIG. 7A to FIG. 7D Examples of analyses that can be determined based on free energy calculations are illustrated.

[0014] Figure 8 An example process for computational drug discovery is illustrated.

[0015] Fig. 9 Examples of computing devices that may be used to accomplish one or more of the computational drug discovery tasks described herein are illustrated. DETAILED DESCRIPTION

[0016] According to the following disclosure provided by example in conjunction with the accompanying drawings, a more detailed understanding can be obtained. The present disclosure includes various aspects comprising tools, features, examples, models, methods, etc. Many aspects in these aspects are specifically described, and at least in order to illustrate individual characteristics, they are often described in a possibly somewhat restrictive manner. However, this is for the purpose of describing clearly, and the application or scope of these aspects may not be limited. In fact, all different aspects can be combined and interchanged to provide other aspects. In addition, these aspects can also be combined and interchanged with the aspects described in the previous files.

[0017] The aspects described and contemplated in this disclosure may be implemented in many different forms. The accompanying drawings provided herein may provide some examples, but other examples are contemplated. The discussion of the accompanying drawings may not limit the breadth of implementation. These and other aspects may be implemented as methods, devices, and / or computer-readable media (e.g., storage media) including (e.g., stored with) instructions for performing any of the described methods.

[0018] Various methods are described herein, and each method includes one or more than one step or action for implementing the described method. Unless a specific order of steps or actions is required for the correct operation of the method, the order and / or use of specific steps and / or actions can be modified or combined. In addition, terms such as "first", "second", etc. can be used in various examples to modify elements, components, steps, operations, etc., such as, for example, "first computing experiment" and "second computing experiment", etc. Unless specifically required, the use of such terms may not mean the sequencing of modified operations. Therefore, in this example, the first computing experiment does not need to be performed before the second computing experiment, and can, for example, occur before the second computing experiment, during the second computing experiment, or in an overlapping time period with the second computing experiment.

[0019] Various numerical values ​​are used in the examples described in this disclosure. These and other specific values ​​are used for the purpose of describing the examples, and the described aspects are not limited to these specific values. When the accompanying drawings are presented as flow charts, it should be understood that a block diagram of the corresponding equipment is also provided. Similarly, when the accompanying drawings are presented as block diagrams, it should be understood that a flow chart of the corresponding method / processing is also provided.

[0020] The implementations and aspects described herein can be implemented in, for example, methods or processes, devices, software programs, or data streams. Even if only discussed in the context of a single implementation form (e.g., discussed only as a method), the implementation of the discussed features can also be implemented in other forms (e.g., devices or programs). Devices can be implemented in, for example, appropriate hardware, software, and firmware. Methods can be implemented in, for example, a processor, which generally refers to a processing device, which includes, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device. The processor also includes a communication device, such as, for example, a computer, a cellular phone, a portable / personal digital assistant ("PDA"), and other devices that facilitate the communication of information between end users.

[0021] Reference to "one example" or "an example" or "one implementation" or "implementation" and other variations thereof means that a particular feature, structure, characteristic, etc. described in connection with the example is included in at least one example. Thus, appearances of the phrases "in one example" or "in an example" or "in one implementation" or "in an implementation" in various places throughout this disclosure are not necessarily all referring to the same example.

[0022] Additionally, the present disclosure may refer to "determining" various information. Determining information may include, for example, one or more of estimating information, calculating information, predicting information, and retrieving information from a memory. Obtaining may include receiving, retrieving, constructing, generating, and / or determining.

[0023] Additionally, the present disclosure may refer to "accessing" various information. Accessing information may include, for example, one or more of receiving information, retrieving information (e.g., from a memory), storing information, moving information, copying information, calculating information, determining information, predicting information, and estimating information.

[0024] Additionally, the present disclosure may refer to "receiving" various information. Receiving, like "accessing," is intended to be a broad term. Receiving information may include, for example, one or more of accessing information and retrieving information (e.g., from a memory). Furthermore, during operations such as, for example, storing information, processing information, sending information, moving information, copying information, erasing information, calculating information, determining information, predicting information, or estimating information, "receiving" is often involved in one way or another.

[0025] It should be understood that the use of any of, for example, " / ", "and / or", and "at least one of..." in the context of "A / B", "A and / or B", and "at least one of A and B" is intended to cover the selection of only the first listed option (A), or only the second listed option (B), or both options (A and B). As another example, in the context of "A, B, and / or C" and "at least one of A, B, and C", such wording is intended to cover the selection of only the first listed option (A), or only the second listed option (B), or only the third listed option (C), or only the first and second listed options (A and B), or only the first and third listed options (A and C), or only the second and third listed options (B and C), or all three options (A and B and C). As will be apparent to those of ordinary skill in this and related arts, this can be extended to as many items as listed.

