50 results about "Cheminformatics" patented technology

The invention relates to the technical field of chemical engineering and machine learning, and provides a method for predicting the solubility of carbon dioxide in a deep-eutectic solvent based on a machine learning model, and the method comprises the steps: collecting CO2 solubility data, carrying out the coding and feature descriptor conversion of a deep-eutectic solvent molecular structure through an open-source chemoinformatics software package RDKit, and constructing a data set. Four algorithms of XGBoost, random forest, a deep neural network and a convolutional neural network are adopted to construct a model, and hyper-parameters are optimized through grid search and k-fold cross validation. After the model is trained, the prediction performance is evaluated, and the optimal model is selected for CO2 solubility prediction. The method improves prediction efficiency and accuracy, has strong adaptability and generalization ability, and provides an effective tool for formula optimization of the eutectic solvent and improvement of CO2 capture efficiency.

The invention belongs to the field of wastewater treatment and industrial automatic control, and particularly relates to an intelligent dosing control method based on water quality online monitoring. The method comprises the following steps: capturing a water body spectrum fingerprint spectrum by using a full-spectrum scanning module to obtain a characteristic signal sequence; a chemical informatics molecular algorithm is introduced for preprocessing and feature enhancement, and molecular fingerprint information is recognized; a convolutional neural network model is constructed to decouple chemical oxygen demand, ammonia nitrogen, total phosphorus and water toxicity indexes from the spectral signals, and multi-parameter parallel inversion is achieved; and based on the inversion index and the spectral dynamic trend, the dosage is calculated through an intelligent decision-making algorithm, and accurate dosing of the medicament is driven. According to the invention, through fusion of full spectrum scanning and deep learning, high-precision anti-interference water quality perception is realized, potential risks can be identified and preventive intervention can be carried out, and the intelligent level and economical efficiency of dosing control are improved.

The invention belongs to the field of material science and chemical informatics, and particularly discloses a polymer graph convolutional neural network method based on uncertainty guidance, and the method comprises the steps: carrying out the graph structure coding of a polymer monomer; performing property prediction and uncertainty evaluation on the data set; training an uncertainly guided active learning acceleration model; a global attention mechanism identifies key functional groups; and high-throughput screening of polymers with specific target properties is carried out. According to the method, through the representation capability of the complex topological relation of the graph neural network and uncertainty quantification of evidence deep learning, the polymer features are effectively extracted, efficient sample training is achieved, and the method is particularly suitable for the data scarcity condition; the relation between the polymer structure and the performance can be deeply excavated, and recognition of high-quality polymers is accelerated.

The invention provides a method for extracting chemical reaction information from journal papers, and relates to the technical field of chemical informatics. The method comprises the following steps: firstly, extracting a reaction information unit from a journal paper PDF (Portable Document Format) file; carrying out reaction information modal segmentation on the reaction information units; automatically segmenting the extracted reaction information unit into three types of picture information including a chemical reaction expression, a table and a conditional description text; respectively analyzing the segmented chemical reaction expression picture, the table picture and the conditional description text picture; and finally, integrating the analyzed information into complete chemical reaction information. Meanwhile, the extracted, segmented, analyzed and integrated reaction information is stored in an optimized mode, and subsequent use and updating are facilitated. The method can be used for efficiently and accurately extracting reactants, products, reaction conditions, yield, remarks and other information related to the chemical reaction in the journal paper.

The invention discloses a UNIFI compound identification result classification and model construction method based on machine learning, and belongs to the field of natural product chemical informatics. The method comprises the following steps: acquiring a compound identification result table output by a UNIFI platform; candidate compounds in the table are manually rechecked and dichotomy supervision labels are labeled to form a training data set; extracting a group of specific nine predetermined numeric features from the information of each candidate compound; and training a machine learning algorithm to obtain a classification model by taking the nine features as input and taking a corresponding label as a target. During application, nine identical features are extracted from a new identification result table and input into the model, and then identification credibility categories and matching degree scores can be output. According to the method, rapid, objective and automatic classification of UNIFI high-throughput identification results is realized through a fixed label and lightweight feature strategy, and the re-checking efficiency and the result consistency are remarkably improved.

