CN114141312A - VR-based virtual drug effect group screening system and method - Google Patents

Abstract

The invention discloses a virtual drug effect group screening system and method based on VR, wherein the system comprises a file loading module, a molecule visualization module, a molecule model rendering module, a virtual drug effect group screening module, an interaction analysis module and an operation guide module; in a VR environment, protein molecule data are loaded and analyzed, after sequentially passing through a visualization module and a model rendering module under the guidance of an operation guide module, a pharmacophore model is extracted from a pharmacophore virtual screening module and screened, a ligand molecule list with potential biological activity is finally obtained, and an optimization result is displayed in an interaction analysis module. The method can effectively improve the reliability and scientific research efficiency of virtual screening of the pharmacophore, and enables new drug research and development; on the other hand, the interactive guidance of the invention can enable beginners to quickly start to understand relevant research and development processes and provide reliable and easy-to-use tools for culturing drug research and development talents.

Description

VR-based virtual drug effect group screening system and method

Technical Field

The invention relates to a Virtual screening method based on a pharmacophore in the field of computer-aided drug design, in particular to a Virtual screening system based on Virtual Reality (VR) technology.

Background

The research and development of new drugs are important means for preventing and treating diseases, and are closely related to national health. However, the development process tends to be costly, long-lived and accompanied by high risks. According to statistics, the time from research and development to market of a new medicine is 10 to 15 years, the research and development cost can reach $ 26 hundred million, and the success rate is less than 10 percent. Computer-aided drug design (CADD) has become an important method for reducing the development cost of new drugs by utilizing the analytical calculation, quick access and three-dimensional simulation modeling capabilities of modern computers. In CADD, virtual screening based on pharmacophores has been increasingly used successfully in recent years, and is receiving wide attention from the industry and researchers. The method deduces and induces the characteristic of a pharmacophore with commonality by analyzing the pharmacodynamic characteristics of one or more active small molecule compounds or the structural characteristics of receptor protein molecules, and then screens out small molecules with potential biological activity from a database by a pharmacophore matching algorithm. Research shows that compared with the traditional CADD screening method, the method has the advantages of rapidness and accuracy.

The key of the virtual screening method based on the pharmacophore is to recognize the spatial characteristics of the pharmacophore, the active sites of molecules and the binding postures of target molecules and ligands. During the development process, the positions of the structures in the three-dimensional space are updated continuously and iteratively, and the accuracy is strictly required. Even minor positional deviations can lead to problems such as missed binding sites or false positives for ligands. However, the existing virtual screening platforms based on pharmacophores, such as LigandScout and PharmMapper, display the three-dimensional structures of the target molecules, ligands and pharmacophores models and the processes of virtual screening through two-dimensional screens. The display mode draws a three-dimensional structure by monocular perspective projection, and is limited by shielding and insufficient depth clues, so that the accurate perception of a user on depth information and relative distance is influenced. On the other hand, the existing virtual screening platform uses traditional input devices such as a keyboard and a mouse to edit and move the three-dimensional molecular model, which also brings about the problems of insufficient accuracy and complicated operation. As the number of atoms of target protein compounds in virtual pharmacophore-based screening is increasing and the structural complexity of molecules is gradually increasing, the influence of the limitations on the research and development efficiency and reliability of new drugs is also amplified.

Virtual Reality (VR) based technologies provide a class of solutions for accurate three-dimensional visualization. VRs can generate immersive, simulated environments for users. The head-mounted display carries out three-dimensional imaging through binocular vision and combines motion parallax to manufacture the sense of depth in space. Research shows that the use of VR can effectively improve the positioning and depth perception of users. Meanwhile, the VR handle brings a brand new interaction means for users. Work such as ChemPreview, Nanome, combining VR with molecular visualization has all yielded satisfactory user feedback. However, these works are limited to performing simple operations such as molecular display and design independently. In virtual screening based on the pharmacophore, no work and VR technology are combined at home and abroad, and a perfect solution can be provided for complex workflow of pharmacophore model and molecular structure display and matching scoring in frequent user interaction. Therefore, the virtual screening system based on the pharmacophore is realized by combining the VR technology, the related vacancy is made up, and the virtual screening system has important research and practical significance.

Disclosure of Invention

The invention aims to provide a virtual drug effect group screening system and method based on VR (virtual reality), aiming at the defects of the prior art, the whole process of virtual drug effect group screening interactive calculation is displayed in an immersive manner by utilizing VR technology, and the problems of low visualization accuracy of a three-dimensional structure of a molecule and inconvenience in man-machine interactive operation in the virtual drug effect group screening process are effectively solved.