[0026] Figure 1 An example of a computational drug discovery system 100 (also referred to herein as a computational drug discovery platform) according to one or more embodiments of the present disclosure is illustrated. Figure 1As shown, the computational drug discovery system may include one or more first computing devices 102 (e.g., one or more web servers), one or more second computing devices 104 (e.g., a server cluster), and / or one or more databases 106, which may be connected via a first communication network 108 (e.g., a private enterprise network) and configured to exchange information through the first communication network. The first computing device 102 may also be configured to interact with a client device 110 through a second communication network 112 (e.g., a public network such as the Internet, etc.) (e.g., through a web browser of a client device), and receive instructions related to one or more drug discovery related services from the client device 110 (e.g., from a user of the client device 110). The first computing device 102 may analyze the received instructions and determine one or more specific computing tasks initiated by the client device 110 based on these instructions (e.g., free energy perturbation (FEP) calculations, molecular dynamics (MD) simulations, prediction calculations for binding affinity, etc.). First computing device 102 may pass these tasks to second computing device 104 for execution, and may track the progress of execution, for example, based on feedback (e.g., computational results, status reports, logs, etc.) provided by second computing device 104 (e.g., directly) and / or information stored in database 106 by second computing device 104. First computing device 102 may provide client device 110 with a means to monitor the status of task execution, and may present results of task execution (such as, for example, binding free energy, protein-ligand interactions, etc.) to client device 104. Results may be presented via various forms of reports and graphical representations, including, for example, three-dimensional (3D) poses, graphs, charts, heat maps, movies, etc.

[0027] Figure 1 The computational drug discovery system shown can accelerate the lead optimization process and the hit-to-lead process by predicting and optimizing ligand binding affinity and selectivity, identifying key residues involved in ligand binding and specific interactions, revealing the structure-activity relationship (SAR) of ligands, and displaying the dynamic behavior of protein-ligand complexes. Compared with traditional drug discovery techniques, Figure 1 The computational drug discovery system shown can allow researchers to identify the most promising candidates for further synthetic work and explore a greatly expanded chemical space. For example, by calculating the free energy difference between two ligands, the computational drug discovery system can identify which modifications are likely to improve or impair binding affinity, allowing researchers to focus their synthetic efforts on the most promising candidates.

[0028] It should be noted that Figure 1The system diagram shown is only one example of how computation-based drug discovery can be performed according to embodiments of the present disclosure. Other system architectures may also be used, including, for example, an architecture in which the client device 110 can interact directly with the second computing device 104 and / or the database 106 without the first computing device 102.

[0029] In some embodiments of the present disclosure, computational experiments may be organized into projects. Projects may be created by users with certain privileges (e.g., system administrators assigned project creation privileges), and additional users (e.g., member users) may be added to projects and given the ability to create, modify, and / or view computational experiments (e.g., also referred to herein as runs) associated with the project. Computational experiments (e.g., FEP runs) may be created based at least on a dataset and a receptor. A dataset may include one or more than one ligand, and each ligand in a dataset may also belong to (one or more than one) other datasets (e.g., there may be a many-to-many relationship between a dataset and a ligand).

[0030] Figure 2A An example user interface (UI) 200a that can be used to create a computational experiment according to an embodiment of the present disclosure is illustrated. The UI 200a can be composed of Figure 1 The (one or more) first computing device 102 or (one or more) second computing device 104 shown is provided and can be Figure 1 The client device 110 shown in FIG. 1 is displayed (for example, in a browser). Figure 2A As shown, UI 200a can be used to create a data set within a project named "DemoProject". The data set can include one or more ligands, which can be added to the data set using a ligand drawing tool provided by the computational drug discovery system described herein or based on a file uploaded by a user. For example, a user can (e.g., using control component 202) upload a .sdf or .smi file from a local or cloud-based storage area, and the computational drug discovery system can analyze the file contents, validate the ligands contained in the file (e.g., to ensure that all structures are valid), and as in Figure 2B As shown on panel 204 of example UI 200b in FIG. 1 , the ligands (including their 2D molecular structures) are displayed to the user. The ligand information determined to be missing during the verification process can be highlighted on panel 204, and a prompt can be displayed to the user to provide the missing information. Upon completion of the verification process, as shown in FIG. Figure 2B As shown, a user may create a data set by giving the data set a name (eg, “Test”) or an ID using controls 206 and clicking a “Create” button 208 .

