Method for screening potential functional raw materials of cosmetics by establishing mathematical model based on molecular docking and physical and chemical properties of substances

Abstract

The invention discloses a method for screening potential efficacy raw materials of cosmetics by establishing a mathematical model based on molecular docking and physical and chemical properties of substances, and the method comprises the following steps: S1, aiming at semi-flexible molecular docking of a specific efficacy target, including pretreatment and simulated docking of a compound ligand and a target receptor; s2, querying and selecting cosmetic application related indexes in a database, wherein the specific indexes are molecular weight, the number of hydrogen bond receptors and donors, fat-water distribution coefficient, topological polar surface area and skin permeability; and S3, alternative compound screening, including data preprocessing, definition of optimal / inferior values and median values, primary screening of principal component analysis, weight coefficient determination by an entropy weight method and Topsis comprehensive evaluation. According to the method, reasonable non-experimental type indexes are found, so that the screening efficiency is effectively improved, the cost is effectively reduced, the scientific large-batch raw material screening requirement can be met, and besides the screening purpose, the potential size of alternative compounds can be predicted and quantified.

Description

Technical Field

[0001] The present invention belongs to the field of cosmetic raw material development and verification, and specifically relates to a method for screening potential cosmetic raw materials by establishing a mathematical model based on molecular docking and the physicochemical properties of substances. Background Art

[0002] When verifying the efficacy of cosmetics or raw materials, a single in vivo or in vitro efficacy evaluation index is often used as part of the evidence for cosmetic efficacy claims, but this type of investigation often has a single factor and the evaluation is too one-sided. This requires us to examine as many factors as possible when developing efficacy raw materials. On the one hand, the increase in the number of factors to be examined undoubtedly poses a challenge to the efficiency and cost of experimental efficacy verification. On the other hand, simultaneously examining multiple related indicators rather than a single indicator also creates new difficulties for the reasonable and comprehensive evaluation of the potential of raw materials.

[0003] However, authoritative compound property databases and molecular docking technology are widely used in the field of drug development, but are still rarely used in cosmetics research and development. Therefore, it is of great significance to find a method to efficiently obtain efficacy-related indicators and establish a method that can quickly screen and evaluate the application potential of raw materials for the development of cosmetic raw materials. Summary of the invention

[0004] The present invention aims to at least solve the technical problems existing in the above-mentioned prior art, and provides a method for screening potential cosmetic raw materials by finding reasonable non-experimental indicators to effectively improve screening efficiency and reduce costs, which can meet the needs of scientific large-scale raw material screening. In addition to the screening purpose, it can also predict and quantify the potential size of alternative compounds and establish a mathematical model based on molecular docking and the physical and chemical properties of materials to screen potential cosmetic raw materials.

[0005] The technical solution of the present invention is a method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances. The steps are as follows:

[0006] S1. Semi-flexible molecular docking for specific efficacy targets, including pre-processing and simulated docking of compound ligands and target receptors;

[0007] S2. Query and select indicators related to cosmetic applications in the database. The specific indicators are as follows: molecular weight (MV), number of hydrogen bond acceptors and donors (Hacc & Hdor), lipid-water partition coefficient (XlogP), topological polar surface area (TPSA) and skin permeability (logK p );

[0008] S3. Screening of candidate compounds, including data preprocessing, definition of optimal / inferior values ​​and median values, preliminary screening by principal component analysis, determination of weight coefficients by entropy weight method, calculation of optimal / inferior solution distance and Topsis comprehensive evaluation.

[0009] Preferably, the pre-treatment of the compound ligand and the target receptor in step S1 refers to: using Chem 3D to simulate the optimal energy conformation of the compound molecule; using AutodockTool to perform hydrogenation, charge calculation and rotation bond definition steps on the optimized conformation; using Pymol to remove heteroatoms such as homologous chains, unique ligands and water molecules from the target receptor, and then performing hydrogenation and charge calculation steps.

[0010] Preferably, the simulated docking in step S1 refers to: importing the compound ligand and the target receptor into the Autodock Tool to start the simulated binding process, adjusting the docking Box parameters to ensure that the docking sites are covered, starting the docking, and outputting the docking results, which include the best binding site and the lowest binding energy, so as to obtain the binding activity index of the compound and the target receptor.

