A method for calculating small molecule transmembrane permeation free energy and a method for establishing a data set

Abstract

This invention provides a method for calculating the transmembrane permeation free energy of small molecules and a method for establishing a dataset, belonging to the field of free energy calculation technology. The method for calculating the transmembrane permeation free energy of small molecules uses a molecular bond network algorithm to coarsely parameterize small molecules, constructs a water-membrane system, and utilizes molecular dynamics simulation software to perform tensile dynamics simulation and umbrella sampling simulation, thereby obtaining the free energy change curve of the compound's transmembrane permeation process and the water-membrane distribution free energy. This not only significantly saves experimental time but also generates a large batch of high-quality transmembrane permeation capacity data. The method for establishing the small molecule transmembrane permeation free energy dataset uses chemical feature space to screen molecules, making the dataset widely applicable and highly representative. Drug transmembrane free energy prediction models trained using this dataset can more accurately predict the transmembrane capacity of drugs, providing a scientific basis for drug development.

Description

Technical Field

[0001] This invention relates to a method for calculating the transmembrane permeation free energy of small molecules and a method for establishing a dataset, belonging to the field of free energy calculation technology. Background Technology

[0002] In the field of drug development, the transmembrane permeability of small molecule compounds is one of the key factors in assessing their drug potential. The human cell membrane, as the basic structural unit of organisms, not only maintains the stability of the intracellular environment but also regulates the process of substances entering and leaving the cell. Therefore, small molecule compounds need to penetrate at least one cell membrane in the human body to reach their target site and exert a therapeutic effect. However, accurately and efficiently assessing the transmembrane permeability of small molecule compounds is a crucial factor in determining their drug potential.

[0003] Currently, the determination of the transmembrane permeability of small molecule compounds mainly relies on experimental methods. These methods typically assess transmembrane capacity by simulating a water-membrane environment in vitro and measuring the distribution of the drug across the membrane after a certain period of time. Although these experimental methods can provide relatively accurate data, they have significant limitations: they are time-consuming, costly, produce limited data, and have poor reproducibility.

[0004] With the rapid development of computer technology and computational chemistry, computational simulation-based methods are being increasingly widely used in drug development. Computer simulations allow for in-depth research into the interaction mechanisms between small molecules and cell membranes at the molecular level, enabling accurate prediction of their transmembrane permeability. This method not only saves time and costs but also generates large volumes of high-quality transmembrane permeability data, providing strong support for drug development.

[0005] However, existing computational simulation methods still have some limitations. For example, while traditional all-atom simulations offer high accuracy, they are computationally intensive and difficult to apply to screening large batches of compounds. Coarse-grained simulation methods, although significantly reducing computational load, still require improvement in accuracy and reliability when dealing with complex molecular systems. Therefore, developing a computational simulation method that is both efficient and accurate is crucial for advancing drug development. Summary of the Invention

[0006] The purpose of this invention is to provide a method for calculating the transmembrane permeation free energy of small molecules and a method for establishing a dataset. This method not only significantly saves experimental time but also generates a large amount of high-quality transmembrane permeation capacity data. At the same time, a corresponding dataset is established. The drug transmembrane free energy prediction model trained on this dataset can more accurately predict the transmembrane capacity of drugs, providing a scientific basis for drug development.

[0007] To achieve the above objectives, the present invention employs the following technical solution:

[0008] Firstly, a method for calculating the transmembrane permeation free energy of small molecules is provided, including the following steps:

[0009] Obtain the SMILES format description of the target small molecule;

[0010] A graph-based molecular bond network algorithm is used to perform coarse-grained parameterization of molecules to generate coarse-grained molecular models.

[0011] A coarse-grained water-film system was constructed, and the coarse-grained molecular model was dynamically assembled with the water-film system to position the molecules at a predetermined location.

[0012] Perform a stretching dynamics simulation to pull small molecules across the membrane using a center-of-mass traction method, generating transmembrane reaction coordinate trajectories;

[0013] Sampling windows are drawn at fixed intervals along the reaction coordinate trajectory to simulate umbrella-shaped sampling;

[0014] The WHAM algorithm was used to perform weighted histogram analysis on the umbrella-shaped sampling results, and the transmembrane free energy variation curve and water-membrane partition free energy were output.

