Virtual screening method of small-molecule inhibitor targeting YTHDF1 protein

Abstract

The invention belongs to the field of drug design, and particularly discloses a virtual screening method of a small-molecule inhibitor targeting YTHDF1 protein, which comprises the following steps: acquiring a YTHDF1 protein structure, and carrying out homologous modeling to generate various conformation proteins; small molecules combined with the YTHDF1 protein are obtained, bait molecules are generated, and ligand molecules are formed; a ligand molecule is in butt joint with a binding site on the YTHDF1, and the compound is characterized through PLEC fingerprints; the method comprises the following steps: constructing a specificity scoring function by using a machine learning algorithm, carrying out scoring screening on a compound to be screened and a compound butted with YTHDF1 molecules, carrying out affinity prediction, and screening out the small-molecule inhibitor targeting YTHDF1. According to the virtual screening method provided by the invention, the YTHDF1 targeting molecules with high affinity can be accurately screened on a large scale, candidate molecules are provided for research and development of medicines for treating YTHDF1 related diseases, and a foundation is laid for clinical tests and medicine optimization.

Description

Technical Field

[0001] The present invention relates to the technical field of drug design, and in particular to a virtual screening method for small molecule inhibitors targeting YTHDF1 protein. Background Art

[0002] Tumor is one of the major diseases that threaten human health. Its occurrence and development not only involve gene mutation and deletion, but are also closely related to the imbalance of epigenetic regulation. Epigenetic modification mainly regulates gene expression through pathways such as DNA methylation, histone modification and RNA modification, and abnormalities in these modifications often trigger the occurrence and progression of tumors. N6-methyladenosine (m6A) is one of the most common RNA modifications, which regulates gene expression by affecting the splicing, translation and degradation of mRNA. Its recognition depends on m6A reader proteins, and YTHDF1 is one of the key reader proteins, which is widely involved in the pathological processes of diseases such as inflammation and cancer.

[0003] The key role of YTHDF1 in cancer and its potential as a potential therapeutic target have prompted researchers to try to develop new anti-tumor drugs by screening its small molecule inhibitors. However, the number of existing YTHDF1 inhibitors is limited, and their screening and identification processes are relatively complicated. Ebselen is the first reported small molecule inhibitor that can bind to the YTHDF1 protein and interfere with its recognition function for m6A-modified RNA. However, the mode of action of Ebselen depends on environmental factors (such as a reducing environment) and may interact with YTHDF1 in a reversible manner under certain conditions. Therefore, it is of great significance to further search for and develop efficient and specific YTHDF1 inhibitors.

[0004] Traditional drug screening usually uses scoring functions based on force fields, knowledge bases or empirical formulas. Although these methods can identify some active molecules, they cannot fully capture the complex interactions in protein-ligand binding, especially in the case of protein structural diversity and binding site flexibility. These limitations significantly reduce the efficiency and accuracy of virtual screening methods.

[0005] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0006] In view of the above-mentioned deficiencies in the prior art, the object of the present invention is to provide a virtual screening method for small molecule inhibitors targeting YTHDF1 protein, aiming to solve the problem of how to effectively improve the accuracy and efficiency of screening small molecule inhibitors targeting YTHDF1.

[0007] The technical solution of the present invention is as follows:

[0008] The present invention provides a virtual screening method for small molecule inhibitors targeting YTHDF1 protein, comprising the steps of:

[0009] (a) YTHDF1 protein structure preparation: The crystal structure of YTHDF1 protein was obtained, and homology modeling of YTHDF1 protein was performed using the SWISS-MODEL platform to generate YTHDF1 protein with multiple conformations;

[0010] (b) Construction of ligand molecule data set: Active small molecules and inactive small molecules that bind to YTHDF1 protein are obtained, and inactive bait molecules are generated based on the active small molecules to construct a ligand molecule set;

[0011] (c) Molecular docking and complex characterization: high-throughput molecular docking is performed on the molecules in the ligand molecule set and the binding sites on the YTHDF1 protein in various conformations using a molecular docking program, and the generated protein-ligand complex is characterized by PLEC fingerprinting;

[0012] (d) Construction of specificity scoring function: Based on the obtained PLEC fingerprint dataset, a specificity scoring function for YTHDF1 protein was constructed using a machine learning algorithm;

[0013] (e) Virtual screening and affinity prediction: The compounds in the compound library to be screened are docked onto YTHDF1 proteins of various conformations to obtain the PLEC fingerprint of the complex. After scoring and sorting using the specific scoring function, affinity calculation is performed to finally screen out the small molecule inhibitors targeting the YTHDF1 protein.