[0031] Once a data set is created, a user may further create a computational experiment (eg, a run of a computational experiment) based on the data set. Figure 2C An example UI 200c for accomplishing such a task is illustrated. Figure 2C As shown, a user may use UI 200c to specify a receptor (e.g., a protein receptor) that may bind to a ligand of a data set. For example, a user may use text field 210 to provide a name for the receptor, and / or use control 212 to upload a file (e.g., a protein data bank (pdb) file) that may describe the properties of the receptor (e.g., atomic coordinates, disulfide bonds, helices, sheets, active sites, etc.). Upon receiving a successfully uploaded receptor file, such as in Figure 2C As shown on panel 214 of the present invention, the computational drug discovery system described herein can display a 3D structure diagram of a receptor. In the example, UI 200c can also include a control 216 for a user to upload an experimental data file (e.g., an Expdat file) of a ligand and / or receptor, which can include additional information about the ligand and / or receptor obtained through laboratory experiments. UI 200c can also allow the user to specify how the current run of the computational experiment will be performed (e.g., automatically or manually) and / or the tasks to be included in the run (e.g., functional modules). For example, the computational drug discovery system described herein may assign a workflow (e.g., a FEP workflow) to a computational experiment by default, which may include a posture calculation task (e.g., illustrated in the accompanying drawings as "TandemPose"), a force calculation task (e.g., illustrated in the accompanying drawings as "TandemForce"), a network mapping task (e.g., illustrated in the accompanying drawings as "TandemNetMap"), a free energy calculation task (e.g., illustrated in the accompanying drawings as "TandemEnergy"), and an analysis collection task (e.g., illustrated in the accompanying drawings as "TandemAnalytics"). The computational drug discovery system may allow a user to change the default workflow and include only a subset of the above tasks in the computational experiment.

[0032] In an example, UI 200c may include an automatic run option 218, which, when selected, may indicate that the user wishes to make the computational drug discovery system automatically (e.g., sequentially or in parallel) perform all tasks (e.g., functional modules) in the current run. In an example, if the user selects the automatic run option, the computational drug discovery system may further prompt the user to provide a reference ligand, for which experimental data (e.g., experimental structure information) may be known, and other ligands of the data set may be docked against the reference ligand (e.g., for relative binding free energy (RBFE) determination, for generating reasonable binding modes of other ligands, etc.). The user may choose not to use the automatic run option, but to manually run one or more than one of the tasks described herein. In this case, the computational drug discovery system may allow the user to specify a reference ligand for (one or more than one) specific task selected by the user.

[0033] The user can, for example, use Figure 2C The text field 220 shown gives the run a unique name to distinguish the new run of the computational experiment from other runs, and can be changed by clicking Figure 2C 222 to start a new run. This can trigger a computational drug discovery system (e.g., hosted on Figure 1The computational engine of the system on the (one or more than one) second computing device 104) performs a computational experiment by performing tasks associated with the operation (e.g., all or a subset of the tasks described herein). For example, the computational drug discovery system can perform the free energy calculation tasks described herein by determining the respective binding free energies between the receptor and multiple ligands associated with the operation, and perform the posture calculation tasks described herein by determining the respective ways in which the ligand can bind to the receptor that result in the binding free energy determined by the free energy calculation tasks. The ways of ligand-receptor binding can include the respective positions and / or orientations of the ligand relative to the receptor at the corresponding binding site or binding pocket. The binding free energies described herein may include absolute binding free energies (ABFEs) or relative binding free energies (RBFEs) that can indicate the respective affinities of the ligand to the receptor in absolute or relative terms (e.g., the RBFE can indicate the difference between the binding free energies of two compounds). Various techniques can be used to determine the binding free energy and / or binding position between the ligand and the receptor. For example, molecular docking techniques can be used to determine the binding site between a ligand and a receptor, and then the corresponding relative binding free energy can be determined by calculating the free energy change associated with the alchemical transformation of one ligand into another ligand using molecular dynamics simulations and free energy perturbation theory or thermodynamic integration (TI) methods. The binding free energy between a ligand and a receptor can also be determined using machine learning models (such as deep learning or regression models, etc.), which can be trained on molecular features and experimental binding data to predict binding free energy.

[0034] The computational drug discovery system can perform the force calculation task described herein (called "TandemForce") by generating a force field (e.g., a small molecule force field) for the ligands included in the data set, and comparing the fit between the potential energy dihedral angle scans predicted by the quantum mechanical predictions for each dihedral angle in the data set. The computational drug discovery system can perform the net (network) mapping task described herein by mapping the corresponding atoms of two or more ligands involved in the free energy calculation (e.g., RBFE calculation) and generating an alchemical network for user visualization and editing mapping. In addition, the computational drug discovery system can perform the analysis task described herein by analyzing the calculation and / or simulation results from the computational experiment and presenting the analysis with a series of data rankings and graphical representations, including, for example, a dynamic 3D representation of the mode of binding of one or more ligands to the receptor, a table / list / chart showing each binding free energy between a receptor and one or more ligands, a trajectory movie showing the interaction between a receptor and one or more ligands, a comparison of the mode of interaction between a receptor and one or more ligands, etc.

[0035] The tasks (or modules) associated with the operation of the computational experiment can form a workflow for the operation. Tasks (or modules) can be connected in a coherent manner, and the results of a task (or module) can be used in the next one. When the task is being performed and / or completed, (e.g., via the labels shown in some of the accompanying drawings) feedback can be provided to indicate the current state of the workflow (e.g., status and / or results can be displayed when they become available). For example, after the user creates and starts the operation, the computational drug discovery system described herein can allow the user to monitor the progress and / or results of the operation. Figure 2D An example of monitoring the progress of an operation according to an embodiment of the present disclosure is illustrated. Figure 2D As shown, while a run is in progress, the computational drug discovery system can provide a real-time status indicator of the run, which can be (e.g., Figure 2D 224) indicates the ligands submitted to the various calculations to be run, and (e.g., via Figure 2D 226) of the computations that are running, queued, or finished. Once the run is complete, e.g. Figure 2E As illustrated by the panel 228 and progress bar 230, the computational drug discovery system may provide an indication of the complete status. Figure 2E The computational drug discovery system can provide a running log, the contents of which can be as follows: Figure 2F exemplified.

