Ultrahigh-flux drug small molecule virtual screening method, storage medium and equipment

Abstract

The invention relates to an ultrahigh-flux drug small molecule virtual screening method, a storage medium and equipment, and the method comprises the following steps: obtaining a target protein structure and a small molecule compound library to be screened, and carrying out preliminary screening from the small molecule compound library to obtain a first small molecule compound set; adopting at least more than two check tools to screen from the first small molecule compound set to obtain a second small molecule compound set; performing molecular dynamics simulation on the target protein structure, extracting a plurality of representative conformations from a molecular dynamics simulation trajectory, and screening from the second small molecule compound set based on the representative conformations to obtain a third small molecule compound set; performing clustering analysis on the third small molecule compound set, and selecting representative small molecule compounds as candidate compounds; and selecting a compound with relatively low binding free energy as a final screening result. Compared with the prior art, the method has the advantages that the screening efficiency and precision of small molecule compounds in the new drug development process can be improved, and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of drug research and development, and in particular to an ultra-high throughput virtual screening method for small drug molecules, a storage medium and a device. Background Art

[0002] Drug development is a complex and time-consuming process, and early lead compound screening is one of the key steps in determining the success of new drug development. In order to find lead compounds with druggable potential, researchers usually need to screen a small number of candidate compounds from a library of millions or even billions of small molecule compounds. Existing high-throughput screening (HTS) and virtual screening (VS) technologies still face many challenges when facing large-scale compound libraries, which limits their practical application effects.

[0003] 1. Limitations of high-throughput screening

[0004] Traditional high-throughput screening relies on automated screening systems in the laboratory, which can screen hundreds of thousands to millions of small molecule compounds in a short period of time. However, this method has significant limitations, including high costs and complex experimental conditions. For example, high-throughput screening usually requires expensive automated equipment and a large number of reagents, and the false positive rate of the screening results is high, resulting in a significant increase in the workload and cost of subsequent verification. For compound libraries of hundreds of millions, experimental screening faces the problems of long time and high cost, which restricts its wide application. Although high-throughput screening can screen out some candidate compounds to a certain extent, its efficiency and cost-effectiveness are still insufficient when faced with ultra-large-scale compound libraries.

[0005] 2. Limitations of Virtual Screening

[0006] Virtual screening technology relies on computer simulation to screen large-scale compound libraries to reduce experimental costs and improve screening efficiency. However, existing virtual screening technologies are usually based on a single protein crystal structure for molecular docking. This method ignores the dynamic changes of proteins and the existence of multiple conformational states. Proteins are highly flexible and have multiple conformational states under physiological conditions, which has a significant effect on the binding ability of small molecule compounds. Ignoring the diversity of protein conformations may cause the screened candidate compounds to perform poorly in actual biological environments, reducing the reliability and hit rate of screening.

[0007] 3. Impact of protein conformational diversity on screening results

[0008] Proteins usually have multiple conformational states in the body, and the flexibility and dynamic changes under physiological conditions significantly affect the binding effect of small molecule compounds. Traditional virtual screening usually performs molecular docking based on a single crystal structure, ignoring the dynamics and conformational diversity of proteins, which may lead to inconsistency between screening results and binding conditions in actual biological environments. In order to improve the reliability of screening, researchers have tried to generate a variety of representative protein conformations through molecular dynamics simulations to reflect the true conformational changes of proteins as comprehensively as possible. This method significantly improves the accuracy and reliability of screening, but it also increases the demand for computing resources. How to balance computing resources and screening accuracy is still a major challenge in current virtual screening.

[0009] 4. Further optimization and evaluation of small molecule compounds after screening

[0010] After the initial screening of potential small molecule compounds, they need to be further refined to reduce the risk of experimental verification failure. Traditional energy evaluation methods are usually based on molecular docking scores, but this cannot accurately predict the binding free energy of compounds to proteins, thus affecting the screening quality of candidate compounds. Binding free energy calculations (such as molecular dynamics binding free energy calculations, MM-PBSA, and MM-GBSA methods) can more accurately evaluate the binding strength between compounds and targets. However, these methods have high computational resource requirements.

[0011] In recent years, researchers have paid more and more attention to the screening of ultra-large small molecule libraries. Improving screening efficiency and accuracy through advanced computing technology and screening strategies is an important research and development direction in this field. Although high-throughput screening and virtual screening technologies have achieved some results, their limitations are still significant. To overcome these challenges, it is necessary to develop a more efficient ultra-high-throughput virtual screening method. Summary of the invention

[0012] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide an ultra-high throughput virtual screening method, storage medium and equipment for small molecule drug screening, which can effectively improve the efficiency and success rate of large-scale small molecule screening through multi-level optimization screening process, and provide more efficient technical support for the discovery of lead compounds in new drug development.