[0014] Preferably, in step (a), the crystal structure of the YTHDF1 protein is from the Protein Data Bank database, numbered 4RCJ, and up to ten different conformations of the YTHDF1 protein are generated by homology modeling.

[0015] Traditional virtual screening often only performs molecular docking based on proteins with a single conformation, ignoring the dynamics and conformational changes of proteins. The present invention generates 10 different conformations by homology modeling of the YTHDF1 protein. These conformations represent the natural flexibility and dynamics of the protein, covering the dynamic changes of the protein, so that molecular docking and virtual screening can fully consider the performance of the protein in different states, and can more comprehensively evaluate the binding ability of the compound to the YTHDF1 protein, greatly improving the accuracy and credibility of the screening results.

[0016] Optionally, in step (b), the active small molecules and inactive small molecules that bind to the YTHDF1 protein are from PubChem, BindingDB and Chembl databases, and duplicate molecules are removed according to the SMILE string.

[0017] Preferably, in step (b), inactive bait molecules are generated based on active small molecules, specifically comprising: using the DeepCoy algorithm to generate inactive bait molecules with similar physical and chemical properties to the active small molecules but different structures, using the SMILES of the active small molecules as input, first generating 100 bait molecules based on each active small molecule, and then retaining 50 bait molecules for each active small molecule based on the degree of embedding of the active small molecules in space.

[0018] Normally, directly obtaining active molecules and inactive molecules may only rely on simple structural similarities or collecting known molecules from the literature. However, the present invention uses the deep learning tool DeepCoy to generate inactive bait molecules with dissimilar structures but matching physical and chemical properties. These bait molecules are used to expand the data set of inactive molecules to establish a large-scale equilibrium data set containing active molecules and inactive molecules, thereby ensuring that the model has sufficient robustness and discrimination when identifying active small molecules, and further enhancing the ability to distinguish between active and inactive molecules. Therefore, the bait molecules generated by DeepCoy in the present invention can effectively improve the machine learning model's ability to discriminate active small molecules, making virtual screening more challenging, thereby improving the model's ability to capture real active molecules.

[0019] Preferably, in step (c), the molecular docking program includes the Smina molecular docking program, and in the Smina molecular docking program, the exhaustiveness parameter is set to 8 and the num_modes value is set to 10.

[0020] Preferably, in step (c), when the protein-ligand complex is characterized by using PLEC fingerprint, the parameter setting of the PLEC fingerprint includes: a cutoff distance of The protein depth is 4, the ligand depth is 2, and the vector length is 1092.

[0021] Unlike traditional 2D or 3D structural features, the present invention uses extended connectivity fingerprints to perform high-dimensional feature characterization of protein-ligand complexes. The PLEC fingerprint feature can capture the spatial information of protein-ligand interactions and represent the structural details of the complex with a 2048-dimensional high-dimensional feature vector. This method is more representative than traditional molecular descriptors and can capture more complex molecular interactions. Therefore, through the use of PLEC features, these high-dimensional features can more comprehensively characterize the spatial configuration and interaction pattern of protein-ligand complexes, providing high-quality input features for subsequent machine learning models, thereby having higher accuracy and reliability than traditional virtual screening methods.

[0022] Optionally, in step (d), the machine learning algorithm includes at least one of random forest, extreme gradient boosting, support vector machine and artificial neural network.

[0023] The present invention proposes to construct a scoring function by combining at least one machine learning method among random forest (RF), extreme gradient boosting (XGBoost), support vector machine (SVM) and artificial neural network (ANN). It can also combine multiple machine learning algorithms. After different models are trained through multiple rounds of cross-validation, an integrated learning method is adopted to select the optimal model as the final scoring function. The strategy of multi-model integration can improve the generalization ability and prediction accuracy of the model.