[0036] The computational drug discovery system may allow a user to adjust the parameters of an ongoing run of a computational experiment and create another run of the computational experiment using the same data set. Figure 2E 234) indicates that a parameter of the ongoing run is desired to be changed, the computational drug discovery system may open a new panel (e.g., Figure 2G As another example, in response to a user (e.g., via Figure 2E If a control 236 is displayed indicating that a change is desired to the ongoing run, the computational drug discovery system may provide a graphical representation of the current status of the ongoing run and allow the user to edit task(s) associated with the ongoing run. Figure 2H Example response to user activation Figure 2E Example of changing the operation by clicking the "Change Operation" control 236. Figure 2HAs shown, in response to the "change run" control being activated, the computational drug discovery system can summarize the tasks of the ongoing run (e.g., posture, force, mapping, energy, etc.) via a graph 238, in which each task can be represented as a node, and the relationship between two sequential tasks can be represented by an edge. The user can indicate the task to be changed, for example, by selecting a corresponding node in the graph or by activating a corresponding label 240 in the lower panel. Upon receiving such an indication, the computational drug discovery system can (e.g., in a panel 242) display the parameters associated with the selected task and allow the user to change these parameters.

[0037] Once a user (for example, via Figure 2G The "Edit Parameters" action illustrated or Figure 2H The computational drug discovery system may submit new or updated parameters for an ongoing run (e.g., a "change run" operation illustrated in the example), and if the computational drug discovery system determines that the new or changed parameters may cause the ongoing run (e.g., for the same data set and receptor) to generate different results at one or more stages of the run, the computational drug discovery system may create a new run based on the new or changed parameters, and may associate the new run and the ongoing run (e.g., also referred to as the original run) as two related forked tasks or operations. Fig.2I As illustrated, the computational drug discovery system may present newly created runs and ongoing runs as separate branches of a graph. Fig.2I As shown, the computational drug discovery system can organize and present (e.g., related runs of computational experiments for the same dataset and receptor) via a multi-level structure 244, in which the levels can correspond to project->dataset->receptor->run. The computational drug discovery system can also provide a graph 246 in which receptors (e.g., "THROMBIN") and tasks (e.g., "TandemPose", "TandemForce", etc.) can be represented as nodes, and the sequential relationships between the nodes can be represented by edges. For example, two runs "thrombin_allsdf_auto" and "run2" associated with the dataset "thrombin_all" and the receptor "THROMBIN" can be represented as a bifurcated network that shows where the runs branch (e.g., due to submitting new tasks with different parameters), which may affect the results of one run relative to another run. As shown in FIG. Fig.2I As illustrated by the example in , the graph (or bifurcated network) 246 may indicate that the user re-ran “TandemPose” with different parameters as part of “run2” related to another run “thrombin_allsdf_auto”. Figure 2JAnother example is shown as follows: a graphic (e.g., a run explorer or fork network) 248 may be provided to indicate that multiple pose, force, and energy calculations are created and started for the dataset "Test-CC-TYK-3_70" and the receptor "TYK2". Figure 2J 2. In the illustrated interface, a user may select a run (eg, in the left pane 250) and the corresponding branch is highlighted in the graph 248 to visualize where the selected run branches off from the other runs.

[0038] The computational drug discovery system can allow runs to fork at any point in the workflow, and runs for the same dataset and receptor in a forked network can be grouped. For example, a run can fork in response to a user creating a new run, or starting a new task (e.g., gesture or energy calculation) for an existing run (in which a similar task may already be executing), (e.g., if a run already has a gesture task and the user starts a new gesture task, this can create a new forked run). A run can also fork based on a user canceling a task, reinitializing a task, editing the alchemical map of the current run, etc.

[0039] Figure 2K to Figure 2O Examples of tools (eg, graphical user interfaces) that may be used to monitor and / or manage projects, data sets, or computational experiments after they have been created are illustrated. Figure 2K The following example user interface is illustrated, using which a user can select a project (e.g., "DemoProject") from a project list 252 (e.g., a drop-down box) and obtain an overview of the basic information of the project (e.g., members, creation time, etc.) and the number of datasets and / or ligands included in the project. Figure 2L An example user interface is illustrated below, using which a user can select a particular dataset ("BD_TYK2_DEMO") and view the ligands included in the dataset and the two-dimensional (2D) molecular structures of these ligands (e.g., as shown in panel 254). Figure 2M An example user interface is illustrated below, with which a user can download a source data file associated with a particular ligand, for example, by clicking control 256 for the particular ligand. Figure 2N An example user interface is illustrated below, using which a user can obtain a summary of the datasets / ligands included in a project, and / or computational experiments or (one or more) runs (e.g., "TYK2 (3 runs)") associated with a particular dataset (e.g., "TYK2_DEMO_BD") (as shown in 258). Additionally, in response to a user selecting (one or more) runs associated with a particular dataset, as shown in 258. Fig.2O As illustrated, the computational drug discovery system may provide a summary of the run that includes various calculations 260 (e.g., poses, forces, energies, etc.) performed on a data set of ligands (e.g., Fig.2O Panel 262 shows the ligands for which calculation 260 was completed.) The user can switch between runs associated with a particular data set, for example, using a drop-down box 264 listing all runs.

