42 results about "Atomic coordinates" patented technology

Systems and methods for characterizing an interaction between a compound and a polymer include obtaining a plurality of sets of atomic coordinates. Each set of atomic coordinates comprises the compound bound to the polymer in a corresponding pose in a plurality of poses. Each respective set of atomic coordinates, or an encoding thereof, is sequentially inputted into a neural network, to obtain a corresponding initial embedding as output, thereby obtaining a plurality of initial embeddings. Each initial embedding corresponds to a set of atomic coordinates in the plurality of sets of atomic coordinates. An attention mechanism is applied to the plurality of initial embeddings, in concatenated form, to obtain an attention embedding. A pooling function is applied to the attention embedding to derive a pooled embedding. The pooled embedding is inputted into a model to obtain an interaction score of the interaction between the compound and the polymer.

The invention relates to the technical field of semiconductor manufacturing, and discloses a wafer doping control method based on intelligent feedback adjustment, and the method comprises the steps: carrying out the atomic scale imaging of the surface of a wafer, so as to recognize a target doping site, and generating an atomic coordinate mapping graph, secondly, inputting the atomic coordinate mapping graph into a multi-scale simulation model to generate an initial process parameter set, and processing the initial process parameter set by combining a transformer model with a reinforcement learning algorithm to obtain a dynamic process data matrix; performing optimization processing on the dynamic process data matrix through a cross-material process migration graph neural network to obtain a process parameter migration mapping relation, comparing measured data with a multi-scale model prediction result through closed-loop feedback iteration, and outputting a corrected process parameter migration mapping relation and a multi-scale model updating instruction; according to the wafer doping method, the stability, the consistency and the yield of the wafer doping process are improved, and high accuracy and high uniformity of wafer doping are realized.

The invention discloses a method for identifying a molecular structure and constructing a reaction network based on atomic coordinates, and belongs to the technical field of material simulation. According to the method for recognizing the molecular structure based on the atomic coordinates, by obtaining trajectory data, atoms outside a box are converted into atoms inside the box through minimum mirror agreement; grids are divided according to a space truncation radius, potential bonding atoms are accurately screened, and redundancy of distance calculation between every two atoms of the whole system is avoided; atomic clusters are quickly divided in combination with depth-first search, and molecular structure recognition is achieved. The method only depends on basic data such as atomic coordinates and does not depend on a simulation software post-processing module, and the limitation that only a single simulation system is adapted in the prior art is broken through. According to the method for constructing the reaction network, molecular structure description is compared frame by frame, and invalid reactions are offset and filtered through positive and negative reactions to obtain net reactions; counting the net reaction times to determine the reaction probability so as to quantify the tendency of the reaction; and constructing a reaction network by taking the molecular structure as a node and the reaction frequency as an edge weight.

Embodiments of the present application provide a three-dimensional molecular structure generation method, device, equipment and storage medium, at least applied to the field of artificial intelligence and the field of drug synthesis, wherein the method comprises: obtaining atomic type noise and atomic coordinate noise at a current time step; constructing a full connection adjacency matrix based on the atomic type noise; performing three-dimensional isometry processing on the atomic type noise, the atomic coordinate noise and the full connection adjacency matrix to obtain a molecular structure distribution average value at the current time step; generating a molecular structure of the molecule iteratively based on the molecular structure distribution average value at the current time step to obtain a molecular representation of the molecule; and constructing the three-dimensional molecular structure through the molecular representation. Through the present application, the efficiency of three-dimensional molecular structure generation can be accelerated and cumulative errors can be avoided, so as to accurately generate a three-dimensional molecular structure of an effective molecule.

The application relates to the technical field of semiconductor manufacturing, and discloses a wafer doping control method based on intelligent feedback adjustment, which comprises the following steps: atomic scale imaging is performed on the surface of a wafer to identify target doping sites and generate an atomic coordinate mapping diagram, the atomic coordinate mapping diagram is input into a multi-scale simulation model to generate an initial process parameter set, a dynamic process data matrix is obtained by combining a transformer model with a reinforcement learning algorithm to process the initial process parameter set, then, the dynamic process data matrix is optimized by a cross-material process migration graph neural network to obtain a process parameter migration mapping relationship, and measured data and multi-scale model prediction results are compared through closed-loop feedback iteration and output of a corrected process parameter migration mapping relationship and a multi-scale model update instruction. The application improves the stability, consistency and yield of wafer doping processes, and realizes high precision and high uniformity of wafer doping.