[0015] Preferably, the molecular bond network algorithm includes:

[0016] Map atoms onto coarse-grained beads;

[0017] Parameterizing nonbonded interactions between atoms;

[0018] Parameterizing interatomic bonding interactions;

[0019] Initial three-dimensional molecular structures were generated and optimized based on the MMFF94 force field.

[0020] Generate molecular structure and parameter files based on atom-bead mapping.

[0021] Preferably, the dynamic assembly includes:

[0022] Define the small molecule simulation box and center;

[0023] Small molecules are inserted 1 nm above the membrane system and replace the water molecules at that position;

[0024] Calculate the number of water molecules replaced and add 0.1M NaCl solution;

[0025] Energy optimization and multi-stage balance simulation are performed.

[0026] Preferably, the equilibrium simulation process uses a v-rescale temperature coupling algorithm to control the system equilibrium temperature to 303.15K, and uses the Berendsen method to control the system pressure to 1 bar.

[0027] Preferably, the specific method for performing the umbrella-shaped sampling simulation is as follows:

[0028] The sampling window is sampled along the Z-axis at 0.1 nm intervals;

[0029] Each system has 60 windows, and each window has a simulation duration of 100 ns.

[0030] Secondly, a method for establishing a dataset of small molecule transmembrane permeation free energy is provided, including:

[0031] Neutral small molecules were screened from the database, and their molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar surface area were extracted to form a high-dimensional feature space.

[0032] The t-SNE algorithm is used to reduce the dimensionality of the high-dimensional feature space to a two-dimensional plane;

[0033] Randomly select a preset number of small molecules until the distribution uniformity condition is met;

[0034] The transmembrane permeation free energy of the selected small molecule is calculated using the aforementioned method for calculating the transmembrane permeation free energy.

[0035] By correlating the molecular SMILES format description with the calculated transmembrane permeation free energy, a structured dataset is generated.

[0036] Preferably, the distribution uniformity condition is: the difference in two-dimensional spatial distribution density is less than or equal to 20% and the overall coverage is greater than or equal to 95%.

[0037] Thirdly, a system for calculating the transmembrane permeation free energy of small molecules is provided, including:

[0038] Molecular coarsening module: Configured to receive SMILES input, and use a graph-based molecular bond network algorithm to coarse-grained parameterize the molecule to generate a coarse-grained molecular model;

[0039] Membrane environment construction module: configured to generate a coarse-grained water-membrane system;

[0040] Dynamic assembly engine: configured to dynamically assemble coarse-grained molecular models with water-film systems, positioning molecules to preset positions;

[0041] Transmembrane trajectory generator: configured to drive molecules to move along the membrane normal according to the stretching kinetics model, generating transmembrane reaction coordinate trajectories;

[0042] Free energy calculation unit: configured to extract sampling windows at fixed intervals along the reaction coordinate trajectory, perform umbrella-shaped sampling simulation, and output free energy curves and water-film distribution free energy through the WHAM algorithm.

[0043] Fourthly, a system for establishing a dataset of small molecule transmembrane permeation free energy is provided, including:

[0044] Chemical feature extractor: configured to screen neutral small molecules and extract molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar surface area to form a high-dimensional feature space;

[0045] Feature space dimensionality reducer: configured to use the t-SNE algorithm to map high-dimensional features to a two-dimensional plane;

[0046] Molecular screening engine: Configured to randomly select a preset number of small molecules until the distribution uniformity condition is met;

[0047] Data synthesis terminal: configured to calculate transmembrane permeation free energy using the small molecule transmembrane permeation free energy calculation method, and associate the molecule SMILES with the transmembrane free energy calculation results to generate a structured dataset.

[0048] Fifthly, a device for calculating the transmembrane permeation free energy of small molecules is provided, comprising a processor and a memory storing program instructions, wherein the processor is configured to execute the method for calculating the transmembrane permeation free energy of small molecules when running the program instructions.

[0049] The advantages of this invention are:

[0050] This invention proposes a method for calculating the transmembrane permeation free energy of small molecules. This method uses a molecular bond network algorithm to coarsely parameterize small molecules, constructs a water-membrane system, and utilizes molecular dynamics simulation software to perform tensile dynamics simulation and umbrella sampling simulation, thereby obtaining the free energy change curve of the compound's permeation process and the water-membrane partition free energy. This method not only significantly saves experimental time but also generates large amounts of high-quality transmembrane permeation capacity data, providing strong support for drug development.