[0024] Preferably, the machine learning algorithm is an artificial neural network, and the artificial neural network is trained and evaluated by cross-validation.

[0025] ANN has significant advantages in processing complex protein-ligand interactions, automatic feature extraction, processing large-scale data, and generalization capabilities. ANN automatically extracts complex features and patterns through multi-layer neurons and nonlinear activation functions, which is very important in the YTHDF1 protein-ligand complex. This is because the three-dimensional structure of the complex and its interaction with molecules usually have highly complex nonlinear characteristics. ANN extracts meaningful information from these complex structural relationships, thereby improving the accuracy of prediction. In order to improve the generalization ability of the model, the present invention selects an artificial neural network model for training as a scoring function for predicting the activity of the compound, which can more accurately predict the binding activity of the compound with the YTHDF1 target, significantly improving the effectiveness of virtual screening.

[0026] Optionally, in step (e), the compound library to be screened is selected from the ZINC20 database, comprising 45,000 small molecule compounds.

[0027] Optionally, in step (e), the complexes are scored and ranked using the specificity scoring function to screen out the top 1% of candidate compounds, and then the affinity of the candidate compounds is calculated using a logistic regression model to ultimately screen out small molecule inhibitors targeting YTHDF1 with high affinity.

[0028] According to the specific scoring function constructed above, 45,000 compounds in the compound library to be screened were screened for activity. The specific scoring function is based on the previously trained ANN model, which is specifically optimized for the target protein and can more accurately predict the activity of the compound. In order to further confirm the reliability of the screening results and the practical application potential of the compounds, the present invention also predicts the binding affinity of the top 1% of the screened molecules. The binding affinity prediction can help evaluate the binding strength between these compound molecules and the target protein, and then judge their drug development potential. Therefore, introducing affinity prediction into the later stage of virtual screening, in conjunction with the scoring function, can ensure that the screened molecules not only have high affinity, but also have good stability.

[0029] Beneficial effects:

[0030] The present invention provides a virtual screening method for small molecule inhibitors targeting YTHDF1 protein, wherein the method uses homology modeling of multi-conformation proteins and uses the DeepCoy algorithm to generate bait molecules, thereby enhancing the discrimination and capture capabilities of the specificity scoring function for active small molecules, thereby enhancing the diversity and accuracy of virtual screening. Moreover, by introducing PLEC fingerprint features and combining them with machine learning algorithms, the generalization ability and predictive performance of the target-specific scoring function are significantly improved. At the same time, affinity prediction is introduced into the later stage of virtual screening, and cooperates with the specificity scoring function to ensure that the screened molecules not only have high affinity but also have good stability.

[0031] Therefore, the present invention proposes an effective virtual screening method for small molecule inhibitors targeting YTHDF1 protein by combining advanced technologies of bioinformatics, machine learning and cheminformatics. It can screen out YTHDF1 targeting molecules with high affinity on a large scale and accurately, providing important candidate molecules for the drug development of YTHDF1-related diseases (such as cancer), and laying a solid foundation for subsequent clinical trials and drug optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is the chemical structural formula of the small molecule inhibitor targeting YTHDF1 protein finally screened out in the examples of the present invention.

[0033] Figure 2 This is a docking diagram of the small molecule inhibitor targeting YTHDF1 protein and YTHDF1 protein that was finally screened out in the examples of the present invention. DETAILED DESCRIPTION

[0034] The present invention provides a virtual screening method for small molecule inhibitors targeting YTHDF1 protein. In order to make the purpose, technical scheme and effect of the present invention clearer and more specific, the present invention is further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0035] Unless otherwise defined, all technical terms and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0036] The following describes it in detail through specific embodiments.

[0037] Example 1

[0038] (1) Preparation of protein structure

[0039] Protein structure preparation: The crystal structure of YTHDF1 (PDB ID: 4RCJ) was selected as the target, and the protein structure was further optimized by homology modeling. The SWISS-MODEL platform was used to generate up to 10 different conformations of YTHDF1 proteins, which represent the natural flexibility and dynamics of the protein, can capture the different activity states of the protein, and ensure that the conformational diversity of YTHDF1 can be taken into account during molecular docking. This step ensures that the model is not limited to a single conformation, and enhances the reliability and pertinence of small molecule screening.

