Polyamino acid high polymer material, preparation method, application, performance evaluation system and method

Abstract

The invention provides a polyamino acid high polymer material, the chemical formula of which is as shown in the following formula: # imgabs0 #, R1 is a hydrophilic amino acid repeated fragment or a combination of hydrophilic and hydrophobic amino acid repeated fragments; r2 is a linker and is used for connecting a hydrophilic group and a hydrophobic group; r3 is a transition element and is used for reinforcing connection of hydrophobic groups and preventing substitution reaction; the six-membered ring in the formula is selected from at least one of benzene ring, cyclohexane, heterocyclohexane, cyclohexene, cyclohexene, pyridine or pyran; n and m are natural numbers. A polyamino acid high polymer material is used for replacing PEG lipid, the influence of a pre-stored anti-PEG antibody can be effectively avoided, good biodegradability is achieved, the physical dispersion stability of the nanoparticles is effectively improved, the long circulation of the nanoparticles in the body is increased, four components of traditional lipid nanoparticles can be reduced into three components, and the preparation process is simplified. The prepared lipidosome and nanocrystal preparation have excellent stability within a certain time, and the drug loading advantage of the lipidosome is reserved.

Description

Technical Field

[0001] The present invention relates to a technology for encapsulating and solubilizing drugs using liposomes as raw materials, and particularly relates to polymer materials for drug loading and solubilization. Background Art

[0002] A liposome is a bilayer lipid vesicle that encapsulates drugs with lipid materials. It is mainly used in drug delivery systems, which can enhance the efficacy of drugs and reduce toxic and side effects. Its characteristics include targeting and lymphatic targeting, prolonging the blood circulation time of drugs, and improving stability. Among them, cationic liposomes are tiny vesicles composed of cationic lipids and are widely used especially in gene therapy and mRNA therapy. Under acidic conditions, ionizable lipids such as ALC-0315, MC3, and SM-102 are protonated to form cationic lipids, which bind to negatively charged nucleic acids through electrostatic interactions. After the lipids are mixed with the aqueous solution containing nucleic acids and precipitate out, they self-assemble to form nanoparticles loaded with nucleic acids. In addition, cationic liposomes can also be used as efficient delivery carriers for antigens, polypeptides, or small molecule drugs, but cationic liposomes have certain toxic and side effects.

[0003] In addition to liposomes, in recent years, with the development of new drug formulation technologies, formulation technologies such as drug-loaded nanoparticles and drug nanocrystals have been adopted. This type of formulation mainly solves problems such as low bioavailability of poorly soluble drugs and poor drug absorption. However, due to the differences in the physicochemical properties of poorly soluble drugs, there is a lack of high-performance pharmaceutical polymer excipients for the solubility and bioavailability of poorly soluble drugs at present. In terms of drug nanocrystals, the commonly used pharmaceutical excipient polyethylene glycol vitamin E succinate (D-α-Tocopherol polyethylene glycol 1000 succinate, TPGS) has an amphiphilic structure and has certain emulsifying and solubilizing effects, and has good stabilizing and dispersing effects on drug-loaded nanoparticles or drug nanocrystals. However, it has the following disadvantages: 1) TPGS has physiological activities such as P-glycoprotein inhibition, which may change the behavior of some drugs in the body and cause side effects; 2) PEG materials have inherent immunogenicity; 3) Due to the fixed length of the PEG block, its effects in solubilization, nano-dispersion, etc. are limited. The physicochemical properties such as the surface charge and particle size of liposomes affect the formation of their protein corona. For example, cationic liposomes tend to adsorb plasma proteins with an isoelectric point (PI) less than 5.5, while anionic liposomes are more likely to interact with plasma proteins with a PI greater than 5.5. Liposomes with a large particle size have a larger surface curvature and surface area, and can achieve a greater degree of plasma protein adsorption. The particle size of liposomes affects their blood circulation time and tissue distribution, etc. When the particle size of liposomes exceeds the endothelial gap, their tissue penetration ability will decrease.

[0004] Therefore, it is of great practical significance to prepare a new type of liposome to solve the problems of solubilization, emulsification, dispersion, and stability of poorly soluble drugs. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a polyamino acid polymer material, a preparation method, and an application, aiming to solve the problems existing in the background technology.

[0006] The technical solution adopted by the present invention to solve the above technical problems is:

[0007] In the first aspect of the present invention, the present invention provides a polyamino acid polymer material, and its chemical formula is shown as the following formula:

[0008] (Formula I)

[0009] Among them, R1 is a hydrophilic amino acid repeat fragment or a combination of hydrophilic and hydrophobic amino acid repeat fragments;

[0010] R2 is a linker for connecting a hydrophilic group and a hydrophobic group;

[0011] R3 is a transition group for strengthening the connection of the hydrophobic group and preventing the occurrence of substitution reactions;

[0012] The six-membered ring in the formula is selected from at least one of benzene ring, cyclohexane, heterocyclohexane, cyclohexene, heterocyclohexene, pyridine, or pyran;

[0013] n and m are natural numbers.

[0014] Furthermore, R1 is selected from at least one of Structural Formula II, Formula III, or Formula IV:

[0015] (Formula II); (Formula III); (Formula IV);

[0016] Among them, R 21 and R 22 are independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen, or silyl.

[0017] Furthermore, R2 is selected from at least one of Structural Formulas V to XI:

[0018] (Formula V); (Formula VI);

[0019] (Formula VII); (Formula VIII);

[0020] (Formula VIIII); (Formula X);

[0021] (Formula XI);

[0022] wherein, R 31 and R 32 are independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

[0023] Furthermore, R4, R5, R7, R8, R9, R 10 , R 11 are each independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl. R3 is selected from at least one of oxygen, sulfur, selenium, alkyl, heteroalkyl or silyl.

[0024] Furthermore, R6 is selected from at least one of the following structural formulas;

[0025] (Formula XII);

[0026] (Formula XIII);

[0027] (Formula XIIII);

[0028] (Formula XV).

[0029] Furthermore, the excipients of the polyamino acid polymer material include at least one selected from organic solvents DMSO, DMF, DMAc, DCM, ether, methyl tert-butyl ether, isopropyl ether, n-heptane, n-hexane and petroleum ether. The excipients referred to in the present invention are substances that provide reaction conditions such as the possibility and convenience for chemical reactions, such as solvents, catalysts, etc.

