CN114520022A - GPU parallel computing molecular similarity method, device, system and medium - Google Patents

Abstract

The invention discloses a method, a device, a system and a medium for GPU parallel computing molecular similarity, wherein the method comprises the following steps: inputting a query compound; calculating a centroid of the query compound; translating and/or rotating the query compound based on the centroid; performing parallel calculation through a GPU, and comparing query compounds before and after conversion as initial search points with molecular compounds in a molecular library; and filtering according to the comparison calculation result, and outputting the molecular compound similar to the query compound.

Description

A kind of GPU parallel computing molecular similarity method, device, system and medium

technical field

The invention relates to the technical field of computer-aided drug research and development, in particular to a method, device, system and medium for GPU parallel calculation of molecular similarity.

Background technique

Drug R&D has the characteristics of large investment, high risk and long cycle. Usually, a drug R&D cycle is more than 10 years, and the R&D investment is hundreds of millions of dollars, and it shows an upward trend year by year. Drug screening is a key link in drug discovery, and high-throughput drug virtual screening can greatly reduce screening time and cost, which is of great significance for accelerating drug development. The matching and quantitative comparison of the three-dimensional shapes of two molecules and the pharmacophore (a group of atoms with specific properties in a molecule) is a major method in drug design. Molecular volume is related to shape, and shape determines the physicochemical properties of the molecule, and this property determines the biological activity of the molecule.

Molecular shape comparison is a common technique: used to identify spatial features between two or more molecules, a metric commonly used in ligand-based compound discovery efforts. Among them, in the 3D shape similarity search, the best algorithm is "Gaussian superposition based on atomic center". In the software package of OpenEye Scientific Software, ROCS applies the "Gaussian superposition based on atomic center" algorithm for similarity calculation. But even with Gaussian optimization techniques, ROCS takes a long time to complete when scanning large compound databases.

Therefore, the development of a rapid and successful molecular shape comparison method has become an urgent problem to be solved.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a GPU parallel calculation method, device, system and medium for molecular similarity, which is used to improve the calculation speed of 3D Gaussian superposition similarity and shorten the time required for large-scale compound similarity search. Processing improves the success rate and efficiency of searches.

The technical scheme adopted by the present invention to solve the above-mentioned technical problems is as follows:

A GPU parallel computing molecular similarity method, comprising the following steps:

Enter the query compound;

Calculate the centroid of the query compound;

translate and/or rotate the query compound based on the centroid;

Parallel computing is performed by GPU, and the query compounds before and after transformation are used as starting search points to compare with the molecular compounds in the molecular library;

Filter based on the calculated results of the alignment to output molecular compounds similar to the query compound.

In one embodiment, the input query compound includes:

Input the three-dimensional structure information of the query compound, where the three-dimensional structure information includes the type of each atom in the query compound and its coordinate value;

The volume of the query compound is represented by a Gaussian function.

In one embodiment, calculating the centroid of the query compound includes the steps of:

Calculate the center of mass of the query compound based on the three-dimensional structure information of the query compound;

And perform SVD decomposition to calculate 3D rotation matrix.

In one embodiment, the translation and/or rotation transformation of the query compound based on the centroid comprises the following steps:

perform translation and/or rotation transformations on the query compound;

The translation and/or rotation transformation of the query compound is recorded through a set of quadruplets of data.

In one embodiment, performing parallel computing on GPU, and comparing the query compound before and after the transformation with the molecular compound in the molecular library as the initial search point includes the following steps:

Load the Gaussian function of the query compound and the quadruple recording the compound displacement into the GPU memory;

Multiple computing points of the GPU run the optimization algorithm to calculate the 3D Gaussian superposition between the query compound and the molecular compounds in the molecular library.

In one embodiment, the optimization algorithm is BFGS and gradient descent.

Another embodiment of the present invention also provides a GPU parallel computing molecular similarity device, the device includes:

Molecular input module for inputting query compounds;

The centroid calculation module is used to calculate the centroid of the query compound;

Molecular transformation module for translation and/or rotation transformation of query compounds based on centroids;

Molecular alignment module is used for parallel computing through GPU, and the query compound before and after transformation is used as the starting search point to compare with the molecular compounds in the molecular library;

The result output module is used to filter according to the calculation result of the comparison, and output the molecular compounds similar to the query compound.