[0040] As shown in Table 1, the YTHDF1 protein encoding 4RCJ was selected because it contains no missing key amino acids in the loop region that binds to mRNA.

[0041] Table 1

[0042]

[0043]

[0044] Based on the YTHDF1 protein of 4RCJ, 10 other conformations of YTHDF1 protein were constructed by homology modeling, with a total of 11 conformational protein structures, as shown in Table 2:

[0045] Table 2

[0046]

[0047] RMSD (root mean square deviation) is a key indicator for evaluating the structural similarity between proteins and ligands. First, the data is gradually filled based on the RMSD values ​​of proteins with various conformations and 4RCJ, which means that the protein-ligand complex closest to the 4RCJ structure is preferentially selected to ensure the representativeness and diversity of the complexes in the data set. At the same time, the application of RMSD helps to select those conformations that are closest to the experimental structure to enhance the accuracy of subsequent models.

[0048] (2) Construction of ligand molecule set

[0049] The molecules that bind to the YTHDF1 protein were searched from Pubchem, BindingDB, and Chembl databases. These molecules were reported in previous articles. They were divided into active small molecules (30) and inactive small molecules (8) according to their activity, and duplicate molecules were removed according to the SMILE string.

[0050] In order to supplement the number of inactive molecules, better train the scoring function MLSFs, and improve the model's ability to distinguish between active and inactive molecules, the DeepCoy algorithm is used to generate decoy molecules with very similar physical and chemical properties to active molecules but different structures using a graph-generated neural network. Using the SMILES of active molecules as input, 100 decoy molecules are generated for each active small molecule. Finally, according to the degree of embedding of the active small molecule in space, 50 decoy molecules are retained for each active small molecule. The constructed ligand molecule set includes an active molecule set and an inactive molecule set, among which the inactive molecule set includes inactive small molecules and decoy molecules.

[0051] (3) Molecular docking and complex characterization

[0052] In order to generate the interaction patterns of all molecules and cGAS or kRAS, all protein-ligand complexes can be generated using molecular docking programs. Autodock Vina is a well-known molecular docking software, and Smina is its optimized version with better speed and stronger scalability.

[0053] The Smina molecular docking program was used to perform molecular docking on the YTHDF1 protein structures of 11 conformations and the molecules in the ligand molecule set. The ligand search space was defined by the mRNA coordinates bound by the YTHDF1 protein, that is, the molecules were induced to bind at the best binding site through the mRNA coordinates bound by the YTHDF1 protein itself. In the Smina molecular docking program, the exhaustiveness parameter was set to 8, which enables a more detailed search for binding modes. The value of num_modes was set to 10, and multiple conformational complex structures were generated for each molecule docked with the YTHDF1 protein for subsequent training of the enhanced model. When searching and arranging poses, Smina used gradient and Monte Carlo steps, and the final docked poses were saved in SDF format.

[0054] Then, select specific PLEC fingerprint (Protein-Ligand Extended Connectivity Fingerprints) parameter settings, including: cutoff distance The protein depth is 4, the ligand depth is 2, and the vector length is 1092 to ensure accurate modeling of the spatial conformation of the complex. The PLEC fingerprint encodes the extended connectivity environment between atoms within a specific distance range by capturing the spatial correlation between the protein and the ligand. This fingerprint is based on the neighborhood information of the molecular graph and encodes the structural information of each atom and its surroundings in the protein and ligand into a fixed-length binary vector.

[0055] (4) Constructing a target-specific scoring function

[0056] Based on the obtained PLEC fingerprint dataset of protein-ligand complex, an ANN (artificial neural network) was used to construct a specific scoring function for YTHDF1.