The invention discloses a method for screening local vibration modes of an atomic system. The method comprises the following steps of: 1, acquiring an atomic structure file, a target atomic coordinate file and a feature vector file containing a real part and an imaginary part of vibration displacement of each atom; 2, analyzing the coordinate file to extract lattice basis vectors, atomic coordinates and element information; 3, fusing the real parts of the feature vectors to generate a total vibration vector file of each atom in a frequency domain; step 4, generating visual vibration data based on the atomic structure and the total vibration vector; 5, setting an incident light direction, extracting a vibration component in the direction, and generating a transitional vibration vector in combination with the coordinate file; step 6, selecting frequency and fusing transitional vectors to generate Raman / infrared spectrogram data; and step 7, optimizing the spectrogram through Gaussian broadening to complete screening. According to the invention, vector synthesis can be carried out on the displacement of each atom in each direction under different frequencies and visual display is carried out, so that atom vibration can be observed more visually.

The present application relates to the technical field of computers, and relates to a kind of amorphous material atomic structure multiscale automatic analysis method. Including: step S1: with reactant as initial unit, based on first principle molecular dynamics simulation obtains the cluster unit of reaction product;Step S2: the cluster unit is constructed, and its cell parameters and atomic coordinates are optimized using first principle calculation, to obtain cell unit;Step S3: the cell unit is stacked in space, and its cell parameters and atomic coordinates are optimized using first principle calculation, to obtain supercell unit;Step S4: the supercell unit is simulated by first principle molecular dynamics, to obtain accurate machine learning potential function;Step S5: the supercell unit is constructed by cell expansion, and the molecular dynamics simulation is carried out using accurate machine learning potential function, to obtain amorphous material atomic structure;Step S6: steps S1 to S5 are packaged into automation workflow. The method is efficient, and the error is small.

The invention provides a method for reconstructing a three-dimensional atomic configuration of a special interface, and the method comprises the steps: combining actual representation data with theoretical calculation, generating a theoretical template, and carrying out the matching with an actual image, thereby improving the precision of atomic coordinate extraction; the position of each atom is calibrated through Gaussian fitting, and it is ensured that the precision reaches the sub-angstrom scale; atomic phase information in a two-dimensional transmission electron microscope image is utilized, and a direction matrix and coordinate transformation are combined to realize reconstruction from a two-dimensional atomic projection image to a three-dimensional crystal structure.

The application relates to a multi-descriptor and coordinate reversible mapping atomic clustering analysis method and system, which comprises the following steps: inputting a set of multiple descriptors representing atomic arrangement characteristics; selecting single descriptor data or comprehensive descriptor data obtained by multi-descriptor weighted combination according to system characteristics by a user; performing clustering analysis on the selected descriptor data, and generating descriptor data grouping and a sample index set corresponding to each group; extracting the Coulomb matrix of each sample in each group from the set of multiple descriptors based on the sample index set of each group, forming a Coulomb matrix subset of each group and respectively calculating the Coulomb matrix mean value; and applying a multi-dimensional scaling algorithm to reduce the Coulomb matrix mean value of each group to a three-dimensional space to generate corresponding three-dimensional atomic coordinates of each group. Compared with the prior art, the application improves the analysis capability of atomic or molecular arrangement characteristics in the data of simulation outputs such as molecular dynamics.

The application relates to a training method and device of a crystal property prediction model and a computer device. The method comprises the following steps: obtaining a sample crystal and a property label corresponding to the sample crystal; the sample crystal comprises a plurality of sample atoms; determining respective atomic coordinates of each sample atom, and determining orbital characteristics of the sample atoms according to the atomic coordinates; determining respective atomic potential energies of the sample atoms according to the orbital characteristics, and determining a predicted property corresponding to the sample crystal according to the atomic potential energies; training the crystal property prediction model according to the difference between the predicted property and the property label, stopping when a first training stop condition is reached, and obtaining a trained crystal property prediction model; and the trained crystal property prediction model is used to determine the property of the crystal. The method can improve the accuracy of predicting the property of the crystal.