[0030] The synthesis route of the polyamino acid polymer material is as follows:

[0031] .

[0032] On this basis, the present invention provides the application of the polyamino acid polymer material prepared according to the above method in the preparation of drugs.

[0033] On this basis, the present invention also provides the application of the polyamino acid polymer material prepared according to the above method in dissolution or emulsification.

[0034] On this basis, the present invention also provides the application of the polyamino acid polymer material prepared according to the above method as a drug delivery carrier.

[0035] On this basis, the present invention also provides the application of the polyamino acid polymer material prepared according to the above method in terms of heat resistance.

[0036] On this basis, the present invention also provides the application of the polyamino acid polymer material prepared according to the above method in terms of bactericidal property.

[0037] In the first aspect of the present invention, the present invention provides a system and method for evaluating the performance of the above polyamino acid polymer material.

[0038] The system specifically includes:

[0039] The SMILES processing module is configured to generate an initial SMILES representation through a molecular structure conversion method;

[0040] The ChemBERTa encoding module is configured to perform ChemBERTa encoding on the processed SMILES representation to generate a molecular embedding vector suitable for multi-objective regression prediction;

[0041] The model construction and training module is configured to construct a Transformer encoder architecture and pre-train it with a chemical dataset to optimize the semantic representation ability.

[0042] The evaluation module: is configured to perform regression prediction through a multi-objective regression model and use the weighted mean square error loss to supervise the regression prediction results.

[0043] The specific evaluation method of this system is as follows:

[0044] (1) SMILES processing: Generate an initial SMILES representation through a molecular structure conversion method;

[0045] (2) ChemBERTa encoding: Perform ChemBERTa encoding on the processed SMILES representation to generate a molecular embedding vector suitable for multi-objective regression prediction (such as CMC, Z-average, PDI);

[0046] (3) Model construction and training: The backbone network ChemBERTa model structure adopts a Transformer encoder architecture inherited from BERT and is pre-trained with a chemical dataset (such as PubChem SMILES) to optimize the representation ability of chemical semantics;

[0047] (4) Perform regression prediction through a multi-objective regression model and use the weighted mean square error (Weighted MSE) loss to supervise the regression prediction results.

[0048] The beneficial effects of the present invention are:

[0049] (1) The present invention uses a polyamino acid polymer material to replace the PEG lipid, which can effectively avoid the influence of pre-existing anti-PEG antibodies and fundamentally solve the problem. This polyamino acid polymer material contains both poly(sarcosine) and liposomes, has good biodegradability, effectively improves the physical dispersion stability of nanoparticles, increases the long circulation of nanoparticles in vivo, and at the same time can reduce the four-component of traditional lipid nanoparticles to a three-component formulation. The prepared liposomes and nanocrystal preparations have excellent stability within a certain period of time, and greatly retain the drug-loading advantages of liposomes themselves.

[0050] (2) The product synthesized in the present invention is an amphiphilic block copolymer, that is, the lipophilic segment existing in the liposome and the hydrophilic segment of poly(sarcosine) with different degrees of polymerization. This amphiphilic structure endows it with common properties of common amphiphilic materials such as solubilization, emulsification, stabilizer for drug-loaded nanoparticles, stabilizer for drug nanocrystal dispersion, detergency, and wetting.

[0051] (3) Experimental results demonstrate that the product of the present invention shows relatively superior effects in aspects such as emulsifying performance, solubilization, dispersion and stability in drug nanocrystals and nucleic acid lipid nanoparticles.

[0052] (4) The present invention also provides a prediction system and method, which make up for the limitations of experiments, can better infer the functions and effects of compounds, provide valuable references and directions for subsequent research, and can greatly reduce the waste of time and resources in the R & D process. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 are physical pictures of nano-suspensions (TPGS-PTX, CPSS1-PTX, CPSS2-PTX) of the poorly soluble drug paclitaxel (Paclitaxel, PTX) prepared from TPGS in the prior art and the products CPSS1 and CPSS2 of the present invention in Example 3 of the present invention.

[0054] Figure 2 are physical pictures of nano-suspensions (TPGS-SIRO, CPSS1-SIRO, CPSS2-SIRO) of the poorly soluble drug sirolimus (Sirolimus, SIRO) prepared from TPGS in the prior art and the products CPSS1 and CPSS2 of the present invention in Example 3 of the present invention.

[0055] Figure 3 is a physical picture of the bottom precipitate of paclitaxel nanocrystals (CPSS1-PTX) prepared with the product CPSS1 in Example 3 of the present invention.

[0056] Figure 4It is a physical picture of the bottom precipitate of preparing paclitaxel nanocrystals (TPGS-PTX) from TPGS in the prior art in Example 3 of the present invention.

[0057] Figure 5 It is a graph of the particle size and distribution results of the TPGS-PTX nanocrystal suspension in Example 3 of the present invention.

[0058] Figure 6 It is a graph of the particle size and distribution results of the CPSS1-PTX nanocrystal suspension in Example 3 of the present invention.

[0059] Figure 7 It is a graph of the particle size and distribution results of the CPSS2-PTX nanocrystal suspension in Example 3 of the present invention.

[0060] Figure 8 It is a graph of the particle size and distribution results of the TPGS-SIRO nanocrystal suspension in Example 3 of the present invention.

[0061] Figure 9 It is a graph of the particle size and distribution results of the CPSS1-SIRO nanocrystal suspension in Example 3 of the present invention.

[0062] Figure 10 It is a graph of the particle size and distribution results of the CPSS2-SIRO nanocrystal suspension in Example 3 of the present invention.

[0063] Figure 11 It is a fluorescence emission spectrogram for measuring the critical micelle concentration of the product CPSS3 in Example 4 of the present invention.

[0064] Figure 12 It is a graph of the changes in particle size and PDI of the TPGS-PTX, CPSS1-PTX, and CPSS2-PTX nanocrystal suspensions stored at 4°C for 50 days in Example 3 of the present invention.

[0065] Figure 13 It is a graph of the changes in particle size and PDI of the TPGS-SIRO, CPSS1-SIRO, and CPSS2-SIRO nanocrystal suspensions stored at 4°C for 50 days in Example 3 of the present invention.