[0040] As described herein, a computational experiment created and started by a user (e.g., run) can be performed using the disclosed computational drug discovery system by performing tasks or calculations configured by the user. Various feedback related to the results of the computational experiment can be provided during and / or after the computational experiment. For example, as determined by the posture calculations described herein, feedback can include a 3D representation of the way in which multiple ligands bind to a receptor. As another example, as determined by the energy calculations described herein, feedback can include a plot and / or list indicating each binding free energy of multiple ligands to a receptor. As yet another example, feedback can include a representation of an interaction (e.g., a hydrophobic interaction) between a ligand and a receptor. Through feedback, users can predict and optimize ligand binding affinity and selectivity, identify key residues involved in ligand binding and specific interactions, explore the structure-activity relationship (SAR) of the ligand, and study the dynamic behavior of the protein-ligand complex.

[0041] Feedback for a computational experiment may be provided separately for each task associated with the computational experiment (eg, posture, force, net mapping, energy, or analysis). Figure 3 Examples of feedback associated with pose calculations that may be performed by the disclosed computational drug discovery system (e.g., as part of a computational experiment or run) are illustrated. As previously described, such pose calculations may determine individual binding positions of multiple ligands to a receptor at one or more binding sites / pockets, as well as individual interactions between the ligands and the receptor (e.g., hydrogen bonds, halogen bonds, water bridges, salt bridges, hydrophobic contacts, Pi stacking, and / or cation-Pi interactions, etc.). Figure 3As shown, the results of the pose calculation can be visualized by a 3D viewer 302, which can be moved, rotated and / or zoomed in / out to examine the ligand-receptor binding from different angles and / or at different levels of detail. By default, a reference ligand can be loaded with the receptor, but any other ligand in the data set can also be added to the 3D viewer, for example, by selecting the ligand in the left panel 304. And once the ligand is added to the 3D viewer, the user can control what information related to the ligand is displayed in the 3D viewer. For example, a control for switching the visibility of the ligand interaction in the 3D viewer on and off can be provided on the panel 304 for the user. Each ligand tile in the panel 304 can also display basic information related to the ligand (e.g., ligand name) and certain operating results of the ligand (e.g., pose (P) and / or force (F) results). A status icon can be provided for each ligand in the corresponding tile as an indicator of whether the pose calculation has been completed for the ligand. In an example, a user can download results for a ligand and / or modify a ligand (e.g., a non-reference ligand). For example, a user can upload a new file for a ligand (e.g., to replace the original ligand pose), and the ligand file can be used in subsequent calculations / tasks. A user can also upload a ligand file for a ligand for which pose calculations failed, and in response, the computational drug discovery system can redetermine the ligand-receptor interaction for the ligand based on the uploaded file. A user can adjust the display in the 3D viewer 302 by adjusting the viewer settings 306.

[0042] Figure 4 An example of feedback associated with a custom small molecule force field parameterization calculation performed by a computational drug discovery system described herein (e.g., as part of a computational experiment) is illustrated. As described herein, a force field calculation can generate a small molecule force field and compare the fit between potential energy dihedral angle scans as determined by quantum mechanical (QM) calculations, TandemForce, and GAFF2 for each dihedral angle in a data set. Figure 4 As shown, the fragment dihedral angles found in the ligand can be indicated in table 402, which can also provide a summary of all active dihedral angle potential scans, thereby comparing the dihedral angle scan potentials determined by TandemForce, GAFF2, and QM calculations. The root mean square error (RMSE) can be reported for each comparison, where smaller values ​​indicate a higher degree of agreement with the QM prediction, and Figure 4 The plot 404 to the right of may indicate the degree of agreement between the potential energy scan (PES) derived from the TandemForce calculations and the QM PES.

[0043] Figure 5Examples of feedback associated with ligand mapping (also referred to herein as network mapping or NetMap) performed by a computational drug discovery system described herein (e.g., as part of a computational experiment) are illustrated. Mappings (e.g., of corresponding atoms between ligands) can be automatically generated in response to detecting similar chemical and / or structural properties between ligands, or manually generated based on user input connecting two or more of the ligands together (e.g., the computational drug discovery system can first present the ligands as disconnected, and then allow the user to connect a pair of ligands by selecting one ligand and then selecting another ligand). In either scenario, as described herein, a parameter (e.g., connectivity_number) indicating the number of mappings or connections (e.g., 3, 4, etc.) from the ligands can be provided to guide the generation of a mapping network, and the generated mapping can be used to perform free energy calculations (e.g., RBFE calculations). In an example, the computational drug discovery system can be configured to display the binding free energies of one or more ligands with reference to mappings between the ligands. For example, the computational drug discovery system can determine the relative binding free energy (RBFE) between one or more mapped ligand pairs, and via Figure 5 The graph 502 (e.g., a mapping network) shown provides a representation of the determined RBFE along with the mapping. Each node in the graph can represent a corresponding ligand, and each edge (e.g., a forward edge or a reverse edge) can represent the mapping and / or RBFE of two ligands connected by the edge (e.g., in the forward direction or the reverse direction).