[0013] The purpose of the present invention can be achieved by the following technical solutions:

[0014] An ultra-high throughput virtual screening method for small drug molecules comprises the following steps:

[0015] Obtaining a target protein structure and a small molecule compound library to be screened, analyzing the small molecule binding pocket of the target protein structure, and obtaining a first small molecule compound set through a preliminary screening of the small molecule compound library based on the target protein structure;

[0016] Using at least two or more selection tools to finely screen the first small molecule compound set to obtain a second small molecule compound set;

[0017] Performing molecular dynamics simulation on the target protein structure, extracting multiple representative conformations from the molecular dynamics simulation trajectory, and screening the second small molecule compound set to obtain a third small molecule compound set based on the representative conformations;

[0018] performing cluster analysis on the third small molecule compound set, and selecting representative small molecule compounds in each cluster as candidate compounds;

[0019] The binding free energy of each candidate compound with the target protein structure is calculated, and compounds with lower binding free energy are selected as the final screening results.

[0020] Furthermore, the target protein structure is pre-processed, and the pre-processing includes removing water molecules / non-essential ligands and adding essential hydrogen atoms.

[0021] Furthermore, the small molecule compound library is preprocessed, and the preprocessing includes adding necessary hydrogen atoms and generating a compound library file format.

[0022] Furthermore, the virtual screening toolkit in the VFVS software was used to screen and obtain the first small molecule compound set, the second small molecule compound set and the third small molecule compound set.

[0023] Furthermore, the extraction of multiple representative conformations is specifically as follows:

[0024] Cluster analysis is performed on the amino acids around the binding site of the target protein structure, and the molecular dynamics simulation trajectory data are divided into several categories. The most representative structure in each category is selected as the representative conformation of the category.

[0025] Furthermore, performing cluster analysis on the third small molecule compound set specifically includes: performing cluster analysis on the third small molecule compound set using a small molecule clustering algorithm according to chemical structure characteristics and molecular properties.

[0026] Furthermore, the method further comprises:

[0027] The final screening results were experimentally verified.

[0028] Furthermore, it is determined whether each candidate compound has an unfavorable structural annotation. Based on the actual experimental situation, compounds with unfavorable structural annotations are preferentially excluded and then experimental verification is performed.

[0029] The present invention also provides a computer-readable storage medium, comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the ultra-high throughput virtual screening method for small drug molecules as described above.

[0030] The present invention also provides an electronic device comprising one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the ultra-high throughput drug small molecule virtual screening method as described above.

[0031] The present invention adopts a multi-stage screening strategy, combines a variety of computational tools and molecular simulation technologies, gradually narrows the screening scope, and optimizes the selection process of candidate compounds; generates a variety of representative protein conformations through molecular dynamics simulation, fully considers the dynamic changes of proteins in the physiological environment, and combines with free energy calculation and other sophisticated evaluation methods to significantly improve the reliability and success rate of screening, which can solve the shortcomings of existing screening technologies in screening speed, protein structure prediction accuracy and subsequent experimental verification, and improve the efficiency and accuracy of small molecule screening in the process of new drug development.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. Ultra-high throughput screening strategy

[0034] The present invention adopts a multi-stage screening strategy to adapt to the screening needs of ultra-large-scale compound libraries (e.g., millions to hundreds of millions). In the preliminary screening stage, virtual screening tools are combined to quickly screen out potential active compounds through molecular docking and scoring functions. On this basis, the number of compounds is gradually reduced, and the screening criteria are further refined to improve the screening efficiency.

[0035] 2. Considerations of multi-conformational protein structures

[0036] Traditional virtual screening usually performs molecular docking based on a single static protein structure, which ignores the diversity and flexibility of proteins under physiological conditions. The present invention uses molecular dynamics simulation technology to generate a variety of representative protein conformations, as well as a variety of known structures of the same protein, to fully reflect the dynamic changes of proteins in the body, so that the screening results are more in line with the actual biological situation. By introducing proteins with multiple conformations for molecular docking, the accuracy of screening can be significantly improved and the false positive rate in the screening process can be reduced.