[0066] Figure 14 It is a comparative graph of the solubility experiment results of PTX and SIRO in water, TPGS, CPSS1, and CPSS2 respectively in Example 6 of the present invention. Detailed implementation mode

[0067] To enable those skilled in the art to better understand and implement the technical solution of the present invention, the present invention will be specifically, clearly, and completely described below in the form of embodiments in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. And, based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts, any modifications, equivalent replacements, improvements, etc., shall be included within the protection scope of the present invention.

[0068] To enable those skilled in the art to better understand the technical solution of the present invention, some technical terms of the present invention will be explained first. In the text of the present invention, these explanations are mainly to help those skilled in the art understand and implement the technical solution of the present invention, rather than a limitation on the present invention. And, there may be other technical understandings of these technical terms in other disclosures outside the present invention. To avoid technical ambiguities, when understanding the present invention, these explanations of the present invention should be based on.

[0069] Pharmaceutical excipients

[0070] In the present invention, the so-called pharmaceutical excipients refer to excipients and additives used in the production of drugs or the preparation of prescriptions. They are substances that have been reasonably evaluated in terms of safety and are included in drug preparations except for the active ingredients.

[0071] Solubilization and solubilizers

[0072] During the research and development of drug preparations, the solubility of many poorly soluble drugs such as volatile oils, fat-soluble vitamins, and steroid hormones in water is very small and cannot reach the concentration required for treatment. At this time, the method of adding surfactants is often used to increase the solubility of drugs in water. The surfactant that plays the role of solubilization is called a solubilizer.

[0073] Liposomes

[0074] When amphiphilic molecules such as phospholipids are dispersed in an aqueous phase, the hydrophobic tails of the molecules tend to aggregate together, avoiding the aqueous phase, while the hydrophilic heads are exposed to the aqueous phase, forming a closed vesicle with a bilayer structure. Various drugs can be encapsulated in the aqueous phase inside the vesicle and within the bilayer membrane, similar to the structure of an ultra-microcapsule. The ultramicroscopic spherical drug-loaded preparation made by encapsulating drugs in the middle of a lipid bilayer film is called a liposome.

[0075] Cholesterol

[0076] Cholesterol is a neutral lipid and also belongs to amphiphilic molecules, but its lipophilicity is greater than its hydrophilicity. Since the energy of self-association of cholesterol itself is relatively large, it is difficult to bind to proteins and mainly binds to phospholipids to prevent phospholipids from aggregating into crystal structures. In the present invention, cholesterol is used as an embodiment of liposomes.

[0077] Nanocrystal

[0078] Nanocrystals, also known as nanosuspensions, are pure drug submicron colloidal dispersion systems with a particle size of less than 1 μm formed by dispersing drug particles in a liquid medium using surfactants or polymers as stabilizers.

[0079] Surfactant

[0080] Surfactants refer to substances that can significantly reduce the surface tension of a liquid, and are called surfactants of that liquid. Their structures contain polar hydrophilic groups and non-polar hydrophobic groups.

[0081] Emulsions

[0082] Emulsions refer to a heterogeneous liquid dispersion system formed by mixing two immiscible liquids, where one liquid is dispersed in the form of droplets in the other liquid.

[0083] Critical micelle concentration (CMC)

[0084] The critical micelle concentration (CMC) is a characteristic property of amphiphilic polymer materials. That is, when the amphiphilic polymer material reaches a certain concentration in water, further increasing the material concentration can form micelles. The critical micelle concentration (CMC) is the lowest concentration at which surfactant molecules associate to form micelles in a solution. After surfactants reach a certain concentration in an aqueous solution, the solubility of some water-insoluble or slightly soluble drugs in the micellar solution formed by them can increase significantly and form a transparent micellar solution. At this time, the concentration of the surfactant is called the critical micelle concentration. Its formation mechanism is that surfactant molecules usually have an amphiphilic structure, that is, one end is hydrophilic and the other end is hydrophobic. At low concentrations, surfactant molecules exist as individual molecules in water. As the concentration increases, the molecules will accumulate on the water surface to form a monolayer. When the surface adsorption reaches saturation, the excess molecules will aggregate to form micelles. The hydrophobic groups inside the micelles aggregate with each other, and the hydrophilic groups face outwards to contact water, thereby reducing the surface tension of water. In the present invention, the critical micelle concentration can be used to verify the dissolution properties of polymer materials.

[0085] Accelerated blood clearing (ABC)

[0086] Accelerated blood clearing refers to the complement activation caused by the binding of anti-PEG antibodies to the surface of PEGylated nanodrugs, which greatly accelerates the clearance rate of drugs in the body.

[0087] Example 1, a polyamino acid polymer material.

[0088] This example provides a polyamino acid polymer material, the chemical structure of which is shown in Formula I.

[0089] (Formula I)

[0090] Wherein,[[]]

[0091] R1 is a hydrophilic amino acid repeat fragment or a combination of hydrophilic and hydrophobic amino acid repeat fragments; R1 is selected from at least one of the structural formulas II, III or IV:[[]]

[0092] (Formula II);[[]] (Formula III);[[]] (Formula IV);[[]]

[0093] R 21 and R 22 are independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

[0094] R2 is a linker for connecting a hydrophilic group and a hydrophobic group;[[]]

[0095] R2 is selected from at least one of the structural formulas V to XI:[[]]

[0096] (Formula V);[[]] (Formula VI);[[]]

[0097] (Formula VII);[[]] (Formula VIII);[[]]

[0098] (Formula VIIII);[[]] (Formula X);[[]]

[0099] (Formula XI);[[]]

[0100] Wherein, R 31 and R 32 are independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

[0101] R3 is a transitional group for strengthening the connection of hydrophobic groups and preventing the occurrence of substitution reactions; R3 is selected from at least one of oxygen, sulfur, selenium, alkyl, heteroalkyl or silyl.

[0102] R4, R5, R7, R8, R8, R 10 , R 11 are each independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

[0103] R6 is selected from at least one of the following structural formulas;

[0104] (Formula XII);

[0105] (Formula XIII);

[0106] (Formula XIIII);

[0107] (Formula XV).

[0108] The six-membered ring in the formula is selected from at least one of benzene ring, cyclohexane, heterocyclohexane, cyclohexene, heterocyclohexene, pyridine or pyran;

[0109] n and m are natural numbers.