[0011] Preferably, the query of cosmetic application-related indicators in the database in step S2 specifically refers to: the five drug-like principles, which are the five basic rules for screening drug-like molecules proposed by medicinal chemist Christopher Lipinski. Compounds that meet these rules often have better pharmacokinetic properties and higher bioavailability during metabolism in the body; combined with some differences in properties between cosmetics and oral drugs during use, some relevant indicators are simplified or added to ensure that compounds with potential as cosmetic raw materials can be fully evaluated and screened.

[0012] Preferably, the indicators in step S2 are as follows:

[0013] Molecular weight (MV): Molecular weight can usually be accurately calculated. Generally speaking, as the molecular weight increases, the lipophilicity of the compound also increases. For cosmetic raw materials, substances with smaller molecular weight are easier to be absorbed by the body and thus exert their effects;

[0014] Number of hydrogen bond acceptors and donors (Hacc&Hdor): Hydrogen bonds are a type of weak intermolecular interaction. Among molecules with more than 5 hydrogen bond donors and / or more than 10 hydrogen bond acceptors, most are glycosylated molecules or amino acids related to natural products. The values ​​of hydrogen bond acceptors and hydrogen bond donors can affect the lattice energy and melting point, and thus affect the solubility. Some intramolecular hydrogen bonds are also related to the permeability of substances.

[0015] Lipid-water partition coefficient (XlogP): The lipid-water partition coefficient is the concentration ratio of a compound when it reaches thermodynamic equilibrium between the lipid phase and the water phase. Generally, the larger the lipid-water partition coefficient of a compound, the more soluble it is in fat, and vice versa. Substances that are easily soluble in fat are lipophilic or hydrophobic in the body, and the phenomenon of being easily soluble in water is called hydrophilicity. Lipophilicity / hydrophilicity directly affects the permeability / solubility of raw materials, and as the basic properties of cosmetic raw materials, it affects their efficacy to a certain extent.

[0016] Topological polar surface area (TPSA): The topological polar surface area of ​​a molecule refers to the surface area of ​​the non-atomic molecular side blocks around the polar atoms (such as oxygen, nitrogen, etc.) in the molecule. TPSA can be used to evaluate the solubility, membrane permeability, and protein binding properties of the compound. When the TPSA is small, it usually means lower compound solubility and higher membrane permeability. As the molecular polar surface area increases, the possibility of binding to proteins also increases.

[0017] Skin permeability (logK p ): Generally speaking, the larger the logKp value, the stronger the drug's permeability on the skin. According to the special use of cosmetics, skin permeability is an important reference for measuring compounds as cosmetic raw materials. Although we pursue good absorption effects in cosmetics, we also need to pay attention that the larger the value, the better. Because too high skin permeability may cause the drug to be absorbed too quickly on the skin and quickly enter the blood circulation system, thereby increasing the risk of adverse reactions;

[0018] The above indicators are common and clear physical and chemical properties of compounds, and skin permeability data can be retrieved through publicly available online databases.

[0019] Preferably, in step S3, data preprocessing means: before actually selecting indicators for comprehensive evaluation, the evaluation standards of each indicator are often different. In real life, there are extremely large indicators such as performance and growth rate; extremely small indicators such as cost and defect rate; intermediate indicators such as pH value and interval indicators with reasonable numerical range of body temperature. In general, data of the same type are convenient for subsequent calculations, so each indicator is positively processed before comprehensive evaluation to obtain a data matrix containing all indicators of the compounds to be screened;

[0020] At the same time, in order to eliminate the influence of different dimensions on the comprehensive evaluation, each group of data matrices after forward transformation is uniformly standardized to construct a standardized data matrix;

[0021]

[0022] The data contained in the matrix Z are extremely large data after positive and standardized processing, n is the number of evaluation indicators introduced, and m is the number of evaluation candidate raw materials.