[0051] Furthermore, this invention also relates to a method for establishing a small molecule transmembrane permeation free energy dataset. By screening molecules through a chemical feature space, the dataset has a wide applicability and strong representativeness. A drug transmembrane free energy prediction model trained using this dataset can more accurately predict the transmembrane capacity of drugs, providing a scientific basis for drug development.

[0052] In summary, the method for calculating the transmembrane permeation free energy of small molecules based on the molecular bond network algorithm and the method for establishing the dataset proposed in this invention have broad application prospects and important scientific value in the field of drug development. Attached Figure Description

[0053] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0054] Figure 1 This is a flowchart illustrating the method for calculating the transmembrane permeation free energy of small molecules according to an embodiment of this application.

[0055] Figure 2 This is a flowchart illustrating the method for establishing a small molecule transmembrane permeation free energy dataset according to an embodiment of this application.

[0056] Figure 3 This is a schematic diagram of the structure of a small molecule transmembrane permeation free energy calculation system according to an embodiment of this application.

[0057] Figure 4 A schematic diagram of the structure of a system for calculating the transmembrane permeation free energy of small molecules according to an embodiment of this application. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1

[0060] like Figure 1 As shown, a method for calculating the transmembrane permeation free energy of small molecules includes the following steps:

[0061] S1: Obtain the SMILES format description of the target small molecule;

[0062] S2: A graph-based molecular bond network algorithm is used to perform coarse-grained parameterization of the molecule to generate a coarse-grained molecular model;

[0063] S3: Construct a coarse-grained water-film system and dynamically assemble the coarse-grained molecular model with the water-film system to position the molecules to the preset positions;

[0064] S4: Perform a stretching dynamics simulation to pull small molecules across the membrane using a centroid traction method, generating transmembrane reaction coordinate trajectories;

[0065] S5: Sample windows are extracted at fixed intervals along the reaction coordinate trajectory to simulate umbrella-shaped sampling;

[0066] S6: The WHAM algorithm is used to perform weighted histogram analysis on the umbrella-shaped sampling results, and the transmembrane free energy change curve and water-membrane partition free energy are output.

[0067] The above scheme covers all steps from obtaining small molecule information to finally outputting the free energy change curve and water-membrane partition free energy, providing a systematic solution for related research.

[0068] As a refinement of the above embodiment, in step S1, the SMILES (Simplified Molecular Input Line Entry System) format description of the molecule is obtained through the PubChem website; wherein, the SMILES format, namely the simplified molecular linear input specification, is a specification that explicitly describes the molecular structure using ASCII strings. This string can accurately describe the chemical structure of a molecule and can be recognized by most molecular editing software and converted into a two-dimensional or three-dimensional model of the molecule, so as to facilitate compatibility with more software and improve the compatibility of input data.

[0069] As a refinement of the above embodiment, the graph-based molecular bond network algorithm in step S2 is a fast, universal, and symmetric-preserving coarse-grained mapping and parameterization system for small molecules. Its coarse-grained modeling of small molecules mainly involves three stages: 1. Mapping atoms onto coarse-grained beads; 2. Parameterization of non-bonding interactions between atoms; 3. Parameterization of bonding interactions between atoms; 4. To form a more accurate molecular conformation, in this embodiment, the RDKit package is used to generate initial three-dimensional structures of 1000 molecules through the MMFF94 force field, followed by screening and optimization; 5. Finally, the molecular structure and parameter files are generated based on the atom-bead mapping.

[0070] A graph-based molecular bond network algorithm is used to perform coarse-grained parameterization of molecules. Through operations such as atom mapping, interaction parameterization, and structure optimization, coarse-grained molecular models can be generated more accurately, thereby improving computational accuracy.

[0071] As a refinement of the above embodiment, step S3 uses the CHARMM-GUI online tool to create a coarse-grained water-film system; wherein, CHARMM-GUI is an online tool that can quickly construct lipid membrane simulation systems, and the PACECGBuilder module provided by it can be used to construct a coarse-grained water-film system.