[0110] The excipients of the polyamino acid polymer material include at least one selected from organic solvents DMSO, DMF, DMAc, DCM, ether, methyl tert-butyl ether, isopropyl ether, n-heptane, n-hexane and petroleum ether. The excipients referred to in the present invention are substances that provide reaction conditions such as possibilities and conveniences for chemical reactions, such as solvents, catalysts, etc.

[0111] The synthesis route of the polyamino acid polymer material is as follows:

[0112] .

[0113] Example 2, a preparation method of a polyamino acid polymer material.

[0114] This example provides a preparation method of a polyamino acid polymer material. According to the preparation method of this example, the polyamino acid polymer material described in Example 1, that is, CPSS, can be prepared.

[0115]

[0116] The preparation process is as follows:

[0117] I. Preparation of CPSS1 (i.e., n = 14):

[0118] (1)Synthesis intermediate pSar14S:

[0119] (1.1)Dissolution: Add 5.83 g (28 mmol, 14 eq.) of N-phenoxycarbonylsarcosine (SarNPC) to a reaction flask. Weigh 1.29 g (10 mmol, 5 eq.) of N,N-diisopropylethylamine (DIPEA) and add it to the reaction flask. Measure 30 mL of dimethylacetamide (DMAc) and add it to the reaction flask, and stir until dissolved and clear.

[0120] (1.2)Polymerization: Weigh 174 mg (2 mmol, 1 eq.) of neopentylamine and add it to the above reaction flask, and carry out a polymerization reaction at 60 °C for 24 h.

[0121] (1.3)Amidation: Weigh 0.30 g (3 mmol, 1.5 eq.) of succinic anhydride and add it to the above reaction system, and react at room temperature for 24 h.

[0122] (1.4)Purification: Add an appropriate amount of ether to the reaction solution for centrifugal sedimentation, repeat 3 times, and obtain 2.32 g of white solid pSar14S after vacuum drying, with a yield of 98%.

[0123] (2)Synthesis of product CPSS1:

[0124] (2.1)Esterification: Weigh 0.59 g (0.5 mmol, 1 eq.) of pSar14S, 0.39 g (1 mmol, 2 eq.) of cholesterol, 124 mg (0.6 mmol, 1.2 eq.) of dicyclohexylcarbodiimide (DCC), and 15.3 mg (0.125 mmol, 0.25 eq.) of 4-dimethylaminopyridine (DMAP), add them all to a reaction flask at once, then add 6 mL of dichloromethane (DCM), stir at room temperature for 24 h, and then filter by suction to obtain a suction filtrate.

[0125] (2.2)Purification: Add an appropriate amount of ether to the above suction filtrate for centrifugal sedimentation, repeat 3 times, and obtain 705.7 mg of white solid CPSS1 after vacuum drying, with a yield of 91%.

[0126] II. Preparation of CPSS2 (i.e., n = 28):

[0127] (1)Synthesis intermediate pSar28S:

[0128] (1.1)Dissolution: Add 5.83 g (28 mmol, 28 eq.) of N-phenoxycarbonylsarcosine (SarNPC) into a reaction flask, weigh 0.65 g (5 mmol, 5 eq.) of N,N-diisopropylethylamine (DIPEA) and add it into the reaction flask, measure 30 mL of dimethylacetamide (DMAc) and add it into the reaction flask, and stir until it is dissolved and clear.

[0129] (1.2)Polymerization: Weigh 87.2 mg (1 mmol, 1 eq.) of neopentylamine and add it into the above reaction flask, and carry out a polymerization reaction at 60 °C for 24 h.

[0130] (1.3)Amidation: Weigh 0.50 g (5 mmol, 5 eq.) of succinic anhydride and add it into the above reaction system, and react at room temperature for 24 h.

[0131] (1.4)Purification: Add an appropriate amount of ether into the reaction solution for centrifugal sedimentation, repeat 3 times, and obtain 2.13 g of white solid pSar28S after vacuum drying, with a yield of 98%.

[0132] (2)Synthesis of product CPSS2:

[0133] (2.1)Esterification: Weigh 1.09 g (0.5 mmol, 1 eq.) of pSar28S, 0.29 g (0.75 mmol, 1.5 eq.) of cholesterol, 124 mg (0.6 mmol, 1.2 eq.) of dicyclohexylcarbodiimide (DCC), and 30.5 mg (0.25 mmol, 0.5 eq.) of 4-dimethylaminopyridine (DMAP), add them all into a reaction flask at one time, then add 10 mL of dichloromethane (DCM), stir at room temperature for 24 h, and then carry out suction filtration to obtain a suction filtrate.

[0134] (2.2)Purification: Add an appropriate amount of ether into the above suction filtrate for centrifugal sedimentation, repeat 3 times, and obtain 1.21 g of white solid CPSS2 after vacuum drying, with a yield of 95%.

[0135] III. Preparation of CPSS3 (i.e., n = 25):

[0136] (1)Synthesis of intermediate pSar25S:

[0137] (1.1)Dissolution: Add 5.23 g (25 mmol, 25 eq.) of N-phenoxycarbonylsarcosine (SarNPC) into a reaction flask, weigh 0.65 g (5 mmol, 5 eq.) of N,N-diisopropylethylamine (DIPEA) and add it into the reaction flask, measure 30 mL of dimethylacetamide (DMAc) and add it into the reaction flask, and stir until it is dissolved and clear.

[0138]

[0138] (1.2) Polymerization: Weigh 87.2 mg (1 mmol, 1 eq.) of neopentylamine and add it to the above reaction flask, then carry out the polymerization reaction at 60 °C for 24 h.

[0139]

[0139] (1.3) Amidation: Weigh 1 g (10 mmol, 10 eq.) of succinic anhydride and add it to the above reaction system, then react at room temperature for 24 h.

[0140]

[0140] (1.4) Purification: Add an appropriate amount of ether to the reaction solution for centrifugal sedimentation, repeat 3 times, and obtain 1.94 g of white solid pSar25S after vacuum drying, with a yield of 99%.

[0141]

[0141] (2) Synthesis of product CPSS3:

[0142]

[0142] (2.1) Esterification: Weigh 0.98 g (0.5 mmol, 1 eq.) of pSar25S, 232 mg (0.6 mmol, 1.2 eq.) of cholesterol, 124 mg (0.6 mmol, 1.2 eq.) of dicyclohexylcarbodiimide (DCC), and 30.5 mg (0.25 mmol, 0.5 eq.) of 4-dimethylaminopyridine (DMAP), add them all into the reaction flask at one time, then add 10 mL of dichloromethane (DCM), stir at room temperature for 24 h, and then carry out suction filtration to obtain the suction filtrate.