[0044] In addition to mapping network 502, the computational drug discovery system can also display a list of ligands included in the data set (e.g., in panel 504), and / or allow the user to use the UI tools provided on panel 506 to check, add and change the edges in the mapping network 502. For example, in response to the user selecting a ligand from panel 504, the computational drug discovery system can highlight and / or amplify the ligand in the mapping network 502. In response to the user selecting an edge (e.g., edge 508) from the mapping network 502, the computational drug discovery system can display a list of mapped atoms from two ligands connected by the selected edge in region 506a, and / or display a 3D representation of the two ligands in region 506b, wherein the 3D representation can indicate the corresponding atoms of the two ligands mapped, and / or unmapped atoms. The computational drug discovery system can further allow the user to, for example, add or edit the atomic mapping between the two ligands by selecting and mapping the additional atoms of the two ligands via control 506c. The computational drug discovery system can also allow the user to delete an edge or add a new edge (e.g., this may trigger a new free energy run).

[0045] Fig. 6A , Figure 6B and Figure 6C Examples of feedback associated with free energy calculations performed by a computational drug discovery system (e.g., as part of a computational experiment) are illustrated. As shown in these figures, free energy feedback can include Figure 5 The same mapped network and ligand panels as shown in Fig. 6A The mapping network 602 and ligand panel 604 may also include plots, charts, and / or tables summarizing various alchemical properties of the ligands in the mapping network 602. Fig. 6A As shown, these plots, charts, and / or tables can be displayed on panel 606 in response to a user selecting an edge (e.g., edge 608) from the mapped network, and the displayed information can include free energies (e.g., RBFE) and / or other alchemical properties of ligands connected through the selected edge in a forward direction and / or a reverse direction (e.g., over time, in adjacent time windows, etc.). The user can edit the parameters associated with the free energy calculation (e.g., by clicking on an "edit parameters" control 610) and cause the calculation to be rerun for at least the selected edge.

[0046] like Figure 6B and Figure 6C As shown, feedback from the free energy calculation can be presented in the form of (one or more than one) of a chart, plot, list, table, ranking, 3D model, etc. For example, the feedback can include a list 612 of ligands for which free energy calculations were performed, which can be ranked or sorted based on various data fields or criteria, such as ΔG values ​​derived from energy calculations, ΔG values ​​derived from experimental files, similarity scores (e.g., compared to the currently selected left ligand), etc. As another example, if Figure 6B As shown, the analysis may include a plot 614 of ΔG values ​​over a simulation time period in the forward and reverse directions, and a bar chart 616 of pairs of ΔG values ​​and λ values. As yet another example, Figure 6C As shown, the analysis may include plot 618 and plot 620, which indicate the degree of overlap and convergence of the free energy calculations.

[0047] Fig. 7A , Figure 7B and Figure 7CExamples of analyses that can be provided by the computational drug discovery systems described herein are illustrated. For example, as part of a computational experiment, the computational drug discovery system can simulate the interactions between a receptor and a plurality of ligands associated with the computational experiment, and provide analyses indicating the ligand-receptor interactions determined during the simulation. The computational drug discovery system can determine the ligand-receptor interactions based on the chemical groups on the receptor side and the ligand side and their various orientations and distances at each step of the simulation. For example, if a receptor has a positively charged group that is close enough to a negatively charged atom on a ligand, an electrostatic interaction (e.g., which can be referred to as a "salt bridge") can be inferred. The computational drug discovery system can present the determined ligand-receptor interactions in various ways so that users of the system can view the molecular interaction curves of the receptors and ligands involved in the computational experiment. For example, as Fig. 7A As shown, the computational drug discovery system can provide a "cross-ligand" view 702 of each interaction between the receptor and each ligand (e.g., jmc_30, jmc_23, etc.) relative to one or more residues (e.g., Arg:140, Asn:141, Leu:143, etc.), where different shading can indicate the respective frequencies at which the interaction was observed during the simulation. As another example, Figure 7B As illustrated, the computational drug discovery system can allow a user to compare patterns of interactions occurring between a receptor and a first ligand (e.g., jmc_28) during a simulation with patterns of interactions occurring between a receptor and a second ligand (e.g., ejm_45) during a simulation. These patterns can, for example, indicate where the interactions occurred (e.g., with which residues), the nature of the interactions (e.g., hydrophobic, H-bond, water, etc.), and / or the frequency with which the interactions were observed during the simulation (e.g., the intensity of heatmaps 704a and 704b indicates the proportion of frames in the trajectory where the interactions were sampled).