[0037] 3. Optimize the screening process in stages

[0038] The present invention adopts a staged screening process, and different software tools and parameter settings are applied in each stage to make the screening process more flexible and efficient. In the first stage, a fast virtual screening tool is used for preliminary screening to filter out most of the invalid compounds. In the second stage, more detailed docking and scoring are performed to further narrow the screening range. In the third stage, molecular dynamics and cluster analysis are combined to obtain multiple conformations of protein structures for multi-software screening analysis to screen out small molecules with binding potential in multiple conformations. Finally, in the fourth stage, molecular dynamics simulation and binding free energy calculation are used to further optimize the selection of candidate compounds, and small molecule compounds with relatively stable RMSD and low binding free energy are preferentially screened.

[0039] 4. Small molecule cluster analysis and unfavorable structure annotation

[0040] In order to more efficiently analyze the large number of compounds obtained by screening, the present invention performs cluster analysis on the compounds obtained in the third stage of preliminary screening to select representative candidate compounds. At the same time, small molecules with unfavorable structures are marked so that they can be focused on in the subsequent experimental verification process. Cluster analysis and unfavorable structure annotation can help narrow the scope of verification and improve the success rate of verification.

[0041] 5. Elaborate evaluation of binding free energy

[0042] In the final screening stage, the present invention quantitatively evaluates candidate compounds through accurate binding free energy calculations (such as MM-PBSA and MM-GBSA methods) to screen out small molecule compounds with low binding free energy. These calculation methods can better reflect the binding ability between the compound and the target, thereby improving the reliability of the screening results and the success rate of experimental verification, while maximizing resource utilization and improving the success rate of screening.

[0043] 6. Integration and optimization of multiple software tools

[0044] The method of the present invention can integrate a variety of computing software (such as VirtualFlow, AutoDock Vina, qvina, qvina-w, smina, gwovina, gnina, GROMACS, Amber, etc.), and flexibly adjust parameter settings according to the screening requirements at different stages to maximize the screening efficiency. Through the combined use of multiple tools, the different characteristics of the compound can be comprehensively evaluated and screened.

[0045] The method of the present invention is not only suitable for the screening of large-scale compound libraries, but also can significantly improve the screening accuracy of candidate compounds, providing strong technical support for the discovery of lead compounds in new drug development. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the process of the present invention;

[0047] Figure 2 This is the structural diagram of NONO protein;

[0048] Figure 3 The predicted small molecule binding pocket of NONO protein;

[0049] Figure 4 This is a schematic diagram of the preliminary first round screening results in an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the second round of screening results in an embodiment of the present invention;

[0051] Figure 6 The molecular dynamics results of NONO protein;

[0052] Figure 7 This is the clustering result of NONO protein dynamic conformation;

[0053] Figure 8 The results are the molecular dynamics of representative small molecules in receptor-ligand complexes;

[0054] Fig. 9 Representative small molecules interact with NONO protein. DETAILED DESCRIPTION

[0055] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0056] This embodiment provides an ultra-high throughput virtual screening method for small drug molecules. Figure 1 As shown, the following steps are included:

[0057] S1. Preliminary screening stage: obtaining the target protein structure and the small molecule compound library to be screened, and obtaining the first small molecule compound set from the small molecule compound library based on the target protein structure.

[0058] In this embodiment, the Q15233 protein of the NONO gene is taken as an example, and its structure is as follows Figure 2As shown. The small molecule compound library to be screened uses a large-scale small molecule compound library, and the size of the library can reach millions to hundreds of millions of compounds. The compound library can be derived from a commercial compound database (such as Analyticon, Chembridge, ChemDiv, Enamine, InterBioScreen, Lifechemicals, Otava, Specs, Vitas-M, ZINC, etc.) or a self-built compound library.

[0059] In this embodiment, the small molecule compound library needs to be preprocessed, and the preprocessing includes adding necessary hydrogen atoms and compound formatting. Specifically, Open Babel software is used to add necessary hydrogen atoms to the small molecules and convert the compound format to pdbqt format. The converted pdbqt file needs to be organized into a directory structure that can be recognized by VirtualFlow's VFVS software.

[0060] In this example, the target protein structure also needs to be preprocessed, which includes the removal of water molecules / non-essential ligands and the addition of essential hydrogen atoms. The protein binding pocket for small molecules is predicted using the proteins.plus website, such as Figure 3 Specifically, the crystal structure or predicted structure of the target protein is prepared, and molecular editing tools (such as PyMOL or Chimera) are used to remove water molecules and non-essential ligands, and to add necessary hydrogen atoms to ensure the integrity of the structure.