[0023] Preferably, in step S3, the preliminary screening by principal component analysis refers to: importing the preprocessed data matrix containing the best / worst / median solution into SIMCA 14.1 data processing software, taking the compound to be screened as the Principal ID, and using unsupervised PCA principal component analysis to reduce the dimension of the multidimensional data; selecting alternative raw materials between the median and the optimal solution for further screening.

[0024] Preferably, in step S3, the steps of determining the weight coefficient using the entropy weight method are as follows:

[0025] First, calculate the proportion of each indicator P:

[0026]

[0027] Then, in order to quantify the uncertainty of a random variable, the information entropy E of each indicator is calculated:

[0028]

[0029] Correspondingly, the greater the information entropy of an indicator in this screening process, the greater its weight:

[0030]

[0031] After obtaining the weight coefficient, construct the weighted data matrix

[0032]

[0033] Preferably, in step S3, calculating the optimal / worst solution distance means:

[0034] Obviously, there is an ideal optimal solution Z in the data matrix + and the worst solution Z-(compound); use the largest number of each evaluation index in the index after data preprocessing to form the ideal optimal solution matrix, and vice versa:

[0035] Z + =[Z1 + ,Z2 + ,...,Z m + ]=[max{Z 11 ,Z 21 ,...Z n1},max{Z 12 ,Z 22 ,...,Z n2},...,max{Z 1m ,Z 2m ,...Z nm}];

[0036] Z- =[Z1 - ,Z2 - ,...,Z m - ]=[min{Z 11 , Z 21 , …Z n1},min{Z 12 , Z 22 , …Z n2}, ..., min{Z 1m , Z 2m , …Z nm}];

[0037] After obtaining the optimal and worst solutions, calculate the distance D from the optimal solution + And the distance to the worst solution D-:

[0038]

[0039] Preferably, in step S3, Topsis comprehensive evaluation refers to:

[0040] The application potential score S of each compound is calculated according to the Topsis comprehensive evaluation method: The formula is given by Deformation obtained.

[0041] The beneficial effects of the present invention are:

[0042] 1. The present invention effectively improves screening efficiency and reduces costs by finding reasonable non-experimental indicators. Semi-flexible molecular docking simulates the process of small molecules binding to protein receptors, which is the same as the basic principle of target publicity commonly used in current cosmetic efficacy publicity. While examining the binding activity of small molecule candidate raw materials, combined with authoritative network database indicators, it can greatly make up for the shortcomings of molecular docking;

[0043] 2. The present invention can meet the needs of scientific large-scale raw material screening. It uses a multivariate statistical analysis method - principal component analysis (PCA) to map multiple candidate raw materials containing multiple dimensions (multiple evaluation indicators) into a low-dimensional space through linear transformation. This transformation retains the main changes in the data as much as possible, and can screen out potential compounds with application potential in large quantities.

[0044] 3. In addition to the screening purpose, the present invention can also predict and quantify the potential size of candidate compounds. By establishing the Topsis comprehensive evaluation model, the results after PCA statistical screening are evaluated again in detail. This evaluation scores the raw materials after the initial screening through normalization and standardization, and performs decision analysis and sorting on the compounds containing multi-dimensional data, which can further rank the potential size of the compounds in detail. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 Schematic diagram of PCA analysis of candidate compounds in Example 1 of the present invention. DETAILED DESCRIPTION

[0046] In order to make the invention purpose, technical solution and technical effect of the present invention clearer, the present invention is further described in detail below in conjunction with specific implementation methods. It should be understood that the specific implementation methods described in this specification are only for explaining the present invention, not for limiting the present invention.

[0047] Example 1

[0048] Filter Examples

[0049] Taking the screening of anti-inflammatory and soothing cosmetic raw materials as an example, common natural organic compounds were randomly selected from 12, including common natural active substances, such as polyphenols, flavonoids, alkaloids, etc. Before screening, these candidate compounds were used to perform molecular docking with the common inflammatory targets inducible nitric oxide synthase (iNOS) & cyclooxygenase-2 (COX-2) to predict binding activity, and the corresponding physicochemical properties were queried in the SwissADME database, including: molecular weight (MV), number of hydrogen bond donors / acceptors (Hdor / Hacc), lipid-water partition coefficient (xlog P), topological polar molecular area (TPSA) and skin permeability (Log Kp), as shown in Table 1.