[0048] like Figure 7C As illustrated, a user may choose to switch to a "dihedral angle profile" view, and in response, the computational drug discovery system may present plots depicting dihedral angles sampled during a trajectory associated with a selected ligand. The lines in these plots may show the potential energy of each dihedral angle, while the bars in the plots may represent counts of frames in which the dihedral angles are within a particular angle range among all frames of the trajectory file.

[0049] like Fig.7DAs shown, the computational drug discovery system can further provide a 3D viewer 706 for viewing the molecular dynamics simulation (MD) trajectory of the ligand (e.g., ejm_45). The trajectory can be presented as an MD movie including a plurality of frames, each of which can correspond to a step of the simulation and indicate the interaction between the ligand and the receptor at that step (e.g., taking into account the corresponding orientation and distance of the ligand and the receptor). The user can use Fig.7D Controls are shown to pause / play and / or skip different frames to visualize the interaction between ligand and receptor at different binding sites.

[0050] Figure 8 An example process 800 for conducting a computational experiment according to an embodiment of the present disclosure is illustrated. Figure 8 As shown, the process may include: obtaining a data set at 802, wherein the data set may include information related to multiple ligands that a user may want to evaluate. The process may also include: determining a computational experiment to be performed using the data set of ligands at 804, wherein the computational experiment may include free energy calculations and pose calculations. The free energy calculation may be designed to determine the various binding free energies between the receptor and the multiple ligands in the data set, and the pose calculation may be designed to determine the various ways in which the multiple ligands bind to the receptor (which results in the binding free energy determined by the free energy calculation). The process may also include: performing a computational experiment by performing at least a free energy calculation and a pose calculation at 806, and providing feedback related to the computational experiment at 808, wherein the feedback may include a 3D representation of the various ways in which one or more of the multiple ligands bind to the receptor, and an indication of the various binding free energies between the receptor and the one or more ligands.

[0051] Fig. 9 The example computing device 900 (eg, Figure 1An example of a first or second computing device in a drug discovery system 100. As shown, the computing device 900 may include a processor (e.g., one or more processors) 902, which may be a central processing unit (CPU), a graphics processing unit (GPU), a microcontroller, a reduced instruction set computer (RISC) processor, an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a physical processing unit (PPU), a digital signal processor (DSP), a field programmable gate array (FPGA), or any other circuit or processor capable of performing the tasks described herein. The computing device 900 may also include a communication circuit 904, a memory 906, a mass storage device 908, an input device 910, and / or a communication link 912 (e.g., a communication bus), through which one or more components shown in the figure can exchange information.

[0052] The communication circuit 904 may be configured to send and receive information using one or more communication protocols (e.g., TCP / IP) and one or more communication networks (including a local area network (LAN), a wide area network (WAN), the Internet, a wireless data network (e.g., Wi-Fi, 3G, 4G / LTE, or 5G network)). The memory 906 may include a storage medium (e.g., a non-transitory storage medium) configured to store machine-readable instructions that, when executed, cause the processor 902 to perform one or more of the tasks described herein. Examples of machine-readable media may include volatile or non-volatile memory, including, but not limited to, semiconductor memory (e.g., electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM)) and / or flash memory, etc. The mass storage device 908 may include one or more disks, such as one or more internal hard disks, one or more removable disks, one or more magneto-optical disks, one or more CD-ROM or DVD-ROM disks, etc., on which instructions and / or data may be stored to facilitate the operation of the processor 902. The input device 910 may include a keyboard, a mouse, a voice-controlled input device, and / or a touch-sensitive input device (e.g., a touch screen), etc., for receiving user input to the computing device 900.

[0053] It should be noted that computing device 900 may operate as a standalone device, or may be connected (eg, networked or clustered) with other devices to perform the tasks described herein. Fig. 9 Only one instance of each component is shown in FIG. 9 , but those skilled in the art will appreciate that computing device 900 may include multiple instances of one or more than one of the components shown in the figure.

[0054] CROSS-REFERENCE TO RELATED APPLICATIONS

[0055] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 448,387, filed on February 27, 2023, the disclosure of which is incorporated herein by reference in its entirety.

Claims

1. A method for computational drug discovery, the method comprising: obtaining a data set, wherein the data set includes information related to a plurality of ligands; determining a computational experiment to be performed using the data set, wherein the computational experiment comprises a free energy calculation and a pose calculation, wherein the free energy calculation is used to determine respective binding free energies between the receptor and the plurality of ligands, and wherein the pose calculation is used to determine respective modes of binding of the plurality of ligands to the receptor that result in the binding free energies; performing said computational experiment by performing at least said free energy calculation and said pose calculation; and providing feedback related to the computational experiment, wherein the feedback includes a three-dimensional representation (3D representation) of each manner in which one or more of the plurality of ligands bind to the receptor, and the feedback also includes an indication of each binding free energy between the receptor and one or more of the plurality of ligands; The method further includes: after initiating the original run of the computational experiment, receiving an indication of a change to the computational experiment, and while maintaining the original run of the computational experiment, initiating an additional run of the computational experiment based on the change, and associating the additional run and the original run as related forked tasks or operations.