Another embodiment of the present invention also provides a GPU parallel computing molecular similarity system, the system includes at least one processor; and,

a memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described GPU parallel computing molecular similarity method.

Another embodiment of the present invention also provides a non-volatile computer-readable storage medium storing computer-executable instructions, the computer-executable instructions being stored by one or more When executed by the processor, the one or more processors can be made to execute the above-mentioned method for calculating molecular similarity in parallel on the GPU.

Beneficial effects: a method, device, system and medium for parallel computing of molecular similarity based on GPU of the present invention utilizes the advantages of parallel computing of GPU to realize Gaussian superposition calculation of 3D similarity on GPU. The improvement of orders of magnitude shortens the query time of 3D similarity in virtual screening. In the process of calculating the Gaussian superposition, we calculate the overlapping function and its gradient transformation coordinates, optimize the molecular overlap, and perform parallel calculations. At the same time, in order to find the global optimal superposition faster, we perform translation and rotation transformations on the compounds as a starting search. point. The calculation on the GPU can better find the global optimum; the final calculation result is generated according to the similarity threshold that the user is interested in, and the calculation speed is significantly improved compared with the prior art.

Other features and advantages of the invention will be set forth in the description which follows, and in part will become apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description and drawings.

Description of drawings

The present invention will be described in detail below with reference to the accompanying drawings, so as to make the above advantages of the present invention more clear.

FIG. 1 is a flowchart of a GPU parallel computing method for molecular similarity according to the present invention.

2 is a schematic diagram of functional modules of an embodiment of a GPU parallel computing device for molecular similarity of the present invention;

FIG. 3 is a schematic diagram of the hardware structure of an embodiment of a GPU parallel computing device for molecular similarity according to the present invention.

Detailed ways

In order to make the objectives, technical solutions and effects of the present invention clearer and clearer, the present invention will be described in further detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

As shown in Figure 1-2, a GPU parallel computing molecular similarity method includes the following steps:

S100. Input a query compound;

Extract from the database or customize the construction, enter the compound to be queried, extract feature points based on the queried compound, and generate a search formula.

S200. Calculate the centroid of the query compound;

When calculating the 3D Gaussian superposition between the query compound and the compounds in the library, in order to ensure the accuracy of the calculation and improve the calculation speed, the query compound will be adjusted. For the convenience of calculation and comparison, the centroid of the query compound is selected as the calibration point .

S300, performing translation and/or rotation transformation on the query compound based on the centroid;

The query compound is shifted based on the centroid, in order to find the global optimal superposition faster when calculating the Gaussian superposition.

S400, performing parallel computing through GPU, and comparing the query compound before and after transformation with the molecular compound in the molecular library as a starting search point;

The core of the GPU is good at completing tasks with simple control logic, focusing on computing and focusing on parallelism. Compared with CPU, GPU has a lot of computing power. When the query compound is compared with the molecular compounds in the molecular library, the advantage of GPU parallel computing is used to compare the query compounds before and after transformation with the molecular compounds in the molecular library as the starting search point, which greatly improves the performance of the query compound. The efficiency of calculating the 3D Gaussian superposition between the query compound and the compounds in the library shortens the retrieval time; and the parallel computing of the GPU and the translation and/or rotation transformation of the query compound improve the success rate of calculating the 3D Gaussian superposition .

S500 , filtering according to the comparison calculation result, and outputting molecular compounds similar to the query compound. According to the comparison result between the query compound and the molecular compounds in the molecular library, the Tanimoto3D similarity is calculated according to the 3D Gaussian superposition, and the molecular compounds similar to the query compound are output from the molecular library.

In one embodiment, the input query compound includes:

Input the three-dimensional structure information of the query compound, the three-dimensional structure information includes the type of each atom in the query compound and its coordinate value; obtain the corresponding van der Waals radius according to the type of each atom in the query compound, and convert the three-dimensional structure information into a set of representative Query the Gaussian sphere of each atom in the compound, the radius of each Gaussian sphere is the same as the van der Waals radius of the corresponding atom, and the position of each Gaussian sphere is the same as the coordinates of the corresponding atom;

The volume of the query compound is represented by a Gaussian function.