[0072] Molecular dynamics simulation software was used to assemble a coarse-grained model of the molecule and a water-film system, and the molecule was placed in a preset position within the system.

[0073] Specifically, in order to conduct molecular simulation experiments of coarse-grained drug molecules across membranes, the system needs to be assembled and pretreated. In this step, the molecular dynamics simulation software Gromacs was used, and all force fields involved in the molecular dynamics simulation were Martini 2.2 coarse-grained force fields. (1) The coarse-grained drug molecule box and center were redefined using the gmxeditconf tool to facilitate the subsequent insertion steps. (2) The drug was inserted into the coarse-grained DOPC membrane system at a position about 1 nm above the membrane, replacing the water molecules at the corresponding positions in the system. In order to calculate the number of water molecules replaced, the number of particles in the drug molecule structure file, the coarse-grained membrane system structure file, and the assembled system structure file were extracted using the linux command sed and assigned to the variables “lignum”, “martininum”, and “complexnum”. The number of water molecules replaced can be obtained by adding the number of molecular particles to the number of membrane system particles and subtracting the number of particles in the assembled system, and the system topology file of the molecular simulation is changed accordingly. (3) After modifying the topology file, 0.1 M NaCl was added to the system using the gmxgenion tool. (4) The steepest descent method was used to optimize the energy of the assembled system, and six consecutive system equilibrium simulations were performed, gradually reducing the constraints on the coarse-grained lipid beads in the system. During the equilibrium process, the v-rescale temperature coupling algorithm was used to control the system equilibrium temperature to 303.15 K, which is close to the experimental temperature of most laboratories. The Berendsen method was used to control the system pressure to 1 bar.

[0074] This step involves constructing a coarse-grained water-film system and dynamically assembling it with a coarse-grained molecular model. It details operations such as small molecule localization, water molecule replacement, solution addition, energy optimization, and equilibrium simulation to ensure a reasonable system construction and lay the foundation for subsequent simulations. During equilibrium simulation, a specific algorithm is used to precisely control the system's equilibrium temperature and pressure, ensuring the stability of the simulation environment and making the simulation results more reliable.

[0075] As a refinement of the above embodiment, steps S4 and S5 use the molecular dynamics simulation software Gromacs to perform energy optimization and system equilibrium on the simulation system, and perform tensile dynamics simulation and umbrella sampling simulation to obtain the free energy change curve of the compound permeate across the membrane and the water-membrane partition free energy.

[0076] Specifically, the method for performing the umbrella-shaped sampling simulation is as follows:

[0077] The sampling window is sampled along the Z-axis at 0.1 nm intervals;

[0078] Each system has 60 windows, and each window has a simulation duration of 100 ns.

[0079] This step clarifies the specific method of umbrella-shaped sampling simulation, including the sampling window spacing, number, and simulation duration of each window, ensuring the comprehensiveness and accuracy of sampling, thereby making the output free energy change curve and water-film distribution free energy more valuable for reference.

[0080] This embodiment uses the phospholipid bilayer model as an example for specific illustration:

[0081] (1) First, obtain the SMILES format description of the molecule through the PubChem website.

[0082] (2) Then, the molecular structure is coarsely modeled using a graph-based molecular bond network algorithm to generate molecular structure and parameter files.

[0083] (3) In the coarse-grained model of the constructed water-film system, the inner and outer leaflets of the phospholipid bilayer model are each composed of 144 1,2-dioleoyl-sn-propanetriyl-3-choline phosphate (DOPC) molecules, and the water layer thickness is set at 6.0 nm inside and outside the phospholipid bilayer. The phospholipid bilayer model is perpendicular to the z-axis of the system and located at the center of the system, with an initial thickness of about 4.0 nm. Molecular dynamics simulation software is used to assemble the coarse-grained model of the molecules and the water-film system, and the molecules are placed in the preset positions in the system.

[0084] (4) In the molecular dynamics simulation software, the small molecules are pulled through the lipid bilayer membrane by the center of mass traction to complete the stretching dynamics simulation and form the reaction coordinate trajectory.

[0085] Specifically, in the stretching dynamics simulation, it is assumed that a virtual atom is connected to one side of the spring and a drug molecule is connected to the other side. The virtual atom then moves at a constant speed v along a specified direction, thereby pulling the drug molecule across the membrane. In this embodiment, the hypothetical elastic constant is chosen to be 1000 kJ / (mol·nm(2)), the speed of the virtual atom is 0.001 nm / ps, and the simulation duration is 6000 ps. Therefore, the distance the drug molecule moves due to the force is approximately 6 nm.