[0143]

[0143] (2.2) Purification: Add an appropriate amount of ether to the above suction filtrate for centrifugal sedimentation, repeat 3 times, and obtain 1.08 g of white solid CPSS3 after vacuum drying, with a yield of 93%.

[0144]

[0144] Example 3, performance test on the dispersion and stability of product CPSS as drug nanocrystals.

[0145]

[0145] In this example, the application potential of CPSS1 with a sarcosine polymerization degree of 14 (i.e., n = 14) and CPSS2 with a polymerization degree of 28 (i.e., n = 28) in the preparation of drug nanocrystals was tested.

[0146]

[0146] In this example, two representative poorly soluble drugs, namely sirolimus (SIRO) and paclitaxel (PTX), were selected, and their nanocrystals were prepared. The particle size, PDI, and stability of the nanocrystals were used as evaluation indicators to investigate the advantages of using them as stabilizers for drug nanocrystals. First, the nanocrystals of paclitaxel were prepared, and the dispersion and stability of TPGS, CPSS1, and CPSS2 in the nanocrystals were compared.

[0147] The preparation method is as follows: Dissolve paclitaxel in DMSO (40 mg / mL), prepare an aqueous solution of TPGS / CPSS1 / CPSS2 with a certain concentration, and slowly inject the DMSO solution of paclitaxel into the aqueous solution of TPGS / CPSS1 / CPSS2 using a syringe pump while maintaining high-speed magnetic stirring (1200 rpm). As the DMSO solution of paclitaxel is injected, uniform nanoparticles can be seen forming in the aqueous solution. After preparation, sonicate for 15 min using a sonicator (20%, sonicate for 2 s, pause for 2 s). The sonicated nanocrystal suspension is ultracentrifuged at 16,000×g for 30 min. After centrifugation, discard the supernatant, resuspend the nanocrystal precipitate with a small amount of water, and then continue to sonicate for 5 min using a sonicator (20%, sonicate for 2 s, pause for 2 s). Dilute the prepared paclitaxel nanosuspension ( Figure 1 ), and measure the particle size of the nanocrystals using a particle size and zeta potential analyzer after dilution with 10 mM sodium chloride: The particle size of the nanocrystals prepared with TPGS is 465.54 ± 8.13 nm, and the PDI is 0.121 ± 0.040. The particle size of the nanocrystals prepared with CPSS1 is 146.32 ± 2.17 nm, and the PDI is 0.054 ± 0.039. The particle size of the nanocrystals prepared with CPSS2 is 168.74 ± 2.55 nm, and the PDI is 0.044 ± 0.009. It can be seen from the results that the paclitaxel nanocrystals prepared with CPSS1 / CPSS2 have a smaller particle size and more uniform nanocrystal particles. The nanocrystals prepared with TPGS show precipitation at the bottom ( Figure 4 ), and there are spikes at the end of the particle size graph ( Figure 5 、 Figure 6 、 Figure 7 ), indicating that CPSS is more effective than TPGS in dispersing and stabilizing nanocrystals.

[0148] Further investigate the change in the stability of the prepared paclitaxel nanocrystals over time. The paclitaxel nanocrystals prepared with TPGS show unstable PDI indicators in a short time, while the paclitaxel nanocrystals prepared with CPSS1 / CPSS2 can remain stable after being stored at 4°C for 50 days, and there are no obvious changes in the particle size and PDI ( Figure 12 ). From the above results, it can be seen that when different blocks of CPSS are used to prepare paclitaxel nanocrystals, they not only show a better effect of dispersing nanocrystals than TPGS, but also can effectively stabilize the drug nanocrystals and improve the physical stability of their suspensions.

[0149] Investigate the sirolimus nanocrystals prepared with TPGS / CPSS1 / CPSS2 using the same method as above. Disperse the nanocrystal dispersion prepared according to the above method ( Figure 2) After dilution with 10 mM sodium chloride to a certain concentration, the particle size and distribution of each nanocrystal suspension were measured using a particle size potentiometer, and the results are as Figure 4 shown. Further, the stability of the prepared sirolimus nanocrystals over time was investigated. The sirolimus nanocrystals prepared with TPGS / CPSS1 / CPSS2 remained stable after storage at 4°C for 50 days, and there were no significant changes in the particle size and PDI ( Figure 13 ). From the results, it can be seen that the particle size of the nanocrystal suspension prepared with CPSS1 / CPSS2 is smaller than that of the nanocrystal suspension prepared with TPGS.

[0150] Example 4: Determination of the critical micelle concentration value (CMC) of the product CPSS.

[0151] In this example, the pyrene fluorescence method was used to determine the critical micelle concentration (CMC) of CPSS. Pyrene is the most commonly used fluorescent probe molecule, which has very low solubility in water. The core of the polymer micelles formed by CPSS is a hydrophobic region, so pyrene can be solubilized into the CPSS polymer micelles through hydrophobic interaction. When the concentration of CPSS is less than its critical micelle concentration, CPSS molecules exist on the surface of the solution, that is, at the gas-liquid interface. At this time, pyrene is not solubilized in the micelles. When the concentration of CPSS in the solution reaches its CMC, the gas-liquid interface can no longer accommodate CPSS molecules, and the excess CPSS molecules enter the interior of the solution to form micelles, that is, the cholesterol lipophilic end of CPSS is inside the micelles, and the hydrophilic chain of poly(sarcosine) is on the surface of the micelles. When the concentration of CPSS in the solution further increases, that is, > CMC, more and more CPSS micelles will be formed, which will solubilize the poorly soluble pyrene into the interior of the micelles. After excitation at 335 nm, five electronic vibration peaks appear in the fluorescence emission spectrum of pyrene in the solution. The intensity ratio I1 / I3 of the first and the third vibration peaks strongly depends on the polarity of the environment where pyrene is located. Therefore, the polarity of the two environments of pyrene in the solution and in the CPSS micelles is very different, so the value of I1 / I3 will change greatly. A curve can be made with the I1 / I3 ratio and the concentration, and the CMC value of CPSS can be obtained by the tangent method. According to the calculation at the inflection point tangency, the critical micelle concentration of CPSS 3(pSar: n = 25) is 54.65 ± 1.81 μg / mL.