[0057] ANN has significant advantages in dealing with complex protein-ligand interactions, automatic feature extraction, processing large-scale data, and generalization capabilities, making it an ideal choice for building efficient scoring functions in virtual screening tasks. ANN automatically extracts complex features and patterns through multiple layers of neurons and nonlinear activation functions. This is particularly important in the YTHDF1 protein-ligand complex, because the three-dimensional structure of the complex and its interactions with molecules usually have highly complex nonlinear characteristics. ANN is good at extracting meaningful information from these complex structural relationships, thereby improving the accuracy of predictions. Compared with traditional machine learning models that rely on manually extracted features, ANN can automatically learn deep features from input data through a multi-layer network structure. This automatic feature extraction capability reduces the reliance on domain knowledge and can capture potential features in protein-ligand interactions, such as the spatial distribution of binding sites and interactions between atoms. At the same time, ANN can be well expanded to large-scale data sets and process thousands to millions of training samples. For the massive compounds in virtual screening tasks, ANN can perform effective training and reasoning in a large sample space, which has a significant effect on improving virtual screening efficiency and hit rate. In terms of generalization ability, after proper regularization and parameter adjustment, ANN can generalize well to unseen molecules. This means that the ANN model that performs well on the training set can also obtain good prediction results on unknown molecules in actual applications, thereby improving the practicality of the model.

[0058] Based on the performance of the model trained with different amounts of data, the present invention chooses to add all 11 sets of protein-ligand complexes with different conformations to the data set to improve the generalization ability of the model. Compared with using only a single protein structure, this data enhancement strategy significantly improves the prediction accuracy of the model. By adding data from multiple protein complexes, the model can learn more differences between different complexes, thereby improving its prediction ability for unseen data.

[0059] The data set is divided into a training set and a test set according to the PLEC fingerprint similarity of the complex. The diversity of compounds in the training set and the test set is ensured by a 4:1 ratio division. The PLEC fingerprint division method avoids the problem that the training set and the test set may be too similar due to traditional random division, so that the model can generalize better. In the model training stage, classification tasks and regression tasks are trained separately. The goal of the classification task is to distinguish between active molecules and inactive molecules in the compound library. By learning the characteristics of the molecules, the model can quickly screen out potential active molecules. The regression task focuses more on the prediction of the binding affinity of active molecules and quantifies the strength of their interaction with the target protein. Through this combination of classification and regression, the activity of the molecule can be more comprehensively evaluated, and its actual level of biological activity can be further predicted.

[0060] (5) Virtual screening and affinity prediction

[0061] In the virtual screening stage, 45,000 compound molecules in the ZINC20 database were used for high-throughput molecular docking with 11 conformational YTHDF1 proteins. As a widely used public compound library, the ZINC20 database provides a rich resource of screenable molecules, which is suitable for virtual screening tasks in drug development.

[0062] According to the target-specific scoring function constructed previously, the complexes generated after molecular docking were scored and sorted, and potential active compounds were preliminarily screened out. After sorting according to the probability (0-1) of active molecules evaluated by the scoring function, a total of 971 active molecules were screened out from 45,000 compounds, achieving a hit rate of about 2.1%. This is a relatively significant result in the virtual screening task.

[0063] In order to further confirm the reliability of the screening results and the practical application potential of the compounds, the binding affinity of the top 1% of these candidate active molecules was predicted by training a logistic regression model. Binding affinity prediction can help evaluate the binding strength between these molecules and the target protein, further verify their stability and affinity for binding to YTHDF1, and then determine their drug development potential.

[0064] Finally, the 10 most promising small molecule inhibitors of YTHDF1 were screened out. The chemical structures of these 10 small molecule inhibitors targeting YTHDF1 protein are as follows: Figure 1 The corresponding predicted affinity values ​​are shown in Table 3:

[0065] Table 3

[0066]

[0067]

[0068] Figure 2 The docking diagram of the screened small molecule inhibitor targeting YTHDF1 protein and YTHDF1 protein, combined with the affinity prediction results in Table 3, reflects the energy state and interaction mode of the binding of the final screened molecular compound with YTHDF1 protein. The specific analysis is as follows:

[0069] The predicted affinity value of ZINC00024551 was 8.825, and this molecule formed hydrogen bonds and hydrophobic interactions with the amino acid residues around the site. Figure 2 It is shown in the figure that the molecule forms π-π stacking interactions with aromatic amino acids (such as tryptophan, tyrosine, etc.), which help to improve the binding stability of the molecule.