The application discloses a kind of covalent crystal structure intelligent prediction method based on molecular unit topological assembly, by regarding molecular unit as rigid body and using hybrid gene coding system, topological connection and geometric pose decoupling are realized, compared with traditional atomic coordinate random search, the application can keep the internal configuration of molecule stable during evolution, avoid the collapse of geometric configuration of complex molecule during search process, and improve the calculation efficiency of large-scale, non-spherical molecular system;The method does not depend on specific empirical parameters, and can cover the structure prediction of fullerene polymer, covalent organic framework and complex porous material through general topological assembly logic, and has strong cross-field universality.

The application provides an electron density information acquisition method, system, medium, terminal and program product based on machine learning. The method comprises the following steps: first, constructing a real differential density according to a real electron density and a reference electron density; inputting atomic species, atomic coordinates, cell information and uniform grid definition information of a target structure into a machine learning model for training, and taking the real differential density as a supervision target during the training to construct a differential density prediction model; deploying the constructed differential density prediction model to output a predicted differential density of a to-be-tested complex structure according to to-be-tested complex structure information; and performing a recovery operation on the output predicted differential density to obtain total electron density information of the to-be-tested complex structure. The application can acquire electron density information of a complex structure with high precision and high efficiency, and stably support downstream calculations such as photoelectric properties and quantum transport of heterostructures, interface structures and device structures.

The invention relates to a protein binding conformation prediction method and system. The protein binding conformation-based prediction method comprises the following steps: acquiring a static structure file of a target protein, and extracting atomic information to construct a residue index table; performing standardization processing on the atomic coordinates in the residue index table, and calculating a distance matrix between the residue pairs based on the standardized atomic coordinates; based on the conformation sampling data and the residue index table, obtaining a residue motion freedom degree through adaptive weighting calculation and correlation correction; a dynamic distance threshold value is obtained based on the residue motion freedom degree; converting the distance matrix into a binary contact graph based on the dynamic distance threshold, obtaining the contact frequency of the residue pair, and constructing a contact frequency matrix; extending the residue motion degree of freedom into a symmetric matrix, and fusing the contact frequency matrix and the symmetric matrix to obtain an initial dynamic probability matrix; performing normalization processing and probability propagation on the initial probability matrix in sequence to obtain a dynamic probability matrix; protein binding conformations are predicted based on the dynamic probability matrix.

This invention discloses a molecular dynamics calculation method and system based on quantum simulation. The method first applies a discrete quantum Fourier transform to the atomic coordinates of the molecular system, mapping the original plane wave basis to a system containing only O(N) atoms. 2 The invention constructs a plane-wave bibasic quantum mechanics equation; then, it constructs and solves the electronic structure Hamiltonian on a quantum computer to obtain the ground state energy. This quantum energy is combined with the classical empirical potential energy according to predetermined weighting coefficients to calculate the total synthesized energy, from which the atomic forces are determined. The integration time step is then adaptively adjusted based on a comparison of the quantum energy change between consecutive time steps with a preset threshold. This adaptive mechanism shortens the time step to ensure accuracy when quantum effects are significant, and lengthens the step to improve efficiency when changes are gradual. This invention significantly reduces Hamiltonian complexity, balancing quantum accuracy and classical efficiency, and is applicable to various fields such as drug screening, materials design, and catalytic reaction simulation.

The invention discloses a method for exploring the influence of electromagnetic waves on a protein structure directly based on molecular dynamics simulation. The method comprises the following steps: acquiring target protein from a protein database; constructing a simulation system; adding a protein force field and a water molecule model; solvent molecules and ions are added to neutralize system charges, and energy minimization treatment is carried out; pre-balancing is carried out under the NVT ensemble and the NPT ensemble; performing molecular dynamics simulation on the simulation system, and storing an atomic coordinate trajectory; and performing post-processing on the track file, extracting distance time sequence data between atoms or residues, screening out atom pairs or residue pairs meeting conditions, performing Fourier transform analysis, and judging whether the possibility of resonance with the electromagnetic wave with the specific frequency exists or not. According to the method, the interaction between the protein and the electromagnetic field can be revealed under the atomic resolution without consumption of experimental materials, and an efficient and reliable molecular-level theoretical basis is provided for evaluating the biological effect of environmental electromagnetic radiation.