The present invention relates to the technical field of chemical informatics, and in particular to a method for predicting small molecule structures based on equivariant flow matching modeling. The present invention comprises obtaining a data set and real 3D molecular graph information in the data set; constructing a flow matching model, performing flow matching to obtain the flow matching time, the probability path of the real molecular coordinates, and the real vector field; constructing and training a Modified Equiformer model; randomly initializing initial molecular coordinates of the same size that satisfy a Gaussian distribution based on the number of atoms in the real 3D molecular graph information, inputting the initial molecular coordinates, the flow matching time, and the trained Modified Equiformer model into a black box ODE solver to generate predicted 3D molecular graph information, thereby obtaining a predicted molecular structure graph. The present invention improves the accuracy of small molecule structural information and significantly increases the speed and accuracy of structure prediction.

The application discloses a reaction optimization method based on double-view contrast learning and double-layer attention, and relates to the technical field of chemical informatics and artificial intelligence. The method comprises the following steps: constructing a double-view molecular representation model, extracting molecular features from a graph structure view and a path sequence view, and pre-training the model through self-supervised contrast learning; constructing a chemical reaction prediction model; pre-training the chemical reaction prediction model by using the extracted molecular features in a chemical reaction data set; taking the reaction prediction model as a proxy model of Bayesian optimization; and realizing the collaborative and rapid optimization of material variables such as reactants, catalysts and solvents, and numerical condition variables such as temperature and concentration. The application effectively solves the problems of missing molecular representation information and low optimization efficiency of high-dimensional mixed variables in the prior art, and significantly speeds up the chemical research and development process.

The application discloses a chemical reaction diagram analysis method and system, and relates to the technical fields of chemical informatics and multi-agent cooperation, which extracts structured entities of a chemical reaction diagram through a molecular agent, an arrow agent and a text agent, constructs a space-semantic diagram and a chemical perception diagram in parallel based on the entity set, aligns and fuses node and edge weights, generates a unified candidate edge set, and then performs conflict pruning in combination with space consistency, chemical consistency and arrow direction constraints to obtain a unified reaction reasoning diagram; candidate structure search based on a maximum weight generation forest algorithm is executed on the diagram, and the final reaction structure is determined through element conservation and formal charge balance auditing, and a structured reaction equation is output. The application can improve the analysis accuracy under a complex reaction topological layout.

The present invention discloses a method and system suitable for data mining of energetic compounds, belonging to the technical field of structure-activity relationship of energetic compounds in chemical informatics calculation, comprising the steps of: S1, data preprocessing; S2, extracting molecular topological structure features and features of quantum chemical calculation to generate fingerprint features; S3, calculating Pearson correlation and information entropy, screening out linearly correlated and non-correlated features, performing feature dimensionality reduction, and deleting redundant features whose linear correlation between features is higher than a set value; S4, knowledge mining and model verification, exploring the correlation and knowledge association between feature libraries. The present invention can classify data of energetic materials according to skeletons and groups, mine multi-dimensional linear features and non-linear features, embed more features based on chemical expertise into the graph structure model training process, and utilize the data mining engine of the present invention to screen out strongly correlated molecular features, which can make up for the shortcomings of existing models.

This invention relates to the field of cheminformatics. It provides a multi-scale reaction representation learning method based on reaction structure constraints. The method includes the following steps: acquiring chemical reaction data; identifying structural change regions based on the structural identification information and mapping these regions to corresponding structural unit locations to generate structural unit-level annotation information; performing a local structure modeling task on a perceptual model based on the structural representation sequence to enable the perceptual model to perceive structural change regions; and training the perceptual model based on a global structure