Method for simulating, predicting and screening percutaneous external preparation prescriptions based on expert system and molecular dynamics

Abstract

The invention relates to a method for simulating, predicting and screening percutaneous external preparation prescriptions based on an expert system and molecular dynamics. The method is used for predicting and screening the percutaneous external preparation prescriptions. The method comprises the following steps: evaluating the compatibility of medicines and auxiliary materials through an expert system, and screening out candidate auxiliary materials with medium risk or low risk; on the basis of a molecular dynamics model principle, using Materials Studio software to calculate solubility parameters of the medicine and the candidate auxiliary materials, and screening out the auxiliary material with the best compatibility with the medicine; based on a molecular dynamics simulation principle, a GROMACS software package is used for carrying out full-atom molecular dynamics simulation, prescription factor concentrations influencing the percutaneous external preparation are investigated, and the optimal prescription factor concentration is screened; and finally, carrying out quality evaluation on the screened optimal prescription and the reference preparation under the same condition. By means of the method, the optimal prescription suitable for the transdermal external preparation can be efficiently screened out, and the efficiency and quality of preparation research and development are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of pharmaceutical preparations, and in particular relates to a method for predicting and screening transdermal external preparation prescriptions based on an expert system and molecular dynamics simulation. Background Art

[0002] The prescription design of Chinese medicine pharmacy is the process of screening excipients and determining the types and dosages of excipients or excipients according to the dosage, properties, medical requirements, route of administration and characteristics of the dosage form of the drug. Its purpose is to solve the problems of the preparation's formability, safety and stability. Therefore, the preparation prescription screening process can be summarized as the process of determining, selecting and optimizing the concentrations of raw materials and molding matrix excipients and their various prescription factors. Preparation prescription screening is the link between product design and process design, and plays a connecting role. At present, the commonly used method for screening prescriptions of transdermal topical preparations in Chinese medicine pharmacy is manual trial and error experiments based on experience. This method has many disadvantages, mainly manifested in that a large number of repeated experimental operations are required during the screening of transdermal topical preparation prescriptions, which is time-consuming and laborious; a large number of raw materials and experimental equipment are required, which is costly; experimental operations are often limited by experimental conditions and affected by human factors, and experimental results are unpredictable; the experimental process focuses on macroscopic phenomena, and it is difficult to deeply understand the microscopic dynamic processes and molecular interaction mechanisms in the preparation process. Therefore, traditional experimental methods have certain limitations in the screening of prescriptions for transdermal topical preparations in traditional Chinese medicine pharmacy and need to be combined with other methods to improve efficiency and accuracy.

[0003] Although some people have proposed applying computer simulation technologies such as artificial intelligence machine learning algorithms and molecular simulation to pharmaceutical research, researchers are somewhat overwhelmed by the installation and use of a series of complex simulation software. The relevant computer principles, simulation principles, mechanical principles, and machine learning principles involved are complex and difficult, and most researchers find it difficult to understand, so they are discouraged and choose traditional experimental methods, or use a small part of computer technology to conduct scientific research. Overall, there is a lack of a research system and method that is easy to understand and simple to operate. Summary of the invention

[0004] The present invention aims at the problems existing in the above-mentioned research process of transdermal external preparation prescription screening, and aims to provide a method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation can evaluate the optimal compatibility and formulation risk of drugs with specific structural characteristics, accurately calculate the solubility parameters of drugs and candidate excipients, and obtain the best excipient with better preparation molding in theory according to the principle of "like is compatible". The all-atom molecular dynamics simulation method that can describe atomic details is used to perform microscopic dynamic molecular simulation on prescription factors of different concentrations to screen out the concentration of prescription factors with good release effect and stable quality. At the same time, a classic skin model is built to screen the types and optimal transdermal concentrations of transdermal penetration enhancers, and finally the in vitro transdermal performance of transdermal external preparations is studied. The method is flexible, simple and easy to operate, and the prediction results are accurate, economical and efficient.

[0005] The present invention describes a method for predicting and screening transdermal topical preparation prescriptions based on an expert system and molecular dynamics simulation. The drug dosage forms involved are mainly transdermal topical preparations. The method mainly involves the prediction and screening of sinomenine gel prescriptions.

[0006] This method describes a method for predicting and screening prescriptions for transdermal topical preparations based on expert systems and molecular dynamics simulations, which mainly includes three parts: (1) PharmDE: a new expert system for drug excipient compatibility evaluation: a knowledge-driven expert system for drug excipient compatibility risk evaluation. PharmDE first developed an information-rich database to store incompatibility data, covering 532 data items from 228 selected articles. The database is currently open and can be obtained for free at https: / / pharmde.computpharm.org / ; then, based on existing theoretical knowledge and formulation research experience, drug-excipient interaction rules are created; finally, the database search is organically combined with rule-based compatibility risk prediction to make it a compatibility risk expert system, including four main functions: basic search of compatibility database, similarity search and matching, drug compatibility risk evaluation, and formulation compatibility risk evaluation. (2) Molecular dynamics simulation, including the calculation of the solubility parameter δ of drugs and candidate excipients, formulation factor concentration screening, penetration enhancer type screening, and penetration enhancer concentration screening. In this method based on expert system and molecular dynamics simulation to predict and screen transdermal topical preparations, the solubility parameters were calculated using Materials Studio molecular dynamics simulation software, and the concentration of prescription factors was screened using gromacs software. (3) In vitro transdermal performance experimental verification: The modified Franz diffusion cell method was mainly used to study the in vitro transdermal performance of transdermal topical preparations and reference preparations using isolated rat skin as a transdermal permeation model.

[0007] The method for predicting and screening transdermal topical preparation prescriptions based on molecular dynamics simulation provided by the present invention is a systematic, complete, simple to operate, accurate prediction method for transdermal topical preparation prescription screening that combines computer technology with traditional experiments. The invention is economical, efficient and accurate, can replace the trial and error method of traditional prescription screening, and advantageously combines computer technology and traditional experiments to complement each other, and ultimately prepares transdermal topical preparations with good efficacy and high quality.