[0152] The experimental operations are as follows: 50 μL (10 μg / mL) of an acetone solution of pyrene was dropped into a 5.0 mL brown vial and evaporated to dryness in a fume hood. A series of CPSS (pSar: n = 25) solutions with concentration gradients (0.001, 0.01, 0.1, 1, 5, 10, 50, 100, 250, 500, 1000 μg / mL) were prepared and added to the above-mentioned brown vial containing pyrene, and then heated in a 45°C water bath for 1.5 h. After cooling to room temperature, the fluorescence emission spectra of the solution containing pyrene and the surfactant were scanned using a fluorescence spectrometer, with the surfactant solution as the blank control. The temperature was set at 25°C, the excitation wavelength was 335 nm, the excitation slit width was 5 nm, the emission slit width was 2.5 nm, and the emission spectrum scanning range was 365 - 450 nm. A curve was plotted with the logarithm of the concentration of CPSS3 (pSar: n = 25) as the abscissa and the ratio of the intensities of the first and third fluorescence characteristic peaks of pyrene as the ordinate for fitting. The abscissa value at the intersection of the two fitted straight lines is the CMC. As Figure 11 shown, in this example, the average value of three times was taken as the result value.

[0153] Through this example, it was verified that the product CPSS of the present invention has the property of solubilizing drugs.

[0154] Example 5: The product CPSS was used as a drug nanocrystal stabilizer to measure the drug loading.

[0155] In this example, the drug loading measurement experiment was carried out on the drug nanocrystals prepared from CPSS1 (pSar: n = 14) / CPSS2 (pSar: n = 28).

[0156] In this example, two poorly soluble drugs, namely sirolimus (Sirolimus, SIRO) and paclitaxel (Paclitaxel, PTX), were selected. The experimental preparation method was the same as that in Example 3. Subsequently, the prepared nanocrystals were freeze-dried for 24 h, and then the drug loading was measured by HPLC. For sirolimus nanocrystals prepared with the same experimental parameters and procedures, the drug loading rate of TPGS-SIRO nanocrystals was 90.06 ± 10.10%; the drug loading rate of CPSS1-SIRO nanocrystals was 102.36 ± 8.18%; the drug loading rate of CPSS2-SIRO nanocrystals was 104.8 ± 4.64%. From the experimental results, it can be seen that the ability of CPSS1 / CPSS2 to load sirolimus can be as high as nearly 100%, which is significantly higher than 90% of TPGS. Using this type of stabilizer for sirolimus can greatly avoid drug loss; for paclitaxel nanocrystals prepared with the same experimental parameters and procedures, the drug loading rate of TPGS-PTX nanocrystals was 65.87 ± 0.99%; the drug loading rate of CPSS1-PTX nanocrystals was 94.34 ± 0.31%; the drug loading rate of CPSS2-PTX nanocrystals was 96.02 ± 2.00%; from the experimental results, it can be seen that the ability of CPSS1 / CPSS2 to load paclitaxel is similar, both around 95%, which is significantly higher than 65% of TPGS.

[0157] Example 6, solubilization experiment of the product CPSS on poorly soluble drugs.

[0158] In this example, two poorly soluble drugs, namely sirolimus (Sirolimus, SIRO) and paclitaxel (Paclitaxel, PTX), were selected.

[0159] Excess SIRO and PTX were placed in vials containing aqueous solutions of CPSS1 (pSar: n = 14), CPSS2 (pSar: n = 28) or TPGS, and then placed in a thermostatic culture oscillator and shaken at 60 rpm for 24 h to fully dissolve the drugs. Then the samples were centrifuged at high speed, and the supernatant was taken to determine the drug content in the supernatant by high performance liquid chromatography (HPLC). The results are shown in Figure 14 . It can be seen from the figure that the solubility of SIRO and paclitaxel in water is extremely low, and the concentration in the supernatant is lower than the lower limit of quantification of the two drugs by HPLC. For SIRO, both TPGS and CPSS can improve the solubility of SIRO in water to a certain extent, and CPSS2 has the best effect. However, for PTX, no obvious solubilization effect was observed for TPGS, while for CPSS, it can be seen that it can significantly solubilize paclitaxel, indicating that it is significantly superior to TPGS in solubilizing certain drugs.

[0160] Example 7: Test the dispersibility and stability of the product CPSS as a component for preparing lipid nanoparticles.

[0161] The polyamino acid polymer material prepared by the method provided in Example 2 can be used to prepare the composite lipid component in nucleic acid lipid nanoparticles. Through the product (CPSS) in Example 2, the four different lipid component compositions (ionizable lipid, phospholipid, cholesterol, and polyethylene glycol lipid (PEG lipid)) commonly used in the prior art can be replaced by three different lipid components (ionizable lipid, phospholipid, CPSS1 (pSar: n = 14) / CPSS2 (pSar: n = 28)) and four different lipid component compositions (ionizable lipid, phospholipid, cholesterol, and CPSS1 (pSar: n = 14) / CPSS2 (pSar: n = 28)).

[0162] In this example, nucleic acid lipid nanoparticles prepared with CPSS1 (pSar: n = 14) / CPSS2 (pSar: n = 28) were experimentally studied for their application potential. Since factors such as the surface charge, particle size, and morphology of lipid nanoparticles are key factors affecting the in vivo process and pharmacokinetic properties of lipid nanoparticle drugs.

[0163] In this example, pCMV-EGFP-3×Linker-MCS-Neo was selected for encapsulation to prepare nucleic acid lipid nanoparticles. The advantages of lipid nanoparticles were mainly investigated by taking the particle size, PDI, and stability of lipid nanoparticles as evaluation indexes.

[0164] The preparation method is as follows:

[0165] 1) An ionizable lipid (SM-102), a phospholipid (DSPC), cholesterol (optional component), and CPSS1 / CPSS2 were formulated into a 10 mg / mL compound lipid with absolute ethanol as the solvent at a certain molar ratio. According to N / P = 8 and a certain value of the flow rate ratio of citrate buffer to compound phospholipid, the plasmid required in the citrate buffer (or blank lipid nanoparticles were prepared) was calculated, and solutions with the composite lipid in absolute ethanol as the organic phase and nucleic acid in citrate buffer as the aqueous phase were obtained respectively.