In order to achieve the purpose, the technical scheme of the invention is as follows:

a virtual drug effect group screening system based on VR is characterized in that the system comprises a file loading module, a molecule visualization module, a molecule model rendering module, a virtual drug effect group screening module, an interaction analysis module and an operation guide module; in a VR environment, after protein molecule data are loaded and analyzed and sequentially pass through a visualization module and a model rendering module under the guidance of an operation guide module, a pharmacophore model is extracted from a pharmacophore virtual screening module and screened, and finally a ligand molecule list with potential biological activity is obtained and displayed and the result is optimized in an interaction analysis module; wherein:

the file loading module is used for loading local or remote protein molecule data and analyzing the local or remote protein molecule data into a data structure which can be displayed in a virtual reality space;

the molecule visualization module is used for presenting various three-dimensional protein molecule structure characteristics of the loaded molecules in a virtual reality environment;

the molecular model rendering module is used for adjusting the dyeing information and transparency of various characteristics of the loaded molecular model;

the pharmacophore virtual screening module is used for extracting a pharmacophore model according to the loaded molecules and screening ligand molecules in an optimization mode or a scoring mode according to the pharmacophore model; in the optimization mode, a built-in scoring function is applied to calculate the optimal combination posture and the score under the posture one by one for the small molecule database to be screened provided by the user; the scoring mode is used for scoring a single ligand molecule by applying a built-in scoring function according to a spatial position obtained by a user moving a ligand, and simultaneously calculating an optimal binding posture and a score under the posture;

the interactive analysis module is used for displaying the virtual screening result of the optimized mode and providing a pharmacophore re-editing function;

the operation guidance module comprises a prompt box, a highlight information and a UI component of a mutual exclusion button and is used for prompting the execution sequence of the operation panels of the modules.

The file loading module analyzes protein molecules in the PDB format into a data structure of a K-D tree to store information of atoms and bonds.

The molecular visualization module comprises a method for drawing a primary structure, a secondary structure and surface structure characteristics of a molecule in a VR environment; the primary structure consists of a ball stick model and a main chain structure; the secondary structure applies a Catmull-Rom interpolation algorithm to a ribbon model as a drawing method to form spiral, folding and irregular coiled structures; the surface structure takes GPU parallel computing combined with the existing 'improved VCMC algorithm' as a drawing method to form a Solvent exposed structure.

The optimization mode of the pharmacophore virtual screening module calculates the matching scores of all molecules in the small molecule database at one time through triangular Hash operation in a VR environment according to the spatial coordinate position of the ligand molecules, and then displays the results on a three-dimensional user panel and a VR handle in real time; in the scoring mode of the pharmacophore virtual screening module, a user grabs and moves ligand molecules through a VR handle, meanwhile, the matching score of the ligand and the pharmacophore is calculated in real time through a scoring function built in the pharmacophore virtual screening module, and then the result is displayed on a three-dimensional user panel and the VR handle in real time.

The optimal combination posture deduced by the two modes of the pharmacophore virtual screening module is mapped to the molecule visualization module in real time so as to highlight the combination posture displayed in a three-dimensional space.

And the interaction analysis module displays the ligand molecules of the top ten in descending order of the real-time updated three-dimensional histogram in the optimization mode, and after a certain ligand in the histogram is selected through a VR handle, the molecule visualization module correspondingly displays the binding posture of the ligand.

The interaction analysis module allows a user to edit the pharmacophore structural features through the VR handle, thereby iteratively optimizing the virtual screening results.

A method for implementing virtual screening of pharmacophores by using the system is characterized by comprising the following steps:

step 1: analyzing the protein molecule data in the local or remote PDB format through a file loading module according to the operation guidance;

step 2: drawing one or more characteristics of the primary structure, the secondary structure and the surface structure of the read protein molecules in a VR environment through a molecule visualization module according to the operation guide;

and step 3: according to the operation guide, the dyeing information and the transparency of the drawn protein molecular structure characteristics are adjusted through a molecular model rendering module;

and 4, step 4: reading ligand small molecules or a small molecule database through a pharmacophore virtual screening module according to operation guidance, extracting a pharmacophore model according to a protein structure or introducing the existing pharmacophore model, displaying a scoring result on each read molecule by utilizing an optimization mode or a scoring mode aiming at the protein structure and the pharmacophore model, and simultaneously displaying the optimal combination posture deduced by the pharmacophore virtual screening module in a molecule visualization module in real time;

and 5: and editing and optimizing the pharmacophore according to the scoring result through the interactive analysis module, and iteratively executing the step 4 until the score is obtained and the screening result expected by the user is achieved.