[0008] The method of the present invention is that the experimental results can be predicted and analyzed directly using computer simulation software without conducting actual experiments. Compared with the preparation prescription screening method of traditional experiments, the method of the present invention does not require a large amount of experimental materials and time, and can quickly perform simulation calculations on a computer, greatly reducing the research and development cost and time; the performance of the preparation can be predicted before the experiment, which is helpful to find problems in advance and make improvements, and reduce the risk of failure in later experiments; according to the simulation results, the ingredients and proportions in the prescription can be optimized to improve the quality and effect of the preparation; by simulating the molecular behavior under different prescription conditions, the performance of the transdermal preparation, such as drug release rate, permeability, etc., can be predicted, providing reliable guidance for screening the best prescription; molecular dynamics simulation can reveal the details of molecular interactions, help researchers to deeply understand the microscopic properties of the drug in the transdermal preparation, such as the transmission mechanism, solubility, permeability, etc., and provide a more accurate basis for prescription design; the molecular dynamics simulation software is relatively simple to operate, and researchers can easily perform simulation experiments, and can adjust parameters and conditions as needed, with high flexibility. The method for predicting and screening transdermal external preparation prescriptions based on molecular dynamics simulation provided by the present invention is a systematic, complete, simple to operate, accurate prediction method for transdermal external preparation prescription screening that combines computer technology with traditional experiments. It is economical, efficient and accurate, can replace the trial and error method of traditional prescription screening, and provides a clear research idea and scientific research confidence for researchers in the field of pharmaceutical research to apply computer technology to pharmaceutical research. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0010] Figure 1 is the change of solubility parameters of carbomer polymer with degree of polymerization;

[0011] Figure 2 The results of differential scanning calorimetry experiments, (a) to (c) are the results of physical mixing of chitosan, carboxymethyl cellulose and carbomer with sinomenine, respectively, and (d) is the differential scanning calorimetry result of the sample sinomenine;

[0012] Figure 3 The results of ordinary optical microscopy, (a) shows the microscopic state of sinomenine mixed with carbomer matrix, (b) shows the microscopic state of sinomenine mixed with carboxymethyl cellulose matrix, (c) shows the microscopic state of sinomenine mixed with chitosan (magnification: 4×10);

[0013] Figure 4 It is the initial simulation system for the effect of carbomer polymer with a degree of polymerization of 50 on the quality of sinomenine gel;

[0014] Figure 5 The skin bilayer membrane model contains ceramide-sphingosine (purple) and cholesterol (orange) and free fatty acids (green) in a 2:2:1 molar ratio.

[0015] Figure 6 It is the balance result of the effect of the concentration screening of the prescription factors on the quality of sinomenine gel, a1~g1 are the changes of the system energy with time in the carbomer polymer polymerization degree screening, sinomenine concentration screening, carbomer concentration screening, single ethanol concentration screening, prescription factor ethanol concentration screening, glycerol concentration screening, and penetration enhancer concentration screening, a2~g2 are the changes of the average RMSD value of the carbomer polymer chain in the carbomer polymer polymerization degree screening, sinomenine concentration screening, carbomer concentration screening, single ethanol concentration screening, prescription factor ethanol concentration screening, glycerol concentration screening, and penetration enhancer concentration screening system;

[0016] Figure 7 The results of the Rg changes of five random carbomer polymers in the gel simulation system with different polymerization degrees over time, a~f are the simulation systems with polymerization degrees of 15, 30, 40, 50, 60, and 70 respectively;

[0017] Figure 8 The results of the influence of carbomer polymers with different polymerization degrees on the quality of sinomenine gel, g~j respectively represent the results of the average Rg and RMSD of carbomers with different polymerization degrees changing with the polymerization degree, the results of the DA changing with the polymerization degree, the results of the density distribution of carbomer polymers changing with the polymerization degree, and the results of the density distribution of solvent water changing with the polymerization degree;

[0018] Fig. 9The results of the effects of different sinomenine concentrations on the quality of sinomenine gels are shown in Figure 1. a, b, and c are the results of the average Rg value and RMSD value of the polymer matrix chain in different sinomenine concentration systems changing with time, the results of the diffusion rate of drug molecules in different systems changing with time, and the density distribution diagram of the polymer matrix chain network in the Z-axis direction.

[0019] Fig.10 The results of the Rg value variation trend of five random matrix chains in different carbomer polymer concentration systems over time, a~e are gel simulation systems with concentrations of 0.5%, 1%, 1.5%, 2%, and 2.5% respectively;

[0020] Fig.11 The results of the influence of different polymer matrix concentrations on the quality of sinomenine gel, e~g respectively represent the results of the Rg value and RMSD value of the matrix chain in different concentrations changing with the matrix concentration, the results of the DA changing with the matrix concentration, and the results of the average density distribution diagram of the gel system with different concentrations in the Z-axis direction;

[0021] Fig.12 The results of the Rg values ​​of five random polymer matrix chains changing with time in different ethanol concentration systems, a~f are 0%, 20%, 40%, 30%, 60%, 80% ethanol concentration systems respectively;

[0022] Fig.13 is the result of the effect of ethanol concentration on the quality of sinomenine gel, a is the change trend of sinomenine diffusion rate with the change of ethanol concentration, b is the RDF result of the interaction between sinomenine and polymer matrix chain molecules in the simulation system with ethanol concentration of 20% to 30%;

[0023] Fig.14 The trend results of the Rg values ​​of five random polymer matrix chains in the gelling agent system with different glycerol concentrations over time, (a) to (d) are the gelling agent simulation systems with glycerol concentrations of 0%, 5%, 10%, and 15% respectively;

[0024] Fig.15 (a) shows the result of the diffusion rate of sinomenine molecules in gelling agents with different glycerol concentrations changing with the glycerol concentration, and (b) shows the RDF result of the interaction between the oxygen atom on the hydroxyl group of sinomenine and the hydrogen atom in the water molecule in the systems with different glycerol concentrations;

[0025] Fig.16(A) shows the snapshots of the molecular dynamics simulation trajectories of the 0th frame and the last frame saved in the time from 0 to 200 ns when 1,2-propylene glycol and laurocapram were used alone or in combination at different concentrations. (a) to (i) are six concentration conditions, (a) to (b) are 0%, (c) to (d) are 10%, (e) to (f) are 2%, (g) to (h) are 6% + 1.5%, (i) to (j) are 10% + 2%, (k) to (l) are 12% + 3%. Purple represents sinomenine. The middle represents the lipid bilayer, and the penetration enhancer and solvent are not displayed; (B) represents the molecular dynamics simulation trajectory snapshots of the 0th frame and the last frame saved in the time from 0ns to 200ns for different menthol concentration systems, (a) to (b) are menthol concentrations of 0%, (c) to (d) are 2%, (e) to (f) are 5%, (g) to (h) are 10%, (i) to (j) are 15%, purple represents sinomenine, green represents menthol, and the solvent is not displayed. The transparent box in the middle represents the skin lipid bilayer;