[0061] The specific process of screening the first small molecule compound set is as follows: using VirtualFlow's VFVS software, calling a virtual screening toolkit (such as qvina) for preliminary screening. Setting molecular docking parameters, such as docking target, docking mode, etc., in order to speed up the screening speed, this embodiment sets the exhaustiveness parameter to 2. Using high-performance cluster servers for parallel computing, quickly perform molecular docking on compounds. Figure 4 The affinity score distribution is shown (the horizontal axis is affinity and the vertical axis is the number of ligands). According to the docking scores, the top 200,000 potential active compounds with higher scores were screened out to form the first small molecule compound set.

[0062] S2, fine screening stage: using at least two or more screening tools to screen the first small molecule compound set to obtain a second small molecule compound set.

[0063] First, the 200,000 potentially active compounds obtained from the preliminary screening were reorganized into a directory structure recognizable by VirtualFlow's VFVS software to facilitate subsequent screening.

[0064] Then, use VirtualFlow's VFVS software to call multiple virtual screening tools (such as Smina and AutoDockVina) to perform refined screening of the compounds. Set the exhaustiveness parameter to 8 to improve the screening accuracy. Cross-validate the docking results of multiple software to calculate the best ranking of small molecule compounds in multiple screening tools, such as Figure 5 As shown, about 20,000 potential active compounds were screened out based on the ranking to form the second small molecule compound set.

[0065] S3, multi-conformational protein screening stage: performing molecular dynamics simulation on the target protein structure, extracting multiple representative conformations from the molecular dynamics simulation trajectory, and screening the second small molecule compound set to obtain a third small molecule compound set based on the representative conformations.

[0066] This step specifically includes:

[0067] 3.1 Perform molecular dynamics simulation on the target protein structure to reflect its dynamic changes under physiological environment. Use molecular dynamics simulation software such as GROMACS or Amber, set appropriate simulation conditions (such as temperature 300K, pressure 1 bar, solvent as water environment, etc.), and perform long simulation (such as 200ns) to obtain the conformational diversity of the protein. Figure 6 Display molecular dynamics results.

[0068] 3.2 Extract multiple representative conformations from the molecular dynamics simulation trajectory. Perform cluster analysis on the amino acids around the binding site of the target protein. Determine the optimal cluster number k according to the Elbow Method, such as dividing the trajectory data into 89 categories, such as Figure 7 As shown, the most representative structure, i.e., representative conformation, is selected from each category as the protein template for subsequent molecular docking.

[0069] 3.3 Use VirtualFlow's VFVS software to call virtual screening tools (such as vina or smina), dock these representative conformations with each small molecule in turn, and use different software to dock the 20,000 screened compounds to different representative conformations. Calculate the average docking ranking of each compound on all protein conformations, and screen out about 2,000 small molecule compounds that meet multi-conformation binding based on the comprehensive score to form the third small molecule compound set.

[0070] S4, cluster analysis stage: performing cluster analysis on the third small molecule compound set, and selecting representative small molecule compounds in each cluster as candidate compounds.

[0071] In this embodiment, the K-means clustering algorithm is used to cluster according to chemical structure characteristics and molecular properties (such as hydrophobicity, molecular weight, number of rotating bonds, etc.), and similar small molecule compounds are classified into the same category to determine the compound population that may have a pharmacophore. Representative compounds are selected in each cluster as candidate compounds. Compounds that show good binding scores in multiple conformational dockings are preferentially selected to ensure the reliability of the screening results.

[0072] S5, binding free energy calculation stage: Calculate the binding free energy of each candidate compound with the target protein structure, and select the compound with lower binding free energy as the final screening result.

[0073] In this example, the MM-PBSA or MM-GBSA method is used to evaluate the binding free energy between the candidate compounds screened by cluster analysis and the target protein. The GROMACS or Amber software can be used in the calculation process. The stability of the binding of the small molecule in the receptor is calculated based on the GROMACS or Amber molecular dynamics results. Figure 8 According to the calculation results of binding free energy, compounds with lower binding free energy are preferentially selected as the final screening results. These compounds are more likely to show stronger binding activity in biological experiments.

[0074] In a preferred embodiment, the above method further comprises:

[0075] S6. Experimental verification: Conduct experimental verification on the final screening results.

[0076] In this embodiment, the experimental verification includes two steps of verification, specifically including:

[0077] 6.1 The candidate compounds finally screened are verified by in vitro experiments, such as evaluating their inhibitory or activating effects on the target protein through IC50 or Ki determination. Biological activity assays (such as enzyme activity inhibition tests) can be used to preliminarily verify the biological effects of the compounds.