[0166] 2) By the solvent injection method, the lipid mixture was injected into the buffer at a flow rate of 50 μL / min using a syringe pump, and then vortexed for 15 s and sonicated in an ice bath for 10 min (20%, stop for 2 s, sonicate for 2 s).

[0167] 3) The obtained crude product was dialyzed using an 8 - 14 KDa dialysis bag within 10 min. The dialysis solution was 1×PBS (pH 7.4), and dialysis was carried out at 4°C for 2 h (the dialysis solution was at least 50 times the sample).

[0168] 4) The dialysis buffer was changed to 1×PBS (pH 7.4) containing 10% sucrose and dialyzed overnight at 4°C; the final product of the lipid nanoparticles was stored at 4°C.

[0169] As can be seen from Table 1, in the case of no plasmid encapsulation, when the ratio of the organic phase to the aqueous phase was 1:1.5, the lipid nanoparticles composed of the three components of SM-102 / DSPC / CPSS had significantly better average size and PDI than the lipid nanoparticles composed of the four components of SM-102 / DSPC / cholesterol / CPSS.

[0170] Table 1

[0171]

[0172] As can be seen from Table 2, in the case of plasmid encapsulation, when the ratio of the organic phase to the aqueous phase was 1:1.5, the lipid nanoparticles composed of the three components of SM-102 / DSPC / CPSS had significantly better average size and PDI than the lipid nanoparticles composed of the four components of SM-102 / DSPC / cholesterol / CPSS, and the encapsulation efficiency (EE%) of the three-component lipid nanoparticles was greater than 95%.

[0173] Table 2

[0174]

[0175] As can be seen from Table 3, 1 and 2 are lipid nanoparticles without plasmid, 3 and 4 are lipid nanoparticles with plasmid. When the ratio of the organic phase to the aqueous phase was 1:3, the lipid nanoparticles composed of the four components of SM-102 / DSPC / cholesterol / CPSS also showed good performance, but the dispersibility was not as good as that of the three-component lipid nanoparticles.

[0176] Table 3

[0177]

[0178] As can be seen from Table 4, in the case of with or without plasmid encapsulation, when the ratio of the organic phase to the aqueous phase was 1:1.5, the lipid nanoparticles composed of the three components of SM-102 / DSPC / CPSS had good stability.

[0179] Table 4

[0180]

[0181] Example 8: A system and method for evaluating polyamino acid polymers

[0182] The present embodiment provides a system and method for evaluating the indexes of the prepared amino acid polymer compound. According to the invention of the present invention, a variety of compounds with different structures can be prepared, and these compounds can be used for entrapment, volume expansion, emulsification, etc., but in actual production, it is difficult to experiment on all compounds one by one in a short time to verify their effects. In order to enable those skilled in the art to better understand and implement the technical solution of the present invention, and to better demonstrate the technical effect of the present invention, this embodiment designs an evaluation system, which evaluates the compounds through data learning and artificial intelligence algorithms, thereby verifying the effect of the polymer compound under the system of the present invention.

[0183] The system includes:

[0184] SMILES processing module: Generates initial SMILES representation through molecular structure conversion method.

[0185] The specific method is:

[0186] 1. Convert structure to SMILES: Generate the initial SMILES representation through molecular structure conversion method. Specifically, based on the input molecular structure, identify atoms and chemical bonds, and perform topological analysis. Construct a molecular graph with atoms as nodes and chemical bonds as edges, record the connection relationship, and determine the SMILES representation through graph traversal.

[0187] 2. Standardization: Use the standardized SMILES generation method (unified molecular representation, fixed atomic order, such as converting benzene: C1=CC=CC=C1 to c1ccccc1) to eliminate redundant symbols.

[0188] 3. Outlier screening: Remove outliers (e.g., mass <50 or >1000 Da) by molecular weight calculation (calculate total mass based on molecular formula, e.g., ethanol CCO is about 46 Da). Check target value distribution and use interquartile range method (based on upper and lower quartiles of data, remove extreme values, e.g., CMC>1000 μg / mL).

[0189] ChemBERTa encoding module: encodes the processed SMILES representation ChemBERTa to generate molecular embedding vectors suitable for multi-target regression prediction (CMC, Z-average, PDI).

[0190] The specific method is:

[0191] 1. Tokenization: The SMILES string is segmented into chemically meaningful subunits (tokens) by a dedicated chemical sequence tokenizer. This process splits the string into units such as atoms (e.g., C, O), composite atoms, bond symbols, and ring markers. For example, benzene (c1ccccc1) is tokenized into ['c', '1', 'c', 'c', 'c', 'c', 'c', '1'], where each token represents a fragment of the molecular structure. The tokenizer is based on a pre-trained vocabulary to ensure coverage of common SMILES symbols. The [CLS] token is added at the beginning of the sequence for feature aggregation, and the [SEP] token is added at the end to indicate the end, forming an input such as ['[CLS]', 'c', '1', 'c', 'c', 'c', 'c', 'c', '1', '[SEP]'].

[0192] 2. Input Formatting: The tokenized sequence is converted into a numerical representation. Each token is mapped to a unique integer (Token ID) according to the vocabulary. An attention mask is generated, assigning 1 to valid tokens and 0 to padding tokens, enabling the model to focus on the valid input. To unify the sequence length, short sequences are padded with the [PAD] token to the maximum length (512 tokens). These formatted inputs are output in tensor form, ready to be fed into the ChemBERTa model for inference.

[0193] Model Construction and Training Module: The backbone ChemBERTa model structure inherits the Transformer encoder architecture of BERT and is pre-trained on chemical datasets (such as PubChem SMILES) to optimize the representation ability of chemical semantics.

[0194] Specific methods are as follows:

[0195] 1. The processed molecular embedding vectors are input into ChemBERTa, which includes 12 pre-trained Transformer encoder layers, to generate context-related embeddings.

[0196] 2. The context-related embeddings are input into a multi-objective regression model based on a multi-layer perceptron for regression prediction, and the weighted mean squared error (Weighted MSE) loss is used to supervise the regression prediction results.