[0026] Fig.17 The density distribution diagrams of different formulation factors and lipid bilayers in the system along the Z-axis direction of the box, (a) is the simulation system of 1,2-propylene glycol and laurocapram penetration enhancer, (b) is the simulation system of menthol penetration enhancer, where the middle convex (yellow circle) is the density distribution of lipid bilayer, and the lower convex (red circle) is the density distribution of sinomenine and penetration enhancer;

[0027] Fig.18 is the order parameter of the head and tail of CER and FFA in the lipid bilayer membrane at different penetration enhancer concentrations, a and c are the order parameters of the head and tail of CER and FFA in the lipid structure of 1,2-propylene glycol and laurocapram system, b and d are the order parameters of the head and tail of CER and FFA in the lipid structure of menthol system, △SZ is the difference between the maximum and minimum order parameter values ​​in the change of order parameter;

[0028] Fig.19 is the average diffusion rate of sinomenine in different penetration enhancer concentration systems from 0 to 200 ns, Men is menthol, Propy+Lau is 1,2-propylene glycol and laurocapram;

[0029] Fig. 20 The photo of sinomenine gel, the left one is the sinomenine gel obtained by molecular dynamics simulation method, and the right one is the reference preparation of sinomenine gel with the same concentration;

[0030] Fig.21 The in vitro permeation curves of sinomenine gel and reference preparation, A represents sinomenine gel, B represents reference preparation;

[0031] Fig. 22A flowchart of the present invention for predicting and screening transdermal topical preparation prescriptions based on expert system and molecular dynamics simulation;

[0032] Fig.23 This is a flow chart of the method using sinomenine gel prescription screening as an example in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The present invention provides a method for predicting and screening a transdermal external preparation prescription based on an expert system and molecular dynamics simulation, comprising the following steps:

[0034] A) Use an expert system to evaluate the incompatibility of the drug and each excipient, and exclude excipients that are incompatible with the drug; then use the expert system to perform a formulation risk assessment on the drug and the excipients that have passed the incompatibility assessment, and screen out candidate excipients with medium or low risk;

[0035] B) using Materials Studio software based on the all-atom molecular dynamics simulation method to calculate the solubility parameter δ of the drug and each candidate excipient, and screening out the excipient with the smallest solubility parameter difference between the drug and the excipient and δ < 7.5;

[0036] C) obtaining molecular structure files of drugs and excipients, selecting GAFF force field, creating a box, adding drug molecules and excipient molecules according to the volume percentage of components in the transdermal external preparation prescription to the created box to form an initial simulation system, sequentially performing energy minimization and pre-balancing on the initial simulation system, and after the pre-balancing, performing all-atom molecular dynamics simulation on the simulation system under different prescription factors using GROMACS software, evaluating according to the simulation results, and obtaining the desired preparation prescription;

[0037] D) performing quality evaluation on the preparation prescription obtained in step C).

[0038] The method of the present invention mainly includes three steps, which are as follows:

[0039] 1. Study on compatibility of drugs and excipients

[0040] For the special dosage form of percutaneous topical, the selection of excipients is crucial. In addition to the function of giving the preparation molding, the excipients may also change the physicochemical properties of the drug, regulate the release and dissolution rate of the drug in the preparation, and even change the clinical efficacy, safety and stability of the drug. In the field of materials science, compatibility is often used to characterize the ability of the components of the blend to accommodate each other and form a macroscopic uniform material. The present invention first evaluates the formula risk and incompatibility of the drug and the molding matrix excipient when predicting and screening the prescription of the percutaneous topical preparation. The PharmDE expert system is mainly used to predict and evaluate the formula risk and incompatibility of the drug and the excipient, and determine the candidate excipients.

[0041] Then, the solubility parameters of the drug and candidate excipients are further calculated by molecular dynamics simulation. The principle is "like dissolves like". The closer the solubility parameters of the drug and different excipients are, the better the compatibility is, thereby determining the optimal excipient for preparing transdermal topical preparations with drug active ingredients.

[0042] Considering that the matrix excipients of transdermal topical preparations are mostly high molecular weight polymers, which have a large molecular weight, it is unrealistic to simulate their true molecular weight. Therefore, this part uses the molecular dynamics simulation method to calculate the solubility parameters of matrix excipients with different degrees of polymerization, and further calculates the difference in solubility parameters between the drug and the matrix excipient. When the difference in solubility parameters is less than 7.5, it is considered that the two substances have good compatibility after mixing;

[0043] The simulation prediction results of the above-mentioned PharmDE expert system and molecular dynamics simulation were experimentally verified using scanning calorimetry and ordinary microscopy experimental methods to ensure the accuracy of the simulation prediction results.

[0044] 2. Molecular dynamics simulation of the effects of different prescription factor concentrations on the quality of transdermal topical drug preparations

[0045] The molecular dynamics simulation method was used to simulate the effects of different prescription factors on the quality of transdermal topical preparations. The currently popular molecular dynamics simulation software GROMACS software package was mainly used. The software has the advantages of open source, free, and fast calculation speed. The force field used is the GAFF force field compatible with the AMBER force field. The single factor method was mainly used for prescription factor screening. The volume percentage of each prescription factor was calculated according to the drug prescription amount to obtain the number of molecules of the prescription factor.

[0046] Obtain the topological files of each molecule. Different force fields need to be selected for different research purposes and systems. Currently, the Amber force field has high accuracy in the study of biological systems and condensed matter. Therefore, the force field selected in the present invention is the GAFF (Generation Amber Force Field) force field that is compatible with the Amber force field. This force field has advantages in processing various organic small molecules. Since GROMACS does not have its own GAFF force field, the topological parameters of the molecules such as drugs, excipients, solvents, additives, etc. used in the present invention are calculated by GaussianView combined with Multiwfn (the software specifically refers to Lu T, Chen F. Multiwfn: A multifunctional wavefunction analyzer [J]. Journal of computational chemistry, 2012, 33 (5): 580-592.) RESP distribution charge, and finally the topological files of each molecule are generated with the help of Sobtop software and stored for later use;

[0047] In order to study the effects of different prescription factors and concentrations on transdermal drug preparations, an initial molecular dynamics simulation system was further constructed. The molecular dynamics simulation method was used to minimize the energy of the constructed initial system in order to eliminate unreasonable contacts and collisions between molecules in the system, so that the initial structure of the system was in the lowest energy and reasonable state;

[0048] After the energy minimization is completed, the initial system is further pre-balanced at room temperature and pressure to gradually stabilize the temperature and pressure in the system, thereby obtaining an initial simulation system for studying the effects of different prescription factors on transdermal drug preparations.