Compared with the prior art, the invention has the following beneficial effects: the target molecular structure and the screening process of the pharmacophore and the ligand are displayed through the immersive VR environment, so that the problem that the pharmacophore and the molecular structure model are inaccurate in the virtual screening process of the pharmacophore is solved, and meanwhile, the pain point which is fussy to operate and inaccurate in the traditional research and development platform is solved through the interaction mode of operating the three-dimensional panel and the prompt box which are provided by the guide module and utilizing the handle in the VR environment. Therefore, the reliability and scientific research efficiency of virtual screening of the pharmacophore can be effectively improved, and new drug research and development are enabled; on the other hand, the interactive guidance of the invention can enable beginners to quickly start to understand relevant research and development processes and provide reliable and easy-to-use tools for culturing drug research and development talents. In conclusion, the invention can be applied in multiple fields of classroom teaching, academic research, enterprise research and the like, considerable social and economic benefits are created, and the win-win of multi-direction production, study and research is realized.

Drawings

FIG. 1 is a system architecture and method flow of the present invention;

FIG. 2 is a diagram of an embodiment of an optimization mode of the invention under the implementation of a pharmacophore virtual screening module;

FIG. 3 is a diagram of an embodiment of a scoring mode implemented by a pharmacophore virtual screening module according to the invention;

FIG. 4 is a flowchart of an iterative update process for implementing pharmacophore optimization according to the present invention.

Detailed Description

The invention is further described below with reference to the figures and examples, but the invention is not limited in any way.

The invention discloses a VR-based pharmacophore virtual screening system, which comprises the following modules: the system comprises a file loading module, a molecule visualization module, a molecule model rendering module, a pharmacophore virtual screening module, an interaction analysis module and an operation guidance module. The system starts with a file loading module loading target protein molecules, a molecule visualization module draws protein structural characteristics, a molecule model rendering module adjusts model display effects, a pharmacophore virtual screening module screens out a batch of small molecules with potential biological activity and determines the combination posture of each molecule and the target protein, and finally an interactive analysis module edits the pharmacophore model to perform iterative virtual screening to optimize screening results. The operation sequence of each module is guided by the UI of the operation guide module, and the function of each module is scheduled by the operation panel corresponding to each module.

The file loading module comprises the following four sub-modules: the local loading submodule is used for loading the protein file stored in the hard disk of the personal computer of the user; the remote Web request loading submodule is used for remotely requesting a professional protein database and loading a protein file with a specified ID; the working directory loading submodule is used for loading the protein file from the centralized storage directory specified by the program; and the record loading submodule is used for loading the corresponding protein file according to the history of protein loading.

The molecular visualization module in turn comprises the following four sub-modules: the primary structure display submodule is used for displaying a club model and a main chain structure of the target protein molecule; the secondary structure display submodule is used for displaying the characteristic structures of the spiral, folded and irregular curled bands of the target protein molecules, and the band models of the characteristics are obtained by interpolation calculation of a Catmull-Rom spline curve; the surface structure display submodule is used for displaying the molecular surface structure of the target protein and is obtained by combining an improved VCMC algorithm with GPU (graphics processing unit) for parallel calculation; and the scene display sub-module is used for adjusting the environmental characteristics in VR scenes such as illumination, shadow, ground and the like. And the primary structure, the secondary structure and the surface structure are rendered by a Mesh component and a vertex shader of the Unity engine in a unified manner.

The molecular model rendering module is used for customizing the display effects of the ball stick model, the spiral model, the folding model and the ribbon model in the model visualization module, and comprises color and transparency.

The pharmacophore virtual screening module comprises a pharmacophore extraction submodule and a virtual screening and scoring submodule. The pharmacophore extraction modeling module comprises the following steps:

step 1: checking whether a pharmacophore file corresponding to the loaded protein exists in the working directory or not, and if not, performing the step 2; if yes, jumping to step 5;

step 2: checking whether the protein binding pocket file completely exists in the working directory, and if not, performing the step 3; if yes, jumping to the step 4;

and step 3: inputting the loaded protein file into a CAVITY program to generate one or more corresponding protein binding pockets, persisting in a CAVITY format, and storing the pockets in a working directory in a centralized manner;

and 4, step 4: inputting the loaded protein file and the corresponding protein combined Pocket file into an improved Pocket algorithm, generating one or more corresponding pharmacophore files, and then persisting in a hypoedit format and storing in a working directory in a centralized manner;

and 5: corresponding pharmacophore files are loaded and the spatial arrangement of the pharmacophore files is visualized on the basis of VR visualization of proteins.