In one embodiment, calculating the centroid of the query compound includes the steps of:

Calculate the center of mass of the query compound based on the three-dimensional structure information of the query compound;

And perform SVD decomposition to calculate 3D rotation matrix.

Compared with the atom center, the center of mass plays an important role in the subsequent transformation of the query compound. Based on the three-dimensional structure information of the molecule, the center of mass of the query compound is calculated. In order to reduce the interference of other data and the volume of data calculation, after finding the center of mass of the query compound, the 3D rotation matrix is calculated by SVD decomposition.

In one embodiment, the translation and/or rotation transformation of the query compound based on the centroid comprises the following steps:

perform translation and/or rotation transformations on the query compound;

The translation and/or rotation transformation of the query compound is recorded through a set of quadruplets of data.

In three-dimensional space, any coordinate system can be represented by a 4*4 transformation matrix, where the 3*3 matrix in the upper left corner represents the rotation matrix, and the first three in the fourth column represent the x, y, and z coordinates.

The most common representation of a quaternion is: q=s+xi+yj+zk s,x,y,z∈R. For an arbitrary vector, we can use its length and its direction to represent it. Similar to the rotation of the vector, we can also use a quadruple to represent the rotation of the quadruple q=[cosθ, sinθv].

When calculating the 3D Gaussian superposition between the query compound and the compounds in the library, the query compound is often adjusted in order to demand the optimal solution. In this embodiment, the GPU is used for parallel computing, making full use of the high computing power and parallel computing of the GPU. When searching, the query compound is first subjected to displacement transformation, and the rotation vector of the query compound is recorded through a set of quaternions; and the advantage of GPU parallel computing is used to use the query compound before and after transformation as the search starting point and in the molecular library. The molecular compounds are compared to improve the success rate and efficiency of the search.

In one embodiment, performing parallel computing on GPU, and comparing the query compound before and after the transformation with the molecular compound in the molecular library as the initial search point includes the following steps:

Load the Gaussian function of the query compound and the quadruple recording the compound displacement into the GPU memory;

In order to improve the success rate and efficiency of the search, taking advantage of the high computing power and parallel computing of GPU, the query compounds before and after transformation are used as the search starting point.

Multiple computing points of the GPU run the optimization algorithm to calculate the 3D Gaussian superposition between the query compound and the molecular compounds in the molecular library.

The query compound before and after transformation is used as the search starting point, and the 3D Gaussian superposition between the calculation and the compounds in the library is performed at the same time, the overlapping function and its gradient transformation coordinates are calculated, and the molecular overlap is optimized.

The advantages of parallel computing are mainly reflected in two aspects. On the one hand, the query compound can be overlapped with multiple different compounds in the molecular library to improve efficiency. On the other hand, different forms and molecular libraries of the query compound can be used to aggregate the same compound for overlapping comparison, which speeds up the search for the optimal solution and improves the accuracy rate.

In one embodiment, the optimization algorithm is BFGS and gradient descent.

In the process of calculating the Gaussian superposition, we calculate the overlap function and its gradient transformation coordinates, optimize the molecular overlap, and use the BFGS algorithm that supports parallel computing. Newton's method is a method of approximately solving equations in the real and complex fields. The method uses the first few terms of the Taylor series of the function f(x) to find the roots of the equation f(x)=0. The biggest feature of Newton's method is that its convergence speed is very fast.

Gradient descent is the earliest, simplest and most commonly used optimization method. The gradient descent method is simple to implement. When the objective function is a convex function, the solution of the gradient descent method is a global solution. In general, the solution is not guaranteed to be the global optimal solution, and the speed of the gradient descent method is not necessarily the fastest. The optimization idea of the gradient descent method is to use the negative gradient direction of the current position as the search direction, because this direction is the fastest descent direction of the current position, so it is also called the "steepest descent method". The closer the steepest descent method is to the target value, the smaller the step size and the slower the progress.