[0049] GROMACS software was used to simulate different prescription factors and perform molecular dynamics simulations of the initial system of transdermal drug preparations at set times.

[0050] After the simulation, the simulated trajectory was corrected with periodic boundaries, and the trajectory was imported into VMD software for observation and analysis, and the microscopic dynamic kinetic behavior of drug transdermal preparations under molecular dynamics simulation was summarized; the analysis module provided by GROMACS was used to calculate the diffusion coefficient of drug molecules in different systems, the root mean square displacement, gyration radius, density distribution diagram of molding matrix excipients, radial distribution function of drug and molding matrix molecules, radial distribution function of oxygen (O) atoms in drug molecules and hydrogen (H) atoms in water molecules, and the evaluation indicators were compared and analyzed to screen out the best prescription factor ratio of drug transdermal preparations. Finally, the intermolecular interaction laws affecting drug transdermal preparations were summarized.

[0051] 3. Quality evaluation of transdermal external drug preparations

[0052] The therapeutic effect of transdermal external preparations is seriously affected by the transdermal effect. The present invention constructs a skin lipid bilayer membrane structure in the screening of the ratio of permeation enhancers for transdermal external preparations, uses a molecular dynamics simulation method to simulate the interaction between the drug and the skin, observes the actual contact behavior of the drug with the skin in the presence or absence of permeation enhancers, the actual situation of entering the skin, and the effect of permeation enhancers on the skin bilayer membrane structure. The main evaluation methods are to calculate the diffusion coefficient of the drug in the presence of different permeation enhancers, calculate the density distribution diagram of the skin bilayer membrane, drug molecules, and different permeation enhancers in the Z-axis direction, and calculate the order parameter of the skin bilayer membrane. Finally, the principle of the actual permeation enhancement effect of the permeation enhancer in the transdermal external preparation of the drug is explained, and the best permeation enhancer and the prescription ratio of the permeation enhancer are selected.

[0053] The best prescription of the transdermal drug preparation screened by the above method was used to evaluate the quality of the transdermal drug preparation. The modified Franz diffusion cell method was mainly used to study the in vitro permeation performance of the transdermal drug preparation and the reference preparation using the skin of isolated rats as the permeation model. The release rate was kinetically fitted, and finally the kinetic release law of the transdermal drug preparation was evaluated.

[0054] The present invention uses a gel commonly used in transdermal external preparations as a carrier and takes sinomenine gel as an example to illustrate the method of predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation in the present invention.

[0055] 1. Evaluate the incompatibility and formulation risks of drugs and excipients based on expert systems and use Materials Studio conducts all-atom molecular dynamics simulation to calculate solubility parameters and determine the optimal matrix for forming the sinomenine gel

[0056] 1.1. Evaluate the incompatibility of drugs and excipients and formulation risk based on the PharmDE expert system: It is necessary to provide the SMILES standard formula of the drug molecule to be tested. The molecular structure diagram of the sinomenine molecule and the SMILES standard formula (CN1CCC23CC(=O)C(=CC2C1CC4=C3C(=C(C=C4)OC)O)OC) are downloaded from the Pubchem database. Enter the expert system website to enter the PharmaDE online database when the computer is connected to the Internet, select the PREDICT module, first select DRUGINCOMPATIBILITY to analyze the incompatibility of sinomenine with various excipients, and exclude excipients that are incompatible with sinomenine after mixing; then select the FORMULATION RISK function to perform formulation risk assessment. When the incompatibility of drugs and excipients and the formulation risk are both low risk or medium risk in parallel excipients, the drug and excipients are theoretically compatible and safe when used in combination.

[0057] In the present invention, the risk levels such as "low risk" and "medium risk" are automatically defined by the expert system based on the special chemical structure contained in the molecular structure of the input drug, after matching and analyzing with the excipients existing in the database, and based on whether a chemical reaction occurs.

[0058] 1.2. Use Materials Studio software to calculate the solubility parameters (δ) of drugs and candidate excipients based on the all-atom molecular dynamics simulation method: In the calculation method, the compatibility of drugs and polymer carriers is usually evaluated by the solubility parameter (δ). Theoretically, the closer the solubility parameter values ​​of drugs and excipients are to the two substances or polymers, the better the compatibility after mixing. In addition, when the difference in solubility parameters of the two mixtures is less than 7.5, the two substances are compatible and have good compatibility. This part mainly uses the Amorphous Cell and Forcite modules of Materials Studio software. The gel forming matrix excipients screened in the previous step are all polymers. The polymers are made using ChemDraw 2020. When the solubility parameters are initially calculated, the degree of polymerization of the polymers is set to 2. After the theoretically optimal gel forming excipients are screened out, the solubility parameters of polymers with different degrees of polymerization are calculated for the optimal forming excipient polymer using the "alternative solubility parameter method". The difference in solubility parameters between the drug and the excipient is further calculated to confirm whether the solubility parameter difference is less than 7.5. The sinomenine molecule is downloaded from Pubchem. All molecules need to undergo four single-molecule structure optimizations when constructing a periodic system to eliminate unreasonable places in the structure, namely Smart, Steepest descent, and Conjugate gradient. The constructed simulated periodic unit cell is optimized four times in the same way. Then enable the Forcite module to calculate the solubility parameters. The calculation task of the equilibrium geometric configuration is performed using the NVT ensemble. The Ensemble is Velocity Scale, the time step is 0.25fs, the total number of simulation steps is 16000 steps, and the total simulation time is 4ps. After the task is completed, check the Restart checkbox, change the Ensemble to NVE, and keep other settings unchanged. Perform the calculation to generate the molecular dynamics simulation, and finally calculate the cohesive energy of the generated configuration. Change the task to Forcite Cohesive Energy Density to calculate the solubility parameter value of the result of the first frame of the generated configuration. The solubility parameters of carbomer polymers vary with the degree of polymerization. Figure 1 The solubility parameter values ​​and differences of sinomenine and candidate excipients are shown in Table 1. The solubility parameter of carbomer remains unchanged after the degree of polymerization is 30, which is about 24. The solubility parameter difference between sinomenine and carbomer polymer is the smallest and meets the standard that the two substances have good compatibility when the solubility parameter difference is less than 7.5.