The invention discloses a wettability-adjustable porous quartz atomic-scale model generation method based on CT scanning data, and relates to the technical field of petroleum engineering and molecular simulation, and the method comprises the following steps: obtaining template coordinates; generating a quartz substrate; carrying out atom screening and hollowing; performing surface atom cleaning and surface hydroxylation treatment on the initial atomic model; on the basis of a wettability field obtained through CT scanning, the functional group type of a hydroxylation area on the pore surface is automatically adjusted, and a target atom model is obtained; and exporting the atomic coordinates of the target atomic model as a standard atomic structure file. According to the method, an atomic-scale model with geometric fidelity, chemical stability and adjustable wettability can be automatically constructed from CT scanning data, so that the physical consistency and prediction precision of molecular simulation on real rock core wetting behaviors are improved, and a high-quality initial structure is provided for molecular dynamics simulation and oil and gas flow research.

The application discloses a three-dimensional molecular structure generation method and model, relates to the technical field of the cross of computational chemistry and machine learning, and the three-dimensional molecular structure generation method comprises the steps of: generating an initial three-dimensional conformation by adopting a preset distance geometry algorithm in combination with a preset molecular force field optimization based on molecular topological information; constructing three SE(3) isovariant geometric graphs according to the initial three-dimensional conformation and the molecular topological information, and generating isovariant node representations of the graphs based on the SE(3) isovariant geometric graphs; fusing the isovariant node representations of the SE(3) isovariant geometric graphs to obtain a molecular embedding; predicting atomic coordinate offsets according to the molecular embedding and preset DFT true value coordinates; and generating a target three-dimensional molecular structure according to the initial three-dimensional conformation and the atomic coordinate offsets. The application can realize high-precision, high-efficiency and strong-robustness three-dimensional conformation prediction under limited computing resources and data scale.

The invention discloses an execution method and device of a molecular dynamics constraint algorithm, a chip and a medium. The execution method comprises the steps that a GPU chip sequentially determines a molecule as a current molecule in a plurality of substance molecules to be processed; obtaining an atomic coordinate and an atomic velocity corresponding to the current molecule, and executing a molecular dynamics constraint algorithm operation on the current molecule according to the atomic coordinate, the atomic velocity and a plurality of predetermined target parameters; the multiple target parameters comprise atomic position correlation parameters and PV values; judging whether operation on all the molecules is completed or not; and if not, returning to execute the operation of sequentially determining one molecule as the current molecule in the plurality of to-be-processed substance molecules until the operation on all the molecules is completed. According to the technical scheme of the embodiment of the invention, the execution efficiency of the GPU chip on the molecular dynamics constraint algorithm can be improved.

This invention discloses a heterogeneous node electrostatic force calculation device for molecular dynamics simulations, comprising: a CPU node for constructing molecular topology, calculating Gaussian width parameters, initializing short-range force lookup tables, and outputting atomic trajectories and system energy after simulation; an FPGA node connected to the CPU node via a PCIe bus for selecting short-range atom pairs based on truncation radius and atomic coordinates, and calculating short-range electrostatic forces in parallel using a lookup table; and a GPU node connected to both the FPGA and CPU nodes via the same PCIe bus for performing long-range electrostatic force calculations on all atoms in the system, and updating atomic positions and velocities after reducing short-range and long-range forces. The heterogeneous node electrostatic force calculation device for molecular dynamics simulations provided by this invention, through reasonable task allocation, fully utilizes the advantages of GPUs and FPGAs to achieve asynchronous parallel computing between GPUs and FPGAs, thereby improving the overall speed of electrostatic force calculations.

This invention discloses a method for simulating and analyzing the mechanical behavior of atomic-scale amorphous carbon materials, relating to the fields of materials science and computational technology. The method first establishes a diamond supercell with an atomic number equivalent to the target amorphous carbon model and melts it to form liquid carbon. The amorphous carbon model is then obtained by cooling and depressurization. Next, the structural parameters of the amorphous carbon model in its initial state are acquired, strain is applied to the model, and structural relaxation is performed. Then, the Young's modulus of the amorphous carbon model and the structural parameters under different strains are calculated. Using the atomic coordinate information of the initial model and different strains, the von Mises strain of each atom in the model is calculated. Finally, the number of atomic hybridization transitions and their spatial distribution characteristics in the model are statistically analyzed under different strain increments. This invention can accurately characterize the structural evolution and mechanical response of amorphous carbon materials during mechanical loading, providing an effective simulation method for revealing the microscopic mechanisms of the mechanical behavior of amorphous carbon materials.