[0050]

[0051] Table 1 Binding activity and physicochemical properties of candidate compounds

[0052] 1. Data positive processing

[0053] Before actually selecting indicators for comprehensive evaluation, the evaluation criteria of each indicator are often different. For example, there are very large indicators such as performance and growth rate, very small indicators such as cost and defect rate, intermediate indicators such as pH value, and interval indicators with reasonable numerical ranges. In general, the same type of data is convenient for subsequent calculations, so during screening, each indicator is positively processed (converted into very large data) to obtain a data matrix containing the indicator data of all compounds to be screened, such as

[0054] As shown in Table 2.

[0055]

[0056] Table 2 Data of indicators after positive processing

[0057] 2. Data Standardization

[0058] In order to eliminate the influence of different dimensions on the comprehensive evaluation, each group of data matrices after forward transformation are uniformly standardized to construct a standardized data matrix, as shown in Table 3.

[0059]

[0060] Table 3 Data of various indicators after standardization

[0061] 3. Preliminary screening by principal component analysis

[0062] The standardized data matrix containing the best / worst / median solutions was imported into SIMCA, and the compounds to be screened were used as Primary IDs. Unsupervised PCA principal component analysis was used to reduce the dimensionality of multidimensional data, such as Figure 1 The alternative raw materials between the median value and the optimal solution were selected for further screening, as shown in Table 4.

[0063]

[0064] Table 4 Compounds after PCA analysis screening

[0065] 4. Determine the weight coefficient of each indicator by entropy weight method

[0066] After calculating the proportion, finding the information entropy of each group of indicators and calculating the weight coefficient of each indicator, the weight coefficient is obtained and the weighted matrix is ​​constructed, as shown in Table 5:

[0067]

[0068] Table 5 Compound index data after weighting by entropy weight method coefficient

[0069] 5. Calculate the optimal / worst solution distance

[0070] Define the optimal / worst solution and calculate the distance to the optimal / worst solution, as shown in Table 6:

[0071] Name <![CDATA[D + ]]> <![CDATA[D - ]]> Berberine 0.07752 0.049748 Licorice 0.054951 0.040598 Chrysophanol 0.050855 0.045975 Aloe-emodin 0.06616 0.036724 <![CDATA[ α- Bisabolol]]> 0.037187 0.062626 Vanillic acid 0.041281 0.077367 Syringic acid 0.046579 0.064081

[0072] Table 6 Distance between compounds and optimal / worst solutions

[0073] 6.Topsis comprehensive evaluation

[0074] According to Topsis comprehensive evaluation S: The results were then normalized and sorted to obtain the ranking of raw material application potential, as shown in Table 7:

[0075] <![CDATA[ Num ]]> Name Unnormalized Normalization 1 Vanillic acid 1.951514 0.239427 2 α-Bisabolol 1.746715 0.214301 3 Syringic acid 1.439839 0.176651 4 Chrysophanol 0.950001 0.116554 5 Licorice 0.779407 0.095624 6 Berberine 0.691494 0.084838 7 Aloe-emodin 0.591799 0.072607

[0076] Table 7 Application potential and ranking of compounds comprehensively evaluated by Topsis

[0077] 2. Experimental Results and Discussion

[0078] As shown in the results, vanillic acid and α-bisabolol have high application potential. It can be found that the molecular docking results of the two have good results. At the same time, the molecular weight is less than 500, the number of hydrogen bond acceptors is less than 5, the number of hydrogen bond donors is less than 10, and the lipid-water partition coefficient is less than 5, which meets the "five" principles of drug-like substances proposed by Lipinski; at the same time, the polar molecular area shows that the two have good lipophilicity while having a certain solubility and can be well absorbed. Finally, the skin permeability data show that the two may have good transdermal absorption ability.