[0059] Table 1 Solubility parameter values ​​δ and difference △δ of sinomenine and candidate excipients

[0060]

[0061] 1.3. Experimental verification of the prediction results of the expert system and molecular dynamics simulation: The prediction results of the expert system and molecular dynamics simulation were verified by differential scanning calorimetry combined with ordinary optical microscopy. All samples of the differential scanning calorimetry experiment were accurately weighed and packaged in aluminum pans. The test conditions were: the reference was water, the atmosphere was nitrogen, and the heating rate was 10℃·min -1 , the temperature range is 25℃~220℃ for DSC determination of reference substance and physically mixed sample, and the sample is a 1:1 mixture of drug and excipient. The DSC result graph was processed by software Origin 2018. The evaluation criteria are that if the characteristic melting peak shape of sinomenine does not change significantly and does not show obvious shift after physical mixing of sinomenine with each candidate excipient, it can be considered that the compatibility of the drug and the excipient after mixing is good. Ordinary optical microscope experimental verification involves the preparation of blank gel matrix. The specific operation is to weigh 1.0g of each polymer excipient, add 50ml of deionized water, stir evenly, and let it stand for 24h to allow it to fully swell. The swollen gel matrix is ​​stirred magnetically for 30min (500~600r·min -1 ), set aside; then it involves preparing samples, specifically weighing 5.0g of the swollen blank matrix, adding 3mg of sinomenine to each portion, controlling the stirring time to be the same, so that it is mixed evenly, and preparing samples for use. The samples are evenly spread on a glass slide and observed under a microscope. The microscope magnification is 4×10. The evaluation criteria mainly observe the solubility and distribution uniformity of sinomenine in each candidate matrix under the same conditions. The results of the differential scanning calorimetry test and the ordinary optical microscope experiment are shown in Figure 2 , Figure 3 The results are consistent with the results of simulation prediction screening.

[0062] 2. Use molecular dynamics simulation software GROMACS to study the effects of different prescription factors on the quality of sinomenine gel All-atom molecular dynamics simulations

[0063] The research objects involved in this section are the molecular structure models of sinomenine, carbomer, 1,2-propylene glycol, laurocapram, glycerol, ethanol, triethanolamine, menthol, ceramide, cholesterol, and free fatty acids. The molecular models in this section are all 3D molecular models. Except for carbomer polymers, the 3D molecular structure files are downloaded from Pubchem. The 3D molecular model structure of carbomer polymers is drawn using GaussianView5.0.

[0064] 2.1. The present invention involves the generation of the topological file of the force field selection before the formal molecular dynamics simulation: In this study, except for the matrix polymer, the atomic charges of all molecules except the matrix polymer were obtained by fitting the RESP charges using Multiwfn version 3.6 (dev) (Lu T, Chen F. Multiwfn: A multifunctional wavefunction analyzer [J]. Journal of computational chemistry, 2012, 33 (5): 580-592.) software

[34] , and then the final molecular topological file was generated with the help of Sobtop. For the high molecular polymer, Sobtop was used to call OpenBabel to identify the MMFF94 atomic charges to obtain the molecular topological file of the polymer.

[0065] 2.2. Construction of the initial simulation system for the effect of prescription factors on the quality of sinomenine gel based on molecular dynamics simulation: The number of molecules in each initial simulation system is calculated by the volume percentage of each prescription factor in the total gel. The construction of the initial simulation system mainly relies on the gmx insert-molecules, gmx editconf, and gmx solvate commands that come with the molecular dynamics simulation software GROMACS2022.2. The dynamic simulation process visualization and picture production are completed using Visual Molecular Dynamics 1.9.3 (VMD). Take the initial simulation system constructed during the carbomer polymerization degree screening as an example, Figure 4 As shown. Construction of skin lipid bilayer model: Skin keratin bilayer model (ceramide-sphingosine (CER, 24:0) and cholesterol (CHOL) and free fatty acids (FFA, 24:0), the ratio is 2:2:1, using the software SAMSON (http: / / samson-connect.net) for adjusting the structure and the software genmixmem (http: / / sobereva.com / 245) for generating the skin bilayer structure, and then using the GROMACS built-in command with VMD software to jointly construct the skin bilayer membrane model as shown. Figure 5 The difference in box size during the screening of formulation factors is that the pre-cell box size constructed during the investigation of polymerization degree is 10×10×20nm, the pre-cell box size of the system involved in the screening of various formulation factors is 14×14×18nm, and the box size of the system when screening the effect of ethanol concentration on matrix chain stability is 12×12×12.

[0066] Molecular dynamics simulation details: Three-dimensional periodic boundary conditions were applied to all simulation systems. First, 5000 steps of Steep energy minimization were performed with a time step of 1fs and a maximum force of no more than 1000KJ / mol. Then, a 10ns pre-equilibrium simulation was performed under the NPT ensemble. The temperature was increased from 0K to 298.15K within 100ps and then maintained at 298.15K. The temperature control method was V-rescale, the pressure was one standard atmosphere, and the pressure control method was Parrinello-rahman. The velocity Verlet algorithm was used to integrate the Newtonian equations of motion with a time step of 2fs. The particle mesh PME method was used to estimate the long-range electrostatic interaction. The LJ pair interaction was The evaluation is performed within the cutoff distance; finally, the finished product simulation is performed using the same method as the pre-equilibrium stage, with a simulation temperature of 298.15K and a simulation time of 200ns.

[0067] 2.3. Molecular dynamics simulation was performed on the initial simulation system for screening the formulation factors of the constructed sinomenine gel. The formulation factor screening objects of this part of the present invention are mainly the screening of the degree of polymerization of the carbomer polymer, the screening of the sinomenine concentration, the screening of the carbomer polymer concentration, the screening of the ethanol concentration, the screening of the glycerol concentration, and the screening of the penetration enhancer concentration. The number of molecules in the formulation composition of each formulation factor is shown in Tables 2 to 8.