The present disclosure provides a chemical bond stress visualization method, which can be applied to the technical field of cheminformatics. The chemical bond stress visualization method comprises: obtaining an input file corresponding to a target molecule, wherein the input file contains coordinate file identification information corresponding to the target molecule; reading a coordinate file from a target folder according to the coordinate file identification information, wherein the coordinate file comprises atomic coordinates of each atom in the target molecule; determining an atomic distance between each two atoms according to the atomic coordinates to obtain an atomic distance table, wherein the atomic distance is used to represent the chemical bond stress between the two atoms; and drawing a molecular model corresponding to the target molecule according to the atomic coordinates and the atomic distance in the atomic distance table, wherein different chemical bond stresses in the molecular model are represented by different colors. The present disclosure also provides a chemical bond stress visualization device.

This invention discloses a method for simulating phase separation of biological macromolecules. The method includes: constructing a coarse-grained model of biological macromolecules; constructing force field parameters and structural constraints; constructing a two-phase coexistence state calculation model; and conducting molecular dynamics simulations of the two-phase coexistence state to obtain simulated trajectories based on phase separation of biological macromolecules. The coarse-grained model of biological macromolecules defines coarse-grained units according to molecular type, wherein proteins use amino acids as basic units, RNA uses ribonucleotides as basic units, and DNA is split into two basic units by deoxyribonucleotides. It also determines the monomer structure of biological macromolecules, wherein structural proteins and RNA use specific atomic coordinates in their full atomic structure as coarse-grained sphere coordinates, and naturally disordered proteins and RNA generate coarse-grained sphere coordinates through random movement within a cubic box.

The invention discloses a method for calculating the density of a thin film based on molecular dynamics, which belongs to the technical field of detection and can be used for quickly predicting and screening an optimal process window on a computer at low cost before an actual deposition experiment is carried out. By simulating the deposition process under different parameters in batches and calculating the density, research and development personnel can preferentially perform experimental verification in the most potential parameter range, so that the expensive'preparation-measurement 'cycle index is greatly reduced, and the research and development expenditure and the material cost are remarkably saved. Meanwhile, different from a traditional experimental method which can only provide macroscopic average density, the method is based on atomic coordinates for calculation and has atomic-scale resolution capability, the overall average density of the thin film can be given, and the density distribution gradient of the thin film in the growth direction or the density condition of a specified local area can be further analyzed. Therefore, how different deposition parameters affect the atomic deposition behavior is fundamentally revealed.

The invention discloses an atomic element type identification method and related equipment, and relates to the technical field of artificial intelligence and machine learning, and the method comprises the steps: obtaining an initial three-dimensional image of a nanometer material based on an atomic resolution electronic three-dimensional reconstruction technology, carrying out the noise reduction operation of the three-dimensional image through employing a preset three-dimensional global-local residual error U-shaped network, and carrying out the recognition of the atomic element type. The method comprises the steps of obtaining a target three-dimensional image, determining three-dimensional atomic coordinates and a 3D Zernike expansion coefficient corresponding to the target three-dimensional image, and processing the three-dimensional atomic coordinates and the 3D Zernike expansion coefficient based on a preset DAST network model to obtain an atomic element type recognition result corresponding to the nanometer material. On the basis of the GLARE neural network, overall affine transformation operation is carried out on the initial three-dimensional image, the problems of reconstruction volume distortion and the like caused by overall alignment errors are avoided, the accurate and denoised target three-dimensional image is generated, and furthermore, the atomic element type is accurately recognized in combination with the preset DAST network model.

The application discloses a protein essentiality prediction method based on multi-modal double fusion, comprising the following steps: 1, obtaining data containing essential protein sequences and non-essential protein sequences and preprocessing; 2, using a large language model to extract residue embedding matrix, residue contact matrix and three-dimensional atomic coordinates from the preprocessed data; 3, extracting each modal information from the large language model through a feature processing module; 4, realizing the fusion of one-dimensional sequence and two-dimensional structure features based on a fuzzy self-adaptive clustering fusion module; 5, realizing the fusion of one-dimensional sequence and three-dimensional structure features based on a bidirectional co-attention fusion module; 6, constructing a protein essentiality classification layer network to realize the prediction of protein essentiality. The application solves the problems of single feature representation and insufficient sequence and structure feature fusion in the traditional protein essentiality prediction method, and improves the accuracy and robustness of protein essentiality prediction.