The virtual screening scoring submodule includes two types of screening modes, optimization and scoring. The optimization mode is used for rapidly screening a small molecule database to obtain the optimal binding attitude and score of each molecule, and comprises the following steps:

step 1: checking whether the operation of the pharmacophore extraction submodule is finished or not, and if the operation is finished, loading a pharmacophore file; if not, carrying out a pharmacophore extraction step;

step 2: and reading a micromolecule database of the working catalog, sequentially inputting the micromolecules into a PharmMapper program to calculate the pharmacophore matching score and the optimal combination posture, and displaying the score on a three-dimensional user interface and a VR (virtual reality) interactive handle in real time.

The scoring model is used for exploring the binding posture of a specific ligand small molecule and a target protein, and comprises the following steps:

step 1: checking whether the operation of the pharmacophore extraction submodule is finished or not, and if the operation is finished, loading a pharmacophore file; if not, carrying out a pharmacophore extraction step;

step 2: and loading a specific small molecule file from a working catalog, and visualizing the small molecule file into a ball stick model in a VR space. The model can be selected by the handle ray to move, rotate and the like to change the conformation;

and step 3: if the user selects the pharmacophore adaptation function, the system displays the optimal score calculated by PharmMapper, and at the same time, the system recalculates the pharmacophore matching score according to the relative position of the pharmacophore matching score and the target protein every time the spatial position of the small molecule is changed, and the pharmacophore matching score is used as a reference for the quality of the current binding posture;

and 4, step 4: if the user selects the optimal position function, the optimal binding posture of the target protein and the small molecule calculated by PharmMapper is visualized in a highlight form in a VR environment.

The visual interactive analysis module comprises a screening result visual submodule and a pharmacophore model optimization submodule. The screening result visualization module is used for displaying the matching result of the optimal scoring mode, specifically displaying the scores of all small molecules in a descending order in a three-dimensional histogram mode, and performing iterative updating in the process of traversing the database, wherein specific small molecules in the histogram can be selected through a handle and used for observing the combination posture of the specific small molecules. The pharmacophore model optimization module is used for simplifying and adjusting the spatial characteristic structure of the pharmacophore model so as to obtain a better virtual screening result, specifically, the characteristics of the pharmacophore model can be edited by deletion, movement and the like through a handle, and the next round of virtual screening operation is seamlessly input.

The functional operation of the modules is supported by the auxiliary guide module in the whole process. The module comprises three-dimensional panels which respectively correspond to the functions of the modules and are used as user interaction interfaces; and the operation guide module prompts the correct execution sequence of the complicated operations in the virtual selection of the pharmacophore in a UI form with a prompt box, a highlight and a forbidden button.

The invention is supported by VR wearable devices including but not limited to Oculus Rift, HTC VIVE, Pico Neo3, etc. to support SteamVR. This type of device consists primarily of a head-mounted display and an interactive handle. Wherein the head-mounted display is configured to provide accurate and immersive perception of the spatial structure of the molecules in the VR environment by virtue of binocular perspective superimposition and motion parallax; the handle is used for interacting operations such as clicking, dragging and the like with the three-dimensional user interface and the molecular model through a ray function or contact collision in cooperation with the visual angle of the head-mounted display in the virtual drug effect group screening process.

Example 1

The 1db4 protein was used as a target for virtual screening using an optimization model, and the overall flow chart is shown in FIG. 1.

The method comprises the following specific steps:

(1) preparing VR Equipment

The system is started by wearing an HTC Vive VR head-mounted display and an interactive handle, and is wirelessly streamed to VR equipment on a PC through a Steam VR.

(2) Familiarity with VR interaction and System functionality

Entering the immersive VR environment, the operation guidance module provides the system instructions and the use of the keys of the interactive handle in a welcome scene. And then clicking the Home button to enter a working scene.

(3) Parsing target protein files using a file loading module

According to the guidance prompt, a Load File button of a main control panel (Home panel) in a working scene is opened to call a File loading panel, and the File loading panel enters a File loading module. Then click the Remote Files button and enter 1db4 using a soft keyboard to network small molecule structures. Finally, selecting 1db4 in the worklist and clicking the Load loading molecule, and displaying the space club model in the VR environment after the completion.