[0078] 6.2 Compounds that have been verified by preliminary in vitro experiments should be further tested in cell or animal experiments to evaluate their biological activity and safety (such as cytotoxicity, pharmacokinetics, etc.) in more complex biological environments. In addition, further structure-activity relationship studies can be conducted to optimize the structure of the compound. Fig. 9 The binding modes of representative small molecules to NONO protein are shown.

[0079] In a preferred embodiment, after the candidate compounds are screened out by cluster analysis in step S4, structural analysis is performed on the small molecule compounds to mark compounds that may have unfavorable structures (such as strong metabolic activity or potential toxicity) so that they can be focused on in subsequent experimental verification to further improve the reliability of screening.

[0080] The above method includes the following features:

[0081] (1) Multi-stage screening strategy: Combine multiple screening tools and algorithms to gradually optimize the screening process to efficiently screen and lock in high-quality candidate small molecules.

[0082] (2) Consideration of multi-conformational protein structures: Multiple representative protein conformations are generated through molecular dynamics simulation, fully considering the dynamic properties of proteins to make up for the shortcomings of traditional static docking.

[0083] (3) Accurate binding free energy evaluation: The binding free energy calculation method is used to conduct a detailed evaluation of the screened candidate compounds to improve the accuracy of screening and the success rate of experimental verification.

[0084] (4) Cluster analysis and structural annotation: Cluster analysis and unfavorable structural annotation are performed on the screened compounds to facilitate further screening and optimization and narrow the scope of experimental verification.

[0085] If the above method is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0086] In other embodiments, an electronic device is also provided, comprising one or more processors, a memory, and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the ultra-high throughput drug small molecule virtual screening method as described above.

[0087] The method of the present invention can effectively improve the efficiency and success rate of large-scale small molecule screening through this multi-level optimization screening process, and provide more efficient technical support for the discovery of lead compounds in new drug development.

[0088] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0089] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0090] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A method for ultra-high throughput virtual screening of small drug molecules, characterized in that: The following steps are involved: Obtaining a target protein structure and a small molecule compound library to be screened, and preliminarily screening the small molecule compound library based on the target protein structure to obtain a first small molecule compound set; Using at least two or more screening tools to screen the first small molecule compound set to obtain a second small molecule compound set; Performing molecular dynamics simulation on the target protein structure, extracting multiple representative conformations from the molecular dynamics simulation trajectory, and screening the second small molecule compound set to obtain a third small molecule compound set based on the representative conformations; performing cluster analysis on the third small molecule compound set, and selecting representative small molecule compounds in each cluster as candidate compounds; The binding free energy of each candidate compound with the target protein structure is calculated, and compounds with lower binding free energy are selected as the final screening results.

2. The ultra-high throughput virtual screening method for small drug molecules according to claim 1, characterized in that: The target protein structure is pre-processed, and the pre-processing includes removing water molecules / non-essential ligands and adding essential hydrogen atoms.

3. The ultra-high throughput virtual screening method for small drug molecules according to claim 1, characterized in that: The small molecule compound library is preprocessed, and the preprocessing includes adding necessary hydrogen atoms and generating a compound library file format.

4. The ultra-high throughput virtual screening method for small drug molecules according to claim 1, characterized in that: The virtual screening toolkit in the VFVS software was used to screen and obtain the first small molecule compound set, the second small molecule compound set and the third small molecule compound set.

5. The ultra-high throughput virtual screening method for small drug molecules according to claim 1, characterized in that: The extraction of multiple representative conformations is specifically as follows: Cluster analysis is performed on the amino acids around the binding site of the target protein structure, and the molecular dynamics simulation trajectory data are divided into several categories. The most representative structure is selected from each category as the representative conformation.

6. The ultra-high throughput virtual screening method for small drug molecules according to claim 1, characterized in that: The cluster analysis of the third small molecule compound set is specifically performed by using a small molecule clustering algorithm to perform cluster analysis on the third small molecule compound set according to chemical structure characteristics and molecular properties.

7. The ultra-high throughput virtual screening method for small drug molecules according to claim 1, characterized in that: The method further includes: The final screening results were experimentally verified.

8. The ultra-high throughput virtual screening method for small drug molecules according to claim 7, characterized in that: Determine whether each candidate compound has an unfavorable structural annotation. Based on the actual experimental situation, give priority to excluding compounds with unfavorable structural annotations and then conduct experimental verification.

9. A computer-readable storage medium, characterized in that: It comprises one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs comprise instructions for executing the ultra-high throughput virtual screening method for small drug molecules as described in any one of claims 1-8.

10. An electronic device, characterized in that: It comprises one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the ultra-high throughput virtual screening method for small drug molecules as described in any one of claims 1-8.

Images