To sum up, the above technical solution takes advantage of GPU parallel computing to realize Gaussian superposition calculation of 3D similarity on GPU. Compared with CPU-based implementation, the speed is improved by an order of magnitude, which shortens the time required for 3D similarity in virtual screening. query time. In the process of calculating the Gaussian superposition, we calculate the overlap function and its gradient transformation coordinates, optimize the molecular overlap, and use the BFGS algorithm that supports parallel computing. Rotation transformation, as the starting search point for BFGS. This makes it possible to better find the global optimum for computing on the GPU; the final computing result is generated according to the similarity threshold that the user is interested in. The practical results show that the method implemented by the parallel GPU has a huge improvement in calculation speed compared with the OpenEye ROCS CPU implementation.

Another embodiment of the present invention also provides a GPU parallel computing molecular similarity device, the device includes:

Molecular input module for inputting query compounds;

The centroid calculation module is used to calculate the centroid of the query compound;

Molecular transformation module for translation and/or rotation transformation of query compounds based on centroids;

Molecular alignment module is used for parallel computing through GPU, and the query compound before and after transformation is used as the starting search point to compare with the molecular compounds in the molecular library;

The result output module is used to filter according to the calculation result of the comparison, and output the molecular compounds similar to the query compound.

Another embodiment of the present invention provides a system for generating a three-dimensional model. As shown in FIG. 3 , the system 50 includes:

One or more processors 510 and the memory 520 are described in FIG. 3 by taking one processor 510 as an example. The processor 510 and the memory 520 may be connected by a bus or in other ways. In FIG. 3 , the connection by a bus is used as an example.

The processor 510 is used to complete various control logics of the system 50, which can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (AcornRISCMachine) ) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components. Also, processor 510 may be any conventional processor, microprocessor or state machine. The processor 510 may also be implemented as a combination of computing devices, eg, a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in combination with a DSP, and/or any other such configuration.

The memory 520, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs and modules, such as the GPU parallel computing molecular similarity method in the embodiment of the present invention. corresponding program instructions. The processor 510 executes various functional applications and data processing of the system 50 by running the non-volatile software programs, instructions and units stored in the memory 520, that is, implementing the GPU parallel computing molecular similarity method in the above method embodiments. .

The memory 520 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the system 50, and the like. Additionally, memory 520 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 520 may optionally include memory located remotely from processor 510, which may be connected to system 50 through a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

One or more units are stored in the memory 520, and when executed by one or more processors 510, execute the GPU parallel computing molecular similarity method in any of the above method embodiments, for example, execute the method in FIG. 1 described above Steps S100 to S400.

Embodiments of the present invention provide a non-volatile computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by one or more processors, for example, to execute the above-described The method steps S100 to S500 in FIG. 1 .

As examples, the nonvolatile storage medium can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) as external cache memory. By way of illustration, and not limitation, RAM can be configured in formats such as Synchronous RAM (SRAM), Dynamic RAM, (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM) And many forms like direct Rambus (Lambas) RAM (DRRAM). The disclosed memory components or memories of the operating environments described herein are intended to include one or more of these and/or any other suitable types of memory.

Another embodiment of the present invention provides a computer program product comprising a computer program stored on a non-volatile computer-readable storage medium, the computer program comprising program instructions that when executed by a processor , causing the processor to execute the GPU parallel computing molecular similarity method of the above method embodiments. For example, the above-described method steps S100 to S400 in FIG. 1 are performed.

To sum up, the present invention discloses a method, device, system and medium for GPU parallel calculation of molecular similarity. The method reduces the weight of the skeleton model, binds the skeleton model and the skin model, and provides fine-tuning and action of the skeleton. Fine-tuning, skin fine-tuning and other tools can create natural and smooth 3D virtual objects, strictly control the volume and transmission volume of data, greatly reduce the loading waiting time and the budget of the program, and finally transfer the file to a general file, It is suitable for different 3D model application platforms and has strong versatility.

The above-described embodiments are only illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, Alternatively, it can be distributed over multiple network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

From the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and certainly can also be implemented by hardware. Based on this understanding, the above-mentioned technical solutions can be embodied in the form of software products in essence, or the parts that make contributions to related technologies. The computer software products can exist in computer-readable storage media, such as ROM/RAM, alkaline disks , optical disc, etc., including several instructions for causing a computer electronic device (which may be a personal computer, a server, or a network electronic device, etc.) to perform various embodiments or methods of certain parts of the embodiments.