[0079] It is worth mentioning that although berberine is widely used as an anti-inflammatory and antibacterial drug in clinical practice and shows good binding activity in molecular docking, it is not difficult to see from the lipid-water partition coefficient and topological polar molecular area that its solubility or membrane permeability is not outstanding, and the lipid-water partition coefficient is even no longer in the range of values ​​recommended by Lipinski's "five" principle. Therefore, the result of this screening of berberine's low cosmetic application potential is reasonable. In addition, in comparison, the raw material that is more widely used in cosmetics is its common derivative dihydroberberine. In comparison, dihydroberberine has higher bioavailability and safety, which also confirms this screening result.

Claims

1. A method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances, characterized in that: Here are the steps: S1. Semi-flexible molecular docking for specific efficacy targets, including pre-processing and simulated docking of compound ligands and target receptors; S2. Query and select indicators related to cosmetic applications in the database. The specific indicators are as follows: molecular weight (MV), number of hydrogen bond acceptors and donors (Hacc & Hdor), lipid-water partition coefficient (XlogP), topological polar surface area (TPSA) and skin permeability (log K p ); S3. Screening of candidate compounds, including data preprocessing, definition of optimal / inferior values ​​and median values, preliminary screening by principal component analysis, determination of weight coefficients by entropy weight method and Topsis comprehensive evaluation.

2. The method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances according to claim 1, characterized in that: The compound ligand and target receptor pretreatment in step S1 refers to: using Chem 3D to simulate the optimal energy conformation of the compound molecule; using AutodockTool to perform hydrogenation, charge calculation and rotation bond definition steps on the optimized conformation; using Pymol to remove heteroatoms such as homologous chains, unique ligands and water molecules from the target receptor, and then performing hydrogenation and charge calculation steps.

3. The method for screening potential cosmetic ingredients based on molecular docking and physical and chemical properties of substances by establishing a mathematical model according to claim 1, characterized in that: The simulated docking in step S1 refers to: importing the compound ligand and the target receptor into the Autodock Tool to start the simulated binding process, adjusting the docking Box parameters to ensure that the docking sites are covered, starting the docking, and outputting the docking results, which include the best binding site and the lowest binding energy, so as to obtain the binding activity index of the compound and the target receptor.

4. The method for screening potential cosmetic ingredients based on molecular docking and physical and chemical properties of substances by establishing a mathematical model according to claim 1, characterized in that: In step S2, the database is queried for indicators related to cosmetic applications, specifically referring to: the "five" principles of drug-like molecules, which are the five basic rules for screening drug-like molecules proposed by medicinal chemist Christopher Lipinski. Compounds that meet these rules often have better pharmacokinetic properties and higher bioavailability during metabolism in the body; combined with some differences in properties between cosmetics and oral drugs during use, some relevant indicators are simplified or added to ensure that compounds with potential as cosmetic raw materials can be fully evaluated and screened.

5. The method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances according to claim 1, characterized in that: The specific indicators in step S2 are as follows: Molecular weight (MV): Molecular weight can usually be accurately calculated. Generally speaking, as the molecular weight increases, the lipophilicity of the compound also increases. For cosmetic raw materials, substances with smaller molecular weight are easier to be absorbed by the body and thus exert their effects; Number of hydrogen bond acceptors and donors (Hacc&Hdor): Hydrogen bonds are a type of weak intermolecular interaction. Among molecules with more than 5 hydrogen bond donors and / or more than 10 hydrogen bond acceptors, most are glycosylated molecules or amino acids related to natural products. The values ​​of hydrogen bond acceptors and hydrogen bond donors can affect the lattice energy and melting point, and thus affect the solubility. Some intramolecular hydrogen bonds are also related to the permeability of substances. Lipid-water partition coefficient (XlogP): The lipid-water partition coefficient is the concentration ratio of a compound when it reaches thermodynamic equilibrium between the lipid phase and the water phase. Generally, the larger the lipid-water partition coefficient of a compound, the more soluble it is in fat, and vice versa. Substances that are easily soluble in fat are lipophilic or hydrophobic in the body, and the phenomenon of being easily soluble in water is called hydrophilicity. Lipophilicity / hydrophilicity directly affects the permeability / solubility of raw materials, and as the basic properties of cosmetic raw materials, it affects their efficacy to a certain extent. Topological polar surface area (TPSA): The topological polar surface area of ​​a molecule refers to the surface area of ​​the non-atomic molecular side blocks around the polar atoms (such as oxygen, nitrogen, etc.) in the molecule. TPSA can be used to evaluate the solubility, membrane permeability, and protein binding properties of the compound. When the TPSA is small, it usually means lower compound solubility and higher membrane permeability. As the molecular polar surface area increases, the possibility of binding to proteins also increases. Skin permeability (log K p ): Generally speaking, the larger the log Kp value, the stronger the drug's permeability on the skin. According to the special use of cosmetics, skin permeability is an important reference for measuring compounds as cosmetic raw materials. Although we pursue good absorption effects in cosmetics, we also need to pay attention that the larger the value, the better. Because too high skin permeability may cause the drug to be absorbed too quickly on the skin and quickly enter the blood circulation system, thereby increasing the risk of adverse reactions; The above indicators are common and clear physical and chemical properties of compounds, and skin permeability data can be retrieved through publicly available online databases.