[0068] Table 2 The number of molecules of each formulation factor in the simulation system with different polymerization degrees (unit: piece)

[0069]

[0070] In Table 2, Cp: carbomer, Sin: sinomenine, Pro: 1,2-propylene glycol, Gly: glycerol, Lau: laurocapram, Tro: triethanolamine, Eth: ethanol, H2O: water.

[0071] Table 3 The number of molecules of each prescription factor in the simulation system of different sinomenine concentrations (unit: piece)

[0072]

[0073] Table 4 Number of molecules of each formulation factor in the simulation system with different polymer matrix concentrations (unit: piece)

[0074]

[0075] Table 5 Number of molecules of each formulation factor in different concentration solvent simulation systems (unit: piece)

[0076]

[0077] Table 6 The number of molecules of each formulation factor in the simulation system with different ethanol concentrations (unit: piece)

[0078]

[0079]

[0080] Table 7 The number of molecules of each formulation factor in the simulation system with different glycerol concentrations (unit: piece)

[0081]

[0082] Table 8 The number of molecules of each formulation factor in the simulation system of different penetration enhancer concentrations (unit: piece)

[0083]

[0084] After the simulation is completed, the simulation trajectory is periodically corrected and the simulation process is analyzed for equilibrium. The main evaluation criteria used are the total energy of the analysis system and the root mean square displacement of the carbomer polymer macromolecules. The results of the equilibrium analysis are as follows: Figure 6 As shown, the equilibrium simulation trajectory is then selected and saved for calculation and analysis of evaluation parameters, including: analysis of the radius of gyration of the carbomer polymer, the root mean square displacement, the density distribution diagram of the system and the carbomer and water solvent in the Z-axis direction, the diffusion coefficient of the drug in the gel matrix network, the order parameters of the ceramide and free fatty acid heads, and based on the analysis data, the optimal concentration of the best prescription factors in the preparation of sinomenine gel is evaluated, so as to quickly and accurately screen out the best sinomenine gel prescription.

[0085] 2.4. In this embodiment, the simulation system for the degree of polymerization screening of carbomer polymers was subjected to RMS displacement analysis. Figure 7 As shown in FIG. 1 , the average Rg and RMSD of carbomers with different polymerization degrees vary with the polymerization degree, the DA varies with the polymerization degree, the density distribution of carbomer polymers varies with the polymerization degree, and the density distribution of solvent water varies with the polymerization degree. Figure 8As shown. The radius of gyration of carbomer polymer fluctuates more and more with the increase of polymerization degree. The larger the radius of gyration, the more obvious the characteristics of carbomer polymer swelling into gel network structure in the system. Therefore, when screening the degree of polymerization, carbomer with a larger degree of polymerization should be considered first to replace the real carbomer polymer. Further analysis of the results of the average Rg and RMSD of carbomers with different polymerization degrees with the degree of polymerization showed that there was no obvious rule. Further analysis of the results of DA with the degree of polymerization, the results of the density distribution of carbomer polymer with the degree of polymerization, and the results of the density distribution of solvent water with the degree of polymerization showed that when the degree of polymerization was 50, sinomenine had a better release rate in the gel network and the density distribution of carbomer polymer and solvent water in the Z-axis direction was inversely proportional, indicating that when the degree of polymerization was 50, the carbomer polymer had a better polymer gel network water storage effect, and it also met the standard that a larger degree of polymerization was more suitable for replacing the real carbomer polymer.

[0086] 2.5. In this embodiment, the simulation system for sinomenine screening was analyzed for the changes in the average Rg value and RMSD value of the polymer matrix chain in different sinomenine concentration systems over time, the changes in the diffusion rate of drug molecules in different systems over time, and the density distribution diagram of the polymer matrix chain network in the Z-axis direction. Fig. 9 As shown. The results of the average Rg and RMSD of carbomer polymers in the sinomenine simulation system with different concentrations of sinomenine were analyzed as the concentration of sinomenine changes, and no obvious pattern was found. Further analysis of the changes in the diffusion rate of drug molecules in different systems over time and the density distribution of the polymer matrix chain network in the Z-axis direction showed that the release rate of sinomenine in the gel network increased with the increase of sinomenine concentration, but when the concentration of sinomenine was greater than 3%, the release rate of sinomenine in the gel network dropped sharply, and the decrease was obvious, indicating that the optimal concentration of sinomenine was 3%, and the density distribution of carbomer polymers at different sinomenine concentrations on the Z-axis did not change significantly, indicating that the concentration of sinomenine did not affect the network structure of carbomer gel polymers.

[0087] 2.6. In this example, the simulation system for screening the concentration of carbomer polymer is analyzed by analyzing the change of Rg of 5 random carbomer polymers over time in the gelling agent simulation system with different concentrations. Fig.10 As shown in the figure, the average Rg value and RMSD value of the carbomer polymer in different carbomer polymer concentrations vary with the carbomer polymer concentration, the DA varies with the carbomer polymer concentration, and the average density distribution diagram of the gel system with different carbomer polymer concentrations varies in the Z-axis direction. Fig.11As shown. In the results of the Rg value change trend of five random matrix chains in different carbomer polymer concentration systems over time, the gyration radius of the carbomer polymer fluctuates greatly with the increase of the carbomer polymer concentration, indicating that the carbomer polymer has expanded when the carbomer polymer concentration is 0.5% to 2.5%, and has a gel network structure. Further analysis of the Rg value and RMSD value of the carbomer polymer in different concentrations with the change of carbomer polymer concentration, the change of DA with matrix concentration, and the gel of different concentrations The results of the change of the average density distribution diagram of the system in the Z-axis direction, the Rg value and the RMSD value are consistent with the change of the carbomer polymer concentration. With the increase of the concentration of the carbomer polymer, the release rate of sinomenine in the gel network gradually slows down, which is not conducive to the release of the drug. The analysis of the change of the average density distribution diagram of the gel system with different concentrations in the Z-axis direction found that when the concentration of the carbomer polymer is 1% to 2%, the system has good uniformity. Combined with the preliminary experimental results, when the concentration of the carbomer polymer is 1%, the prepared sinomenine gel has good quality.