2. The method according to claim 1, further comprising: One or more user inputs are received indicating the receptor and the data set associated with the computational experiment.

3. The method according to claim 1, further comprising: A graphical representation of the status of the computational experiment is provided, wherein the graphical representation depicts the original run and the additional run of the computational experiment as two diverging branches.

4. The method according to claim 1, wherein: The change to the pose calculation or the free energy calculation includes a change to a parameter of the pose calculation or the free energy calculation.

5. The method according to claim 1, wherein: The free energy calculation includes a relative binding free energy calculation, i.e., an RBFE calculation, associated with at least a subset of the multiple ligands, and wherein the feedback also includes a graphical representation of the subset of the multiple ligands, the graphical representation comprising nodes and edges, each node representing a corresponding ligand, and each edge representing a corresponding RBFE associated with two ligands connected by the edge.

6. The method according to claim 5, further comprising: In response to receiving user input indicating a selection of an edge in the graphical representation, providing a 3D representation of two ligands connected by the selected edge, wherein the 3D representation indicates corresponding atoms of the two ligands mapped during the RBFE calculation.

7. The method according to claim 5, further comprising: It is determined that corresponding atoms of two or more ligands have been selected, and a mapping between the corresponding atoms is established, wherein the RBFE calculation is performed based on the established mapping.

8. The method according to claim 1, further comprising: Each interaction between the receptor and each ligand of the plurality of ligands is simulated as part of the computational experiment, wherein the feedback further includes an indication of each interaction between the receptor and each ligand of the plurality of ligands.

9. The method according to claim 8, wherein: The feedback includes a comparison between the pattern of interactions occurring between the receptor and a first ligand and the pattern of interactions occurring between the receptor and a second ligand.

10. The method according to claim 8, wherein: The feedback includes one or more heat maps that compare patterns of interactions occurring between the receptor and each of the plurality of ligands.

11. The method according to claim 8, wherein: For each ligand in the plurality of ligands, the feedback comprises a molecular dynamics trajectory movie, ie, an MD trajectory movie, depicting the interaction between the receptor and the ligand.

12. The method according to claim 11, wherein: The MD trajectory movie includes a plurality of frames, each frame corresponding to a step in the simulation and depicting the interaction between the receptor and the ligand during the step.

13. A system for computational drug discovery, comprising a first set of one or more computing devices and a second set of one or more computing devices, in, The one or more computing devices of the first set are configured to: obtaining a data set comprising information related to a plurality of ligands; as well as determining a computational experiment to be performed using the data set, wherein the computational experiment comprises a free energy calculation and a pose calculation, wherein the free energy calculation is used to determine respective binding free energies between the receptor and the plurality of ligands, and wherein the pose calculation is used to determine respective modes of binding of the plurality of ligands to the receptor that result in the binding free energies; and The one or more computing devices of the second set are configured to: performing said computational experiment by performing at least said posture calculation and said free energy calculation; and providing feedback related to the computational experiment, wherein the feedback includes a three-dimensional representation (3D representation) of each manner in which one or more of the plurality of ligands bind to the receptor, and the feedback also includes an indication of each binding free energy between the receptor and one or more of the plurality of ligands; Wherein, after initiating the original run of the computing experiment, one or more computing devices of the second set are also configured to receive indications of changes to the computing experiment, and while maintaining the original run of the computing experiment, initiate additional runs of the computing experiment based on the changes, and associate the additional runs and the original run as related forked tasks or operations.

14. The system according to claim 13, wherein: The one or more computing devices of the first set are also configured to receive one or more user inputs indicative of the receptor and the data set associated with the computational experiment.

15. The system of claim 13, wherein: The one or more computing devices of the second set are also configured to provide a representation of a status of the computing experiment, the representation depicting the original run and the additional run of the computing experiment as two diverging branches.

16. The system of claim 13, wherein: The free energy calculation includes a relative binding free energy calculation, i.e., an RBFE calculation, associated with at least a subset of the multiple ligands, and wherein the feedback provided by one or more computing devices of the second set also includes a graphical representation of the subset of the multiple ligands, the graphical representation comprising nodes and edges, each node representing a corresponding ligand, and each edge representing a corresponding RBFE associated with two ligands connected by the edge.

17. The system of claim 16, wherein: The one or more computing devices of the second set are further configured to determine that corresponding atoms of two or more ligands have been selected, establish a mapping between the corresponding atoms, and perform the RBFE calculation based on the established mapping.

18. The system of claim 13, wherein: The one or more computing devices of the second set are also configured to simulate individual interactions between the receptor and each of the plurality of ligands as part of the computational experiment, and wherein the feedback also includes an indication of individual interactions between the receptor and each of the plurality of ligands.

19. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 12.

Images