[0197] Specifically, the multi-objective regression model based on a multi-layer perceptron includes:

[0198] Hidden Layer 1: For the input context-related embeddings, the input fully connected layer reduces the dimension to 1 * 512, and then the ReLU activation function is used for operation to obtain the first hidden layer feature;

[0199] Hidden layer 2: The input of the first feature of the hidden layer to the fully connected layer reduces the dimension to 1 * 256, and then the activation function ReLU is used for operation to obtain the second feature of the hidden layer;

[0200] Hidden layer 3: The input of the second feature of the hidden layer to the fully connected layer reduces the dimension to 1 * 128, and then the activation function ReLU is used for operation to obtain the third feature of the hidden layer;

[0201] Regression head - CMC: The third feature of the hidden layer is input to the regression head, and the CMC prediction result with a dimension of 1 * 1 is obtained through the fully connected layer;

[0202] Regression head - Z - average: The third feature of the hidden layer is input to the regression head, and the Z - average prediction result with a dimension of 1 * 1 is obtained through the fully connected layer;

[0203] Regression head - PDI: The third feature of the hidden layer is input to the regression head, and the PDI prediction result with a dimension of 1 * 1 is obtained through the fully connected layer.

[0204] Specifically, for the weighted mean squared error (Weighted MSE) loss, in the case of a multi - target regression task, assuming there are K target variables (e.g., K = 3, corresponding to CMC, Z - average, PDI), the calculation formula of the weighted MSE is:

[0205]

[0206] where represents the th target among all targets, is the corresponding weight, and in this embodiment, the weights are the same and taken as 1. is the mean squared error between the target result and the true value, and the calculation formula is:

[0207]

[0208] where is the number of all samples, and represent the predicted value and the true value respectively.

[0209] After training the fixed parameters of the model with additional pre - training data, the processed embedding vectors of unknown molecules can be used to obtain the predicted result values of CMC, Z - average, and PDI according to SMILES processing - ChemBERTa encoding - model prediction.

[0210] Table 5 lists the different indexes and evaluation results of some compounds.

[0211]

[0212] Note: Chemical formula Ⅰ: ;

[0213] Chemical formula II: .

[0214] As can be seen from the above table, through the evaluation system of the present invention, the core indexes such as CMC, Z-average nanoparticle size and PDI value of the polymer compounds obtained by selecting different groups can be evaluated, providing a preliminary theoretical basis for subsequent experiments, reducing the experimental cost, and being able to more accurately find a suitable preparation direction.

[0215] Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A polyamino acid polymer material, characterized in that, Its chemical formula is shown in Formula I: (Formula I) Wherein, R1 is a hydrophilic amino acid repeat fragment or a combination of hydrophilic and hydrophobic amino acid repeat fragments; R2 is a linker for connecting a hydrophilic group and a hydrophobic group; R3 is a transition group for strengthening the connection of the hydrophobic group and preventing the occurrence of substitution reactions; The six-membered ring in the formula is selected from at least one of benzene ring, cyclohexane, heterocyclohexane, cyclohexene, heterocyclohexene, pyridine or pyran; n and m are natural numbers.

2. The polyamino acid polymer material according to claim 1, wherein R1 is selected from at least one of Structural Formula II, Formula III or Formula IV: (Formula II); (Formula III); (Formula IV); Wherein, R 21 and R 22 are independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

3. The polyamino acid polymer material according to claim 2, wherein R2 is selected from at least one of Structural Formulas V to XI: (Formula V); (Formula VI); (Formula VII); (Formula VIII); (Formula VIIII); (Formula X); (Formula XI); wherein, R 31 and R 32 are each independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

4. A polyamino acid polymer material according to claim 1, characterized in that, R3 is selected from at least one of oxygen, sulfur, selenium, alkyl, heteroalkyl or silyl.

5. A polyamino acid polymer material according to claim 1, characterized in that, R4, R5, R7, R8, R9, R 10 , R 11 are each independently selected from at least one of hydrogen, alkyl, heteroalkyl, aryl, alkenyl, alkynyl, halogen or silyl.

6. The polyamino acid polymer material according to claim 1, wherein R6 is selected from at least one of the following structural formulas; (Formula XII); (Formula XIII); (Formula XIIII); (Formula XV).

7. The polyamino acid polymer material according to claim 1, wherein The excipients of the polyamino acid polymer material include at least one selected from organic solvents DMSO, DMF, DMAc, DCM, ether, methyl tert-butyl ether, isopropyl ether, n-heptane, n-hexane and petroleum ether.

8. A polyamino acid polymer material according to claim 1, characterized in that, The synthesis route of the polyamino acid polymer material is as follows: 。 9. Use of the polyamino acid polymer material according to any one of claims 1-8 in the preparation of drugs.

10. Use of the polyamino acid polymer material according to any one of claims 1-8 in terms of solubility.

11. Use of the polyamino acid polymer material according to any one of claims 1-8 as a drug delivery carrier.

12. Use of the polyamino acid polymer material according to any one of claims 1-8 in terms of heat resistance.

13. Use of the polyamino acid polymer material according to any one of claims 1-8 in terms of bactericidal property.

14. Use of the polyamino acid polymer material according to any one of claims 1-8 in terms of emulsification.

15. The performance evaluation system of the polyamino acid polymer material according to any one of claims 1-8, characterized in that, Including: A SMILES processing module configured to generate an initial SMILES representation by a molecular structure conversion method; A ChemBERTa encoding module configured to ChemBERTa encode the processed SMILES representation to generate a molecular embedding vector suitable for multi-objective regression prediction; A model construction and training module configured to construct a Transformer encoder architecture and pre-train it with a chemical dataset to optimize the semantic representation ability; An evaluation module: configured to perform regression prediction through a multi-objective regression model and use weighted mean square error loss to supervise the regression prediction results.

16. The method for evaluating the performance of the polyamino acid polymer material according to any one of claims 1-8, characterized in that, Including the following steps: (1) SMILES processing: Generate an initial SMILES representation by a molecular structure conversion method; (2) ChemBERTa encoding: ChemBERTa encode the processed SMILES representation to generate a molecular embedding vector suitable for multi-objective regression prediction; (3) Model construction and training: The backbone ChemBERTa model structure adopts a Transformer encoder architecture that inherits BERT and is pre-trained with a chemical dataset to optimize the representation ability of chemical semantics; (4) Perform regression prediction through a multi-objective regression model and use weighted mean square error loss to supervise the regression prediction results.

Images