[0088] 2.7. In this embodiment, not only the concentration of ethanol was screened separately, but also the optimal ethanol concentration was screened based on the results of the ethanol concentration screening to determine the effect of ethanol concentration on the quality of sinomenine gel. The separate ethanol concentration screening mainly analyzed the results of the Rg value changes of 5 random polymer matrix chains over time in different ethanol concentration systems, such as Fig.12 As shown in the figure, the fluctuation range of the gyration radius is too large, indicating that its stability in the solvent is poor. Ethanol with a concentration of 20% to 30% should be selected for further screening of the optimal ethanol concentration. Therefore, the diffusion coefficient of sinomenine in the gel network in the simulation system with an ethanol concentration of 20% to 30% and the RDF results of the interaction between sinomenine and carbomer polymer molecules in the simulation system with an ethanol concentration of 20% to 30% were analyzed. Fig.13 As shown in the figure, the diffusion coefficient increases with the increase of ethanol concentration, indicating that the higher the ethanol content, the more conducive it is to the release of drugs from the gel network. The results of the intermolecular radial distribution function of sinomenine and carbomer polymer molecules are consistent with the diffusion coefficient results. With the increase of ethanol concentration, the interaction strength between sinomenine and carbomer polymer gradually decreases, which is conducive to the release of drugs from the gel network. The optimal ethanol concentration screened is 30%.

[0089] 2.8. In this embodiment, the concentration screening simulation system of moisturizing agent glycerol is used to analyze the trend of Rg values ​​of five random polymer matrix chains in different glycerol concentration gel systems over time. Fig.14As shown in the figure, the diffusion rate results of sinomenine in gel systems with different glycerol concentrations and the RDF results of the interaction between the oxygen atoms on the hydroxyl groups of sinomenine and the hydrogen atoms in water molecules in systems with different glycerol concentrations are analyzed. Fig.15 As shown. Without glycerol, the fluctuation range of the gyration radius of the carbomer polymer is small, while the fluctuation range of the carbomer polymer in the simulation system with glycerol is large, indicating that the addition of glycerol will lead to the instability of the carbomer polymer gel network structure, so the concentration of glycerol should not be too high. The release rate of sinomenine in the gel network and the RDF results of the interaction between the oxygen atoms on the hydroxyl group of sinomenine and the hydrogen atoms in the water molecules also show that the concentration of glycerol should not be too high. Combined with the recommended usage of glycerol in the literature and the predicted results here, the optimal concentration of glycerol is 5%.

[0090] 2.9. In this embodiment, the concentration screening simulation system of the moisturizing agent glycerol is analyzed by analyzing the snapshot results of the molecular dynamics simulation trajectory of the 0th frame and the last frame saved in the time of 0 to 200 ns when 1,2-propylene glycol and laurocapram are used alone and in combination, and menthol is used alone, the density distribution diagram of different prescription factors in the system and the lipid bilayer along the Z axis of the box, the order parameter results of the head and tail of CER and FFA in the lipid bilayer membrane at different penetration enhancer concentrations, and the average diffusion rate results of sinomenine in the system with the same penetration enhancer concentration at 0 to 200 ns, such as Fig.16 , 17 , 18, and 19. The simulation snapshot results show that penetration enhancers generally have a certain effect on promoting the penetration of drugs into the skin, but the effects are different. The simulation trajectory snapshots of the system at 0 to 200 ns when different concentrations of 1,2-propylene glycol and laurocapram are used alone or in combination show that at this penetration enhancer concentration, no matter what concentration of sinomenine is, sinomenine molecules are not all in contact with the skin, but are aggregated and distributed in the space on both sides of the skin, resulting in a part of the drug being unable to contact the skin and being absorbed, which is not conducive to the drug reaching the therapeutic concentration. When menthol is added, the simulation snapshots of 0 to 200 ns are analyzed, and the combination of sinomenine molecules and menthol penetration enhancers has a synergistic penetration-promoting effect, which is conducive to the drug penetrating the skin to achieve a therapeutic effect. The results of the density distribution diagram of different prescription factors and lipid bilayers in the system along the Z axis of the box also show that when menthol is added, the number of drugs entering the skin is large. The larger the difference in the order parameter, the greater the degree of skin damage, which is more conducive to the drug entering the skin and exerting its efficacy to achieve a therapeutic effect. Analysis of the average diffusion rate of sinomenine in different penetration enhancer concentration systems from 0 to 200 ns revealed that the drug penetration enhancer also has an optimal penetration enhancer concentration. The addition of menthol is beneficial to the release of sinomenine. When the menthol concentration is 10%, sinomenine has a better release rate.

[0091] 3. The optimal prescription of sinomenine gel selected by expert system and molecular dynamics simulation method was analyzed. Quality evaluation of sinomenine gel

[0092] This example evaluates the quality of sinomenine gel by using the best formulation of sinomenine gel screened by using the expert system and molecular dynamics simulation method. The main aspects of the evaluation are appearance observation and in vitro transdermal performance. The formulation composition and preparation method of sinomenine gel and reference preparation are as follows:

[0093] (1) The composition of sinomenine gel:

[0094] Carbomer 940: 1g, sinomenine: 3g, anhydrous ethanol: 30ml, glycerol: 5g, menthol: 10g, triethanolamine: appropriate amount, add distilled water to 100g.

[0095] Preparation method:

[0096] Accurately weigh a certain mass of molding matrix auxiliary materials, place it in a 100 ml beaker, add an appropriate amount of distilled water, let it stand, swell overnight, and set aside; accurately weigh free sinomenine in a 100 ml beaker, add anhydrous ethanol to the beaker, put in a rotor with a diameter of 2 mm, seal the cup mouth with plastic wrap to prevent ethanol volatilization, and then place it in a 50°C magnetic stirrer and stir in a water bath for 5 minutes to fully dissolve the free sinomenine; pour the fully dissolved free sinomenine solution into the swollen molding matrix, stir evenly with a glass rod, then add other additives in turn, continue to stir evenly, and finally add triethanolamine, adjust the pH to 6.5-7, add water to make up the mass of sinomenine gel to 100 g, and finally stir the sinomenine gel in a clockwise direction to evenly store at room temperature for later use.

[0097] (2) Composition of the reference preparation:

[0098] Carbomer 980: 2g, sinomenine: 3g, anhydrous ethanol: 30ml, glycerol: 4ml, 1,2-propylene glycol: 10ml, laurocapram: 2g, triethanolamine: 2g, make up with distilled water to 100g.