(4) Mapping protein structures using molecular visualization modules

And opening a Visualization button of a main control panel in a working scene to call a Visualization panel according to the guidance prompt, and entering a molecular Visualization module. Then selecting a Chain check box in a Primary structure (Primary) entry of the panel to display a main Chain structure; selecting Helix, Sheet and Coil in Secondary structure (Secondary) entries to show ribbon models of Helix, fold and random Coil, respectively; selecting a solvent exposed Surface option from a Surface structure (Surface) item to generate a Surface model, and setting the transparency of the Surface model to be 0.5 by using a slider so as to simultaneously display a primary structure, a secondary structure and a Surface structure; finally, the display of the ground in the Scene is closed in the Scene (Scene) entry to obtain a better view.

(5) Adjusting protein rendering effects using a model rendering module

And opening a Render button of the main control panel in the working scene to call up a Render panel according to the guidance prompt, and entering a model Render module. The subsequent adjustment of 4 sliders in the panel ranging from 0 to 255 changes the RGBA channel of the stick model, where the selection does not change the sliders, keeping the default rendering effect.

(6) Drug effect cluster model extraction by using drug effect cluster virtual screening module

And opening a Screening button of the main control panel in the working scene to call up the Screening panel according to the guidance prompt. The binding pocket for the 1db4 protein was first generated in this panel using the Get Cavity button, with the results stored in the working catalog (typically C:. Users \ username \ AppData \ LocalLow \ ECUST \ molecular visualization _ ECUST \ Data). The binding pocket file is then read using the Get pharmacy button to generate one or more Pharmacophore models corresponding to 1db4, the structure also being stored in the working catalog. After completion, the system will search the working catalog and display all the corresponding pharmacophore models, select one model and load the calculation basis of the virtual screening.

(7) Performing virtual screening using an optimization model under a pharmacophore virtual screening module

Selecting a Mode OPT to enter an optimization Mode sub-panel, and after selecting the Start, starting automatic screening by the system by traversing the small molecule database of the working directory. As shown in the left part of fig. 2, the screening result will appear in the VR space in the form of a histogram, and will be updated continuously as the screening progresses. All results are displayed in the histogram in descending order of the best ten small molecules to calculate the score, the real time score will be displayed on the panel and handle; as shown in the right part of fig. 2, the binding attitude will also be displayed in real time in the protein spatial structure. When the database traversal is completed, ten small molecules with the most possible bioactivity can be obtained according to the histogram.

Example 2

Virtual screening was performed using the scoring pattern using the 1db4 protein as a target.

The method comprises the following specific steps:

the steps (1) to (6) are the same as in example 1;

(7) performing virtual screening using scoring patterns

Select Mode Score to enter the Score Mode sub-panel. The small molecule files to be explored are loaded from the file list of the working directory, and the spatial structure represented by the club model appears in the VR environment. As shown in the left part of FIG. 3, the PharmFit button is selected to display the best score of the small molecule as calculated by PharmMapper on the panel and handle, while displaying the current score as calculated from the relative spatial position of the small molecule with respect to the protein molecule. As shown in the right part of FIG. 3, the Best Position button is selected again, and the optimal binding posture of the small molecule and the protein calculated by PharmMapper is displayed in a green highlight form. The small molecules are then grasped, moved, rotated using the buttons of the handle, and the scores for each binding site are observed and compared to the optimal site calculated by the scoring schema.

Example 3

The procedure for optimizing the pharmacophore model extracted from protein 1db4 after virtual screening is shown in fig. 4.

The method comprises the following specific steps:

the steps (1) to (7) are the same as in example 1;

(8) iterative optimization pharmacophore model

The histogram top scoring first through tenth entries were selected to show their predicted best binding poses to proteins in VR space. Any one feature in the pharmacophore model is selected, and the handle button is used for deleting. The Start button is again used in the filter panel to Start a new round of virtual filtering. And after the completion, observing whether the score result of each molecule is improved or not to check the effect of the optimal editing of the pharmacophore. This process is iterated until the optimized mode scoring function results in a sufficiently high score for the user to accept.