Conditional language such as "could," "could," "may," or "could," among others, is generally intended to convey unless specifically stated otherwise or otherwise understood within the context as used Certain embodiments can include, while other embodiments do not, include particular features, elements, and/or operations. Thus, such conditional language is also generally intended to imply that features, elements, and/or operations are required anyway for one or more implementations or that one or more implementations must include for use with or without input or prompting logic to determine whether such features, elements and/or operations are included or to be performed in any particular implementation.

What has been described herein in this specification and the accompanying drawings includes examples that can provide a method, apparatus, system and medium for GPU parallel computing of molecular similarity. Of course, not every conceivable combination of elements and/or methods has been described for the purpose of describing the various features of the present disclosure, but it will be appreciated that many additional combinations and permutations of the disclosed features are possible. Therefore, it will be apparent that various modifications can be made in the present disclosure without departing from the scope or spirit of the disclosure. In addition, or in the alternative, other embodiments of the present disclosure may be apparent from consideration of this specification and drawings, and from practice of the present disclosure as presented herein. It is intended that the examples presented in this specification and drawings are to be regarded in all respects as illustrative and not restrictive. Although specific terms are employed herein, they are used in a generic and descriptive sense and not for purposes of limitation.

Claims (9)

1. a GPU parallel computing molecular similarity method, is characterized in that, comprises the following steps: Enter the query compound; Calculate the centroid of the query compound; translate and/or rotate the query compound based on the centroid; Parallel computing is performed by GPU, and the query compounds before and after transformation are used as starting search points to compare with the molecular compounds in the molecular library; Filter based on the calculated results of the alignment to output molecular compounds similar to the query compound. 2. according to the described GPU parallel computing molecular similarity method of claim 1, it is characterized in that, described input query compound comprises: Input the three-dimensional structure information of the query compound, where the three-dimensional structure information includes the type of each atom in the query compound and its coordinate value; The volume of the query compound is represented by a Gaussian function. 3. according to the described GPU parallel computing molecular similarity method of claim 2, it is characterized in that, described calculating the centroid of query compound comprises the following steps: Calculate the center of mass of the query compound based on the three-dimensional structure information of the query compound; And SVD decomposition is performed to calculate the 3D rotation matrix. 4. according to the described GPU parallel computing molecular similarity method of claim 3, it is characterized in that, described based on centroid, the query compound is carried out translation and/or rotation transformation and comprises the following steps: perform translation and/or rotation transformations on the query compound; The translation and/or rotation transformation of the query compound is recorded through a set of quadruplets of data. 5. according to the described GPU parallel computing molecular similarity method of claim 5, it is characterised in that the described parallel computing is carried out by GPU, the query compound before and after the transformation is all compared with the molecular compound in the molecular library as an initial search point Include the following steps: Load the Gaussian function of the query compound and the quadruple recording the compound displacement into the GPU memory; Multiple computing points of the GPU run the optimization algorithm to calculate the 3D Gaussian superposition between the query compound and the molecular compounds in the molecular library. 6. The GPU parallel computing molecular similarity method of claim 5, wherein the optimization algorithm is BFGS and gradient descent. 7. a GPU parallel computing molecular similarity device, is characterized in that, described device comprises: Molecular input module for inputting query compounds; The centroid calculation module is used to calculate the centroid of the query compound; Molecular transformation module for translation and/or rotation transformation of query compounds based on centroids; Molecular alignment module is used for parallel computing through GPU, and the query compound before and after transformation is used as the starting search point to compare with the molecular compounds in the molecular library; The result output module is used to filter according to the calculation result of the comparison, and output the molecular compounds similar to the query compound. 8. A GPU parallel computing molecular similarity system, wherein the system comprises at least one processor; And, a memory communicatively coupled to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the execution of any of claims 1-7 GPU parallel computing molecular similarity method. 9. A non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores computer-executable instructions that when executed by one or more processors , which can cause the one or more processors to execute the GPU parallel computing molecular similarity method according to any one of claims 1-7.

Images