[0086] (5) Extract the molecular reaction coordinate trajectory at fixed intervals and perform umbrella sampling simulation. Based on the output drug coordinate information, extract the corresponding windows along the Z-axis at intervals of 0.1 nm to perform umbrella sampling simulation. Approximately 60 windows are extracted for each system. The simulation duration for each window is 100 ns, and the total umbrella sampling simulation is approximately 6 ms.

[0087] (6) Finally, the output after umbrella sampling simulation was weighted by WHAM method to obtain the free energy change curve of molecules in the transmembrane process and obtain the water-membrane transfer free energy of molecules.

[0088] Example 2

[0089] like Figure 2 As shown, a method for establishing a dataset of small molecule transmembrane permeation free energy includes:

[0090] S1: Screen neutral small molecules from the database and extract molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds and polar surface area to form a high-dimensional feature space.

[0091] S2: The t-SNE algorithm is used to reduce the high-dimensional feature space to a two-dimensional plane;

[0092] S3: Randomly select a preset number of small molecules until the distribution uniformity condition is met;

[0093] S4: Calculate the transmembrane permeation free energy of the selected small molecule using the method described above;

[0094] S5: Correlate the molecular SMILES format description with the calculated transmembrane permeation free energy to generate a structured dataset.

[0095] As a refinement of the above embodiment, in step S1, all neutral small molecule compounds with a molecular weight less than 500 are selected from the PubChem database, and the molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar area of ​​each small molecule are combined to form the high-dimensional feature space of the molecule.

[0096] PubChem, or Organic Small Molecule Bioactivity Database, is a database of organic small molecule chemical modules used to store compound structural information, biochemical experimental data, and other relevant information. Molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar area are all features that may affect the transmembrane ability of small molecules. Therefore, these features are selected to form a high-dimensional feature space for molecular characteristics.

[0097] As a refinement of the above embodiment, step S2 utilizes the t-SNE method to reduce the dimensionality of the chemical feature space, forming a two-dimensional mapping of molecular features. At this point, the chemical feature space of the molecules can be visualized using a scatter plot. The t-SNE method is a nonlinear dimensionality reduction method used to map high-dimensional data to a two- or three-dimensional space for visualization and analysis. This method is a commonly used dimensionality reduction method widely applied in data mining, machine learning, and visualization. It can preserve the local structure of the data and also uncover the global structure through separation and clustering. In this step, the t-SNE method is mainly used to transform the high-dimensional chemical feature space into a two-dimensional space, thereby facilitating intuitive perception and description of the chemical feature space and making it easier to select representative molecules in subsequent steps.

[0098] As a refinement of the above embodiment, step S3 involves randomly selecting a predetermined number of molecules from the molecules and repeating this process multiple times. A visualization tool is used to inspect the selected molecules until their features are uniformly distributed in a two-dimensional mapping of molecular features.

[0099] Specifically, the selected molecules should have the following characteristics in chemical space:

[0100] 1. There is no obvious clustering of features;

[0101] 2. They are basically evenly distributed in the chemical space;

[0102] 3. There is no obvious missing chemical spatial region coverage. In this case, the selected molecule is considered to be able to represent this chemical space relatively well.

[0103] In this embodiment, the following thresholds are set: when the difference in two-dimensional spatial distribution density is less than or equal to 20%, the selected molecules are considered to be basically uniformly distributed in the chemical space; when the overall coverage is greater than or equal to 95%, there is no obvious missing coverage of the chemical space region.

[0104] As a refinement of the above embodiments, step S4 uses a small molecule transmembrane permeation free energy calculation method described in Example 1 to calculate the transmembrane permeation free energy of all the molecules, and combines it with the SMILES format string of the molecules to form a compound transmembrane free energy dataset. This dataset is presented in CSV format and can be identified and used using mainstream data processing tools.