6. The method for screening potential cosmetic ingredients based on molecular docking and physical and chemical properties of substances by establishing a mathematical model according to claim 1, characterized in that: In step S3, data preprocessing means that before actually selecting indicators for comprehensive evaluation, the evaluation standards of various indicators are often different. In real life, there are extremely large indicators such as performance and growth rate; extremely small indicators such as cost and defect rate; intermediate indicators such as pH value and interval indicators with reasonable numerical ranges of body temperature. In general, data of the same type are convenient for subsequent calculations, so each indicator is positively processed before comprehensive evaluation to obtain a data matrix containing all indicators of the compounds to be screened; At the same time, in order to eliminate the influence of different dimensions on the comprehensive evaluation, each group of data matrices after forward transformation is uniformly standardized to construct a standardized data matrix; The data contained in the matrix Z are extremely large data after positive and standardized processing, n is the number of evaluation indicators introduced, and m is the number of evaluation candidate raw materials.

7. The method for screening potential cosmetic ingredients based on molecular docking and physical and chemical properties of substances by establishing a mathematical model according to claim 1, characterized in that: In step S3, the preliminary screening by principal component analysis refers to: importing the pre-processed data matrix containing the best / worst / median solution into SIMCA 14.1 data processing software, taking the compound to be screened as the Primary ID, and using unsupervised PCA principal component analysis to reduce the dimensionality of the multidimensional data; Select alternative raw materials between the median value and the optimal solution for further screening.

8. The method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances according to claim 1, characterized in that: In step S3, the steps of determining the weight coefficient using the entropy weight method are as follows: First, calculate the proportion of each indicator P: Then, in order to quantify the uncertainty of a random variable, the information entropy E of each indicator is calculated: Correspondingly, the greater the information entropy of an indicator in this screening process, the greater its weight: After obtaining the weight coefficient, construct the weighted data matrix 9. The method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances according to claim 1, characterized in that: In step S3, calculating the optimal / worst solution distance means: Obviously, there is an ideal optimal solution Z in the data matrix + and the worst solution Z - (Compound); Use the largest number of each evaluation index in the index after data preprocessing to form the ideal optimal solution matrix, and vice versa: WITH + =[Z1 + ,Z2 + ,...,WITH m + ]=[max{Z 11 ,WITH 21 ,...WITH n1 },max{Z 12 ,WITH 22 ,...,WITH n2 },...,max{Z 1m ,WITH 2m ,...WITH nm }]; WITH - =[Z1 - ,Z2 - ,...,WITH m - ]=[min{Z 11 ,WITH 21 ,…WITH n1 },min{Z 12 ,WITH 22 ,…WITH n2 },…,min{Z 1m ,WITH 2m ,…WITH nm }]; After obtaining the optimal and worst solutions, calculate the distance D from the optimal solution + and the worst solution distance D - :

10. The method for screening potential cosmetic ingredients based on molecular docking and establishing a mathematical model based on the physicochemical properties of substances according to claim 1, characterized in that: In step S3, Topsis comprehensive evaluation refers to: The application potential score S of each compound is calculated according to the Topsis comprehensive evaluation method: The formula is given by Deformation obtained.

Images