The invention relates to the technical field of computers, in particular to a model training method, a molecule generation method, a data processing device and a computer readable storage medium. The model training method includes: acquiring a training data set including a plurality of original data samples, each original data sample including first atom attribute data for a plurality of atoms of a molecule, first key attribute data for the plurality of atoms, and first atom coordinate data for the plurality of atoms; encoding, by means of an encoder, the raw data sample as a hidden feature for the molecule; decoding, by means of a decoder, the hidden feature into a reconstructed data sample comprising second atomic attribute data for a plurality of atoms of the molecule, second key attribute data for the plurality of atoms, second atomic coordinate data for the plurality of atoms; and training an encoder and a decoder according to the difference between the original data sample and the reconstructed data sample.

This invention relates to the field of computer-aided process design technology, specifically to a method and system for optimizing the formulation of biodegradable tableware materials. The method includes the following steps: generating equilibrium molecular conformations based on atomic coordinates and calculating initial formulation component parameters; converting modulus parameters to generate theoretical elastic recovery strain parameters; mapping models to calculate deformation and marking diffusion characteristic length parameters; calculating diffusion coefficients and flux parameters to update local hydrogen ion concentration indices; updating reaction rate constants and integrating them to generate finalized formulation structural parameters. In this invention, a microscopic molecular dynamics model is constructed to analyze interface voids to quantitatively assess compatibility; relaxation time spectra are used to invert elastic recovery strain to predict deformation risk; autocatalytic reaction and diffusion theories are combined to track hydrogen ion concentration evolution; a dynamic mapping between degradation and failure is established; and full life-cycle verification is completed in a virtual environment, significantly shortening the R&D cycle and reducing trial-and-error costs.

The invention particularly discloses a mechanical behavior simulation analysis method for an atomic-scale amorphous carbon material, and relates to the technical field of material science and calculation. The method comprises the following steps: firstly, establishing diamond supercells equivalent to a target amorphous carbon model in number of atoms, melting to form liquid carbon, and cooling and depressurizing to obtain an amorphous carbon model; secondly, obtaining structural parameters of the amorphous carbon model in an initial state, applying strain to the amorphous carbon model, and performing structural relaxation; then Young modulus of the amorphous carbon model and structural parameters under different strains are calculated, and von Mises strain of each atom in the model is calculated through the initial model and atomic coordinate information under different strains; and finally, counting the transformation quantity of the atomic hybridization mode under different strain increments and spatial distribution characteristics in the model. According to the invention, the structural evolution and mechanical response of the amorphous carbon material in the mechanical loading process can be accurately represented, and an effective simulation means is provided for revealing the microscopic mechanism of the mechanical behavior of the amorphous carbon material.

This invention discloses a method and system for dislocation defect identification based on image recognition and deep learning networks, belonging to the fields of image recognition and machine learning. The method first calibrates the dimensions of a high-resolution electron micrograph, identifies atomic coordinates using a deep learning object detection model, and calculates the local atomic environment vector for each atom. Second, using an atomic model dataset covering various materials, dislocation types, and deformation states, a deep learning network is trained capable of predicting the distance between atoms and dislocations based on the local environment vector. Finally, the image is meshed, and the neighboring atomic information of each mesh node is extracted. Combined with the dislocation distance information output by the trained deep learning network, unsupervised clustering is used to identify the dislocation location. This method achieves automated and high-precision identification of dislocation defects in face-centered cubic materials, while also enabling dynamic characterization of crystal defect deformation behavior.

This invention provides a molecular identification and extraction method based on ReaxFF reaction kinetics simulation, belonging to the field of molecular identification and extraction technology. To obtain molecular structure information from reaction kinetics simulations, this invention obtains a dump source file containing bond orders and coordinates from the ReaxFF simulation output. It stores the bond order information at each time step in the bond source file, extracts and stores the atomic information of all molecules at the same time step, determines the target molecule features, identifies and stores the target molecule features based on the obtained atomic ID information of all molecules at the same time step, stores the coordinate information of the dump source file at each time step, and iterates through the atomic coordinate information and atomic IDs of all target molecules at the same time step to obtain the atomic coordinate information of the target molecule. This invention is applied to the identification and extraction of molecular structure information.