[0099] Preparation method:

[0100] Accurately weigh 2 g of Carbomer 980, place it in a 100 ml beaker, add appropriate amount of distilled water, let it stand and swell overnight for use; accurately weigh 3 g of free sinomenine in a 100 ml beaker, add 30 ml of anhydrous ethanol to the beaker, put in a rotor with a diameter of 2 mm, seal the cup mouth with plastic wrap to prevent ethanol volatilization, and then place it in a 50 ° C magnetic stirrer and stir in a water bath for 5 min to fully dissolve the free sinomenine; pour the fully dissolved free sinomenine solution into the swollen Carbomer 980 matrix, stir evenly with a glass rod, and then add glycerol, 1,2-propylene glycol, and lauryl azone in sequence, continue to stir evenly, and finally add triethanolamine, adjust the pH to 6.5-7, add water to make up the mass of the reference preparation to 100 g, and finally stir the reference preparation in a clockwise direction and store at room temperature for use.

[0101] The appearance of the prepared sinomenine gel and the reference preparation is as follows: Fig. 20 As shown. The sinomenine gel prepared using the sinomenine gel prescription predicted by the expert system and molecular dynamics simulation method is more transparent and has a smoother texture than the reference preparation. It is easy to dissolve on the skin and is absorbed quickly, with a cool feeling. The in vitro transdermal performance was investigated mainly using the modified Franz diffusion cell method. The skin of isolated rats was used as a transdermal permeation model to study the in vitro transdermal performance of sinomenine gel and the reference preparation. The release rate was kinetically fitted, and finally the kinetic release law of sinomenine gel was evaluated. The cumulative permeation amount of sinomenine per unit area (Qn, μg·cm) of sinomenine gel and the reference preparation over a period of 48 hours was calculated. -2 ), the results showed that the cumulative permeation amount of sinomenine per unit area of ​​sinomenine gel and reference preparation was (7945.44±233.62)μg·cm -2 、(6258.25±468.67)μg·cm -2 And the in vitro permeation cumulative permeation curve of sinomenine gel and reference preparation was plotted with Qn versus sampling time (t). The results are as follows Fig.21 As shown. The cumulative permeation curve was also fitted with kinetic linear fitting, mainly zero-order kinetics, first-order kinetics, and Higuchi equation fitting. The fitting results showed that the kinetic processes of sinomenine gel and reference preparation were both first-order kinetic processes, and their release had a sustained release effect. The equation expressions were y1=4440.67*(1-e 0.030t ), R2=0.9959, y2=3976.29*(1-e 0.024t ), R2 = 0.9978, the slope is the permeation rate Js (μg·cm -2 ·h -1 ), so the skin penetration rates of sinomenine gel and reference preparation were (4440.67±250.40) μg·cm -2 ·h-2 、(3976.29±226.91)μg·cm -2 ·h -2 The sinomenine gel selected by the expert system and molecular dynamics simulation method has been improved in appearance and bioavailability. The inventive method is reliable and accurate in the prescription screening of transdermal topical preparations.

[0102] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for predicting and screening a transdermal external preparation prescription based on an expert system and molecular dynamics simulation, comprising the following steps: A) Use an expert system to evaluate the incompatibility of the drug and each excipient, and exclude excipients that are incompatible with the drug; then use the expert system to perform a formulation risk assessment on the drug and the excipients that have passed the incompatibility assessment, and screen out candidate excipients with medium or low risk; B) using Materials Studio software based on the all-atom molecular dynamics simulation method to calculate the solubility parameter δ of the drug and each candidate excipient, and screening out the excipient with the smallest difference in solubility parameter between the drug and the excipient and δ < 7.5; C) obtaining molecular structure files of drugs and excipients, selecting a GAFF force field compatible with the Amber force field, creating a box, adding drug molecules and excipient molecules according to the volume percentage of the components in the transdermal external preparation prescription to the created box to form an initial simulation system, sequentially performing energy minimization and pre-balancing on the initial simulation system, and after the pre-balancing, using GROMACS software to perform 200 ns finished product simulation on the simulation system under different prescription factors, evaluating according to the simulation results, and obtaining the desired preparation prescription; D) performing quality evaluation on the preparation prescription obtained in step C).

2. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: In the step B), the solubility parameter δ is calculated specifically according to the following steps: The molecular structures of drugs and excipients were obtained, and the COMPASS force field was selected. The single-molecule structure energy was minimized four times using Smart, Steepest descent, Conjugate, and gradient methods for all molecular structures. The drug molecules and excipient molecules after energy minimization were used to construct a pure substance simulation periodic unit cell. The simulated periodic unit cell was minimized four times using Smart, Steepest descent, Conjugate, and gradient methods, and then the Forcite module was enabled to calculate the solubility parameters.

3. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 2, characterized in that: The enabling of the Forcite module to calculate the solubility parameters specifically includes: The NVT ensemble is first used to optimize the simulated periodic unit cell after energy minimization, and then the NVE ensemble is used to continue the optimization. Molecular dynamics simulation is performed to generate a molecular dynamics simulation configuration, and finally the cohesive energy calculation is performed on the generated configuration to generate the result of the first frame of the configuration to calculate the solubility parameter value.

4. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: The results of the expert system in step A) and the molecular dynamics simulation in step B) are experimentally verified using differential scanning calorimetry experiments and optical microscopy.

5. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: The step C) also includes constructing a skin lipid bilayer model to simulate the molecular dynamics process between the simulation system containing drugs and excipients and the skin surface layer; The skin lipid bilayer model includes ceramide-sphingosine, cholesterol and free fatty acids in a 2:2:1 molar ratio.

6. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: In the all-atom molecular dynamics simulation using GROMACS software in step C), three-dimensional periodic boundary conditions were applied to all simulation systems. Steep energy minimization was first performed, and then a pre-equilibrium simulation was performed under the NPT ensemble. Finally, a 200 ns finished product simulation was performed according to the pre-equilibrium simulation conditions.

7. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: The different prescription factors in step C) include: screening of drug concentration and screening of concentrations of various excipients.

8. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 7, characterized in that: When the drug and / or excipients include polymers, the different formulation factors in step C) also include screening of the drug polymerization degree and / or excipient polymerization degree.

9. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: The simulation results in step C) include the radius of gyration Rg, root mean square displacement RMSD, density distribution, radial distribution function RDF, diffusion coefficient DA and order parameter.

10. The method for predicting and screening transdermal external preparation prescriptions based on expert system and molecular dynamics simulation according to claim 1, characterized in that: The quality evaluation in step D) includes appearance observation and in vitro skin permeation performance investigation.

Images