[0105] Example 3

[0106] like Figure 3 As shown, a system for calculating the transmembrane permeation free energy of small molecules includes:

[0107] Molecular coarsening module: Configured to receive SMILES input, and use a graph-based molecular bond network algorithm to coarse-grained parameterize the molecule to generate a coarse-grained molecular model;

[0108] Membrane environment construction module: configured to generate coarse-grained water-membrane system; implemented by calling the CHARMM-GUI interface;

[0109] Dynamic Assembly Engine: Configured to dynamically assemble coarse-grained molecular models with water-film systems, positioning molecules to preset positions and performing energy optimization and equilibrium simulations;

[0110] Transmembrane trajectory generator: configured to drive molecules to move along the membrane normal according to the stretching kinetics model, generating transmembrane reaction coordinate trajectories;

[0111] Free energy calculation unit: configured to extract sampling windows at fixed intervals along the reaction coordinate trajectory, perform umbrella-shaped sampling simulation, and output free energy curves and water-film distribution free energy through the WHAM algorithm.

[0112] Specifically, the configuration is to extract trajectory windows at 0.1nm intervals for umbrella-shaped sampling.

[0113] It should be noted that in this embodiment, by inputting the SMILES representation of the molecule into the molecular coarsening module, a coarsened molecular topology is generated based on the molecular bond network; then, the membrane environment construction module is used to build a water-membrane system that meets the experimental requirements, and the constructed molecular topology is inserted into the water-membrane system based on the dynamic assembly engine; the transmembrane trajectory generator and free energy calculation unit are used to output the transmembrane trajectory and free energy curves, as well as the water-membrane distribution free energy data.

[0114] Example 4

[0115] like Figure 4 As shown, a system for establishing a dataset of small molecule transmembrane permeation free energy includes:

[0116] Chemical feature extractor: configured to screen neutral small molecules and extract molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar surface area to form a high-dimensional feature space;

[0117] Feature space dimensionality reducer: configured to use the t-SNE algorithm to map high-dimensional features to a two-dimensional plane;

[0118] Molecular screening engine: Configured to randomly select a preset number of small molecules until the distribution uniformity condition is met;

[0119] Data synthesis terminal: configured to calculate transmembrane permeation free energy using the small molecule transmembrane permeation free energy calculation method, and associate the molecule SMILES with the transmembrane free energy calculation results to generate a structured dataset.

[0120] It should be noted that in this embodiment, the input molecular SMILES representation is parsed by a chemical feature extractor to obtain features such as molecular weight, number of heavy atoms, number of hydrogen bond donors / acceptors, number of rotatable bonds, and polar surface area. Then, a feature space dimensionality reducer is used to reduce the dimensionality of the high-dimensional features and map them to a two-dimensional plane. Molecules are then randomly selected by a molecular screening engine according to preset constraints. The free energy of the selected molecules is calculated according to the method in Example 1, and the calculation results are generated into a structured dataset based on the data synthesis terminal.

[0121] Example 5

[0122] This disclosure provides a device for calculating the transmembrane permeation free energy of small molecules, including a processor and a memory. Optionally, the device may further include a communication interface and a bus. The processor, communication interface, and memory can communicate with each other via the bus. The communication interface can be used for information transmission. The processor can call logical instructions in the memory to execute the method for calculating the transmembrane permeation free energy of small molecules described in the above embodiments.

[0123] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0124] Memory, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as the program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor executes the program instructions / modules stored in the memory to perform functional applications and data processing, thereby realizing the method for calculating the transmembrane permeation free energy of small molecules in the above embodiments.

[0125] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory may include high-speed random access memory and may also include non-volatile memory.

[0126] The technical solutions of this disclosure can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in this disclosure. The aforementioned storage medium can be a non-transitory storage medium, including: a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and other media capable of storing program code. It can also be a transient storage medium.