Claims (8)

1. A virtual drug effect group screening system based on VR is characterized in that the system comprises a file loading module, a molecule visualization module, a molecule model rendering module, a virtual drug effect group screening module, an interaction analysis module and an operation guide module; in a VR environment, after protein molecule data are loaded and analyzed and sequentially pass through a visualization module and a model rendering module under the guidance of an operation guide module, a pharmacophore model is extracted from a pharmacophore virtual screening module and screened, a ligand molecule list with potential biological activity is finally obtained, and an optimization result is displayed in an interaction analysis module; wherein:

the file loading module is used for loading local or remote protein molecule data and analyzing the local or remote protein molecule data into a data structure which can be displayed in a virtual reality space;

the molecule visualization module is used for presenting various three-dimensional protein molecule structure characteristics of the loaded molecules in a virtual reality environment;

the molecular model rendering module is used for adjusting the dyeing information and transparency of various characteristics of the loaded molecular model;

the pharmacophore virtual screening module is used for extracting a pharmacophore model according to the loaded molecules and screening ligand molecules in an optimization mode or a scoring mode according to the pharmacophore model; in the optimization mode, a built-in scoring function is applied to calculate the optimal combination posture and the score under the posture one by one for the small molecule database to be screened provided by the user; the scoring mode is used for scoring a single ligand molecule by applying a built-in scoring function according to a spatial position obtained by a user moving a ligand, and simultaneously calculating an optimal binding posture and a score under the posture;

the interactive analysis module is used for displaying the virtual screening result of the optimized mode and providing a pharmacophore re-editing function;

the operation guidance module comprises a prompt box, a highlight information and a UI component of a mutual exclusion button and is used for prompting the execution sequence of the operation panels of the modules.

2. The VR-based pharmacophore virtual screening system of claim 1, wherein the file loading module parses PDB-formatted protein molecules into a K-D tree data structure to store atom and bond information.

3. The VR-based pharmacophore virtual screening system of claim 1, wherein the molecular visualization module comprises a method for mapping molecular primary structures, secondary structures, and surface structure features in a VR environment; the primary structure consists of a ball stick model and a main chain structure; the secondary structure applies a Catmull-Rom interpolation algorithm to a ribbon model as a drawing method to form spiral, folding and irregular coiled structures; the surface structure takes GPU parallel computing combined with the existing 'improved VCMC algorithm' as a drawing method to form a Solvent exposed structure.

4. The VR-based pharmacophore virtual screening system of claim 1, wherein the optimization mode of the pharmacophore virtual screening module calculates matching scores of all molecules in the small molecule database at one time by triangle hash operations according to ligand molecule space coordinate positions in VR environment, and displays the results in real time on a three-dimensional user panel and VR handle; in the scoring mode of the pharmacophore virtual screening module, a user grabs and moves ligand molecules through a VR handle, meanwhile, the matching score of the ligand and the pharmacophore is calculated in real time through a scoring function built in the pharmacophore virtual screening module, and then the result is displayed on a three-dimensional user panel and the VR handle in real time.

5. The VR-based pharmacophore virtual screening system of claim 1, wherein the inferred optimal binding poses of both modes of the pharmacophore virtual screening module are mapped in real-time to the molecular visualization module to highlight binding poses in three-dimensional space.

6. The VR-based pharmacophore virtual screening system of claim 1, wherein the interaction analysis module displays top ten scored ligand molecules in descending order of a real-time updated three-dimensional histogram in the optimization mode, and the molecule visualization module correspondingly displays binding poses of a ligand upon selection of that ligand by the VR handle.

7. The VR-based pharmacophore virtual screening system of claim 1, wherein the interaction analysis module allows a user to edit pharmacophore structural features through a VR handle to iteratively optimize virtual screening results.

8. A method for performing virtual pharmacophore screening using the system of claim 1, comprising the steps of:

step 1: analyzing the protein molecule data in a local or remote PDB format through a file loading module;

step 2: drawing one or more characteristics of the primary structure, the secondary structure and the surface structure of the read protein molecule in a VR environment through a molecule visualization module;

and step 3: adjusting the dyeing information and transparency of the drawn protein molecular structure characteristics through a molecular model rendering module;

and 4, step 4: reading ligand small molecules or a small molecule database through a pharmacophore virtual screening module, extracting a pharmacophore model according to a protein structure or introducing the pharmacophore model into the existing pharmacophore model, displaying a scoring result on each read molecule by utilizing an optimization mode or a scoring mode aiming at the protein structure and the pharmacophore model, and simultaneously displaying the optimal combination posture deduced by the pharmacophore virtual screening module in a molecule visualization module in real time;

and 5: and editing and optimizing the pharmacophore according to the scoring result through the interactive analysis module, and iteratively executing the step 4 until the score is obtained and the screening result expected by the user is achieved.

Images