[0070] The predicted affinity value of ZINC00025054 is 9.555, which is the highest predicted affinity, showing strong binding ability. This may be due to the fact that the molecule forms hydrogen bonds with multiple polar residues and may interact with the carbonyl groups of the protein backbone. In addition, it also forms favorable van der Waals forces with the surrounding hydrophobic amino acids.

[0071] The predicted affinity value of ZINC00028212 is 9.058, and this molecule is likely to bind to the vicinity of the catalytic residues through multiple polar interactions. Figure 2 Its location in the mitochondria is shown to be close to key catalytic residues, suggesting that the molecule may inhibit the function of YTHDF1 by hindering RNA binding.

[0072] The predicted affinity value of ZINC00039097 is 9.080, Figure 2 The molecule is shown to bind to a key lysine or arginine residue via hydrogen bonds and these interactions play a key role in stabilizing the binding of the inhibitor in the active site.

[0073] The predicted affinity value of ZINC00057125 is 9.484, and this molecule interacts with both polar residues and hydrophobic regions. Figure 2 suggests that it may occupy an important region for RNA recognition, thereby effectively competing with RNA for binding.

[0074] The predicted affinity value of ZINC00057311 is 9.045, and the molecule may be stabilized in the binding site through hydrogen bonds and hydrophobic interactions, especially the conserved residues around the target region, to play a role in blocking the RNA binding function of YTHDF1.

[0075] The predicted affinity value of ZINC00057394 is 9.107, Figure 2 The results show that the molecule interacts with both hydrophobic and polar residues in the binding site, indicating that it may affect the function of YTHDF1 by binding to the RNA binding region.

[0076] The predicted affinity value of ZINC00079466 was 9.176, and the molecule tightly bound to key residues of the binding site through hydrogen bonds and hydrophobic interactions, suggesting that it may inhibit YTHDF1 by competing for the binding of natural RNA.

[0077] The predicted affinity of ZINC00105325 is 9.175, Figure 2 It is shown that the molecule forms multiple hydrogen bonds and van der Waals interactions, tightly binds to the active site, and exhibits strong binding ability.

[0078] The predicted affinity of ZINC00112982 is 8.983, and this molecule forms a complex network of hydrogen bonds and non-polar interactions with residues in the RNA binding domain, which may alter the ability of the YTHDF1 protein to bind its natural substrate.

[0079] According to the structural characteristics of the above 10 molecules, they can be divided into the following five categories:

[0080] (1) Nucleoside analogs: This type of compound has a nucleoside-like structure, usually including sugars and bases, and is easy to interact with the RNA binding site of the YTHDF1 protein, competitively inhibiting the binding of m6A-modified RNA. This type of molecule has a simple structure and often interacts with the target through hydrogen bonds and π-π stacking. For example, ZINC00112982: has a pyrimidine base and sugar structure. ZINC00057311: contains a similar nucleoside structure. ZINC00039097: Its sugar and aromatic base structure has great similarity with the RNA binding domain of the YTHDF1 protein.

[0081] (2) Aromatic rings and their derivatives: This class of compounds contains multiple aromatic rings or heterocyclic structures, which bind to the aromatic residues on the YTHDF1 binding site through π-π interactions and hydrophobic interactions. The structure is relatively complex, and the inhibitory effect is improved through multi-point binding. For example, ZINC00057125: contains multiple aromatic rings and hydroxyl groups. ZINC00079466: has a complex structure, contains multiple heterocyclic rings, and can interact with multiple regions of the protein.

[0082] (3) Sugar derivatives: These molecules contain sugar molecules with multiple hydroxyl groups, which can form stable interactions with polar residues in the YTHDF1 binding domain through hydrogen bonds. For example, ZINC00057394: The sugar backbone is bound to multiple hydroxyl groups. ZINC00024551: The sugar and alkynyl structures provide flexibility. ZINC00028212: The phenyl group is attached to the sugar backbone, which may form hydrophobic and hydrogen bonding interactions with YTHDF1.