[0127] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for calculating the transmembrane permeation free energy of small molecules, characterized in that, Includes the following steps: Obtain the SMILES format description of the target small molecule; A graph-based molecular bond network algorithm is used to perform coarse-grained parameterization of molecules to generate coarse-grained molecular models. A coarse-grained water-film system was constructed, and the coarse-grained molecular model was dynamically assembled with the water-film system to position the molecules at a predetermined location. Perform a stretching dynamics simulation to pull small molecules across the membrane using a center-of-mass traction method, generating transmembrane reaction coordinate trajectories; Sampling windows are drawn at fixed intervals along the reaction coordinate trajectory to simulate umbrella-shaped sampling; The WHAM algorithm was used to perform weighted histogram analysis on the umbrella-shaped sampling results, and the transmembrane free energy variation curve and water-membrane partition free energy were output. The molecular bond network algorithm includes: Map atoms onto coarse-grained beads; Parameterizing nonbonded interactions between atoms; Parameterizing interatomic bonding interactions; Initial three-dimensional molecular structures were generated and optimized based on the MMFF94 force field. Generate molecular structure and parameter files based on atom-bead mapping; The dynamic assembly includes: Define the small molecule simulation box and center; Small molecules are inserted 1 nm above the membrane system and replace the water molecules at that position; Calculate the number of water molecules replaced and add 0.1M NaCl solution; Energy optimization and multi-stage balance simulation are performed.

2. The method for calculating the transmembrane permeation free energy of small molecules according to claim 1, characterized in that, The equilibrium simulation process uses the v-rescale temperature coupling algorithm to control the system equilibrium temperature at 303.15K, and the Berendsen method to control the system pressure at 1 bar.

3. The method for calculating the transmembrane permeation free energy of small molecules according to claim 1, characterized in that, The specific method for performing umbrella-shaped sampling simulation is as follows: The sampling window is sampled along the Z-axis at 0.1 nm intervals; Each system has 60 windows, and each window has a simulation duration of 100 ns.

4. A method for establishing a dataset of transmembrane permeation free energy of small molecules, characterized in that, include: Neutral small molecules were screened from the database, and their molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar surface area were extracted to form a high-dimensional feature space. The t-SNE algorithm is used to reduce the dimensionality of the high-dimensional feature space to a two-dimensional plane; Randomly select a preset number of small molecules until the distribution uniformity condition is met; The transmembrane permeation free energy of a selected small molecule is calculated using the method for calculating the transmembrane permeation free energy of a small molecule as described in any one of claims 1-3. By correlating the molecular SMILES format description with the calculated transmembrane permeation free energy, a structured dataset is generated.

5. The method for establishing a small molecule transmembrane permeation free energy dataset according to claim 4, characterized in that, The distribution uniformity conditions are: the difference in distribution density in two-dimensional space is less than or equal to 20% and the overall coverage is greater than or equal to 95%.

6. A system for calculating the transmembrane permeation free energy of small molecules, characterized in that, The method for calculating the transmembrane permeation free energy of small molecules according to any one of claims 1-3 includes: Molecular coarsening module: Configured to receive SMILES input, and use a graph-based molecular bond network algorithm to coarse-grained parameterize the molecule to generate a coarse-grained molecular model; Membrane environment construction module: configured to generate a coarse-grained water-membrane system; Dynamic assembly engine: configured to dynamically assemble coarse-grained molecular models with water-film systems, positioning molecules to preset positions; Transmembrane trajectory generator: configured to drive molecules to move along the membrane normal according to the stretching kinetics model, generating transmembrane reaction coordinate trajectories; Free energy calculation unit: configured to extract sampling windows at fixed intervals along the reaction coordinate trajectory, perform umbrella-shaped sampling simulation, and output free energy curves and water-film distribution free energy through the WHAM algorithm.

7. A system for establishing a dataset of transmembrane permeation free energy of small molecules, characterized in that, include: Chemical feature extractor: configured to screen neutral small molecules and extract molecular weight, number of heavy atoms, number of hydrogen bond donors, number of hydrogen bond acceptors, number of rotatable bonds, and polar surface area to form a high-dimensional feature space; Feature space dimensionality reducer: configured to use the t-SNE algorithm to map high-dimensional features to a two-dimensional plane; Molecular screening engine: Configured to randomly select a preset number of small molecules until the distribution uniformity condition is met; Data synthesis terminal: configured to calculate transmembrane permeation free energy using the small molecule transmembrane permeation free energy calculation method as described in any one of claims 1-3, and associate the molecule SMILES with the transmembrane free energy calculation results to generate a structured dataset.

8. A device for calculating the transmembrane permeation free energy of small molecules, comprising a processor and a memory storing program instructions, characterized in that, The processor is configured to execute the method for calculating the transmembrane permeation free energy of small molecules as described in any one of claims 1-3 when running the program instructions.

Images