[0083] (4) Molecules containing carboxyl and amide groups: These compounds usually contain carboxyl and amide groups, bind to YTHDF1 through hydrogen bonds, and interfere with its recognition of m6A-modified RNA. For example, ZINC00025054: contains carboxyl and sugar groups, which can form hydrogen bonds with YTHDF1. ZINC00105325: contains a bicyclic and sugar structure, providing multiple interaction points.

[0084] (5) Molecules containing alkyne groups: The alkyne structure gives these compounds flexibility and allows them to adapt to the shape of the protein binding pocket, thereby interfering with the function of the protein. For example, ZINC00024551: its alkyne group provides a special interaction with the YTHDF1 binding pocket.

[0085] It can be seen that the present invention not only significantly improves the accuracy and efficiency of YTHDF1 small molecule inhibitor screening through the virtual screening method, but also provides a new technical path for drug development.

[0086] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A virtual screening method for small molecule inhibitors targeting YTHDF1 protein, characterized in that: Includes steps: (a) YTHDF1 protein structure preparation: The crystal structure of YTHDF1 protein was obtained, and homology modeling of YTHDF1 protein was performed using the SWISS-MODEL platform to generate YTHDF1 protein with multiple conformations; (b) Construction of ligand molecule data set: Active small molecules and inactive small molecules that bind to YTHDF1 protein are obtained, and inactive bait molecules are generated based on the active small molecules to construct a ligand molecule set; (c) Molecular docking and complex characterization: high-throughput molecular docking is performed on the molecules in the ligand molecule set and the binding sites on the YTHDF1 protein in various conformations using a molecular docking program, and the generated protein-ligand complex is characterized by PLEC fingerprinting; (d) Construction of specificity scoring function: Based on the obtained PLEC fingerprint dataset, a specificity scoring function for YTHDF1 protein was constructed using a machine learning algorithm; (e) Virtual screening and affinity prediction: The compounds in the compound library to be screened are docked onto YTHDF1 proteins of various conformations to obtain the PLEC fingerprint of the complex. After scoring and sorting using the specific scoring function, affinity calculation is performed to finally screen out the small molecule inhibitors targeting the YTHDF1 protein.

2. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (a), the crystal structure of the YTHDF1 protein is from the Protein Data Bank database, numbered 4RCJ, and ten YTHDF1 proteins with different conformations are generated by homology modeling.

3. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (b), active small molecules and inactive small molecules that bind to the YTHDF1 protein are from PubChem, BindingDB, and Chembl databases, and duplicate molecules are removed based on the SMILE string.

4. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (b), inactive decoy molecules are generated based on the active small molecules, specifically including: using the DeepCoy algorithm to generate inactive decoy molecules with similar physical and chemical properties to the active small molecules but different structures, using the SMILES of the active small molecules as input, first generating 100 decoy molecules based on each active small molecule, and then retaining 50 decoy molecules for each active small molecule based on the degree of embedding of the active small molecules in space.

5. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (c), the molecular docking program includes the Smina molecular docking program. In the Smina molecular docking program, the exhaustiveness parameter is set to 8 and the num_modes value is set to 10.

6. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (c), when the protein-ligand complex is characterized by using PLEC fingerprint, the parameter settings of the PLEC fingerprint include: a cutoff distance of The protein depth is 4, the ligand depth is 2, and the vector length is 1092.

7. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (d), the machine learning algorithm includes at least one of random forest, extreme gradient boosting, support vector machine and artificial neural network.

8. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (d), the machine learning algorithm is an artificial neural network, and the artificial neural network is trained and evaluated by cross-validation.

9. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (e), the compound library to be screened is selected from the ZINC20 database, which contains 45,000 small molecule compounds.

10. The virtual screening method for small molecule inhibitors targeting YTHDF1 protein according to claim 1, characterized in that: In step (e), the complexes are scored and ranked using the specificity scoring function to screen out the top 1% of candidate compounds, and the affinity of the candidate compounds is calculated using a logistic regression model to ultimately screen out small molecule inhibitors targeting YTHDF1 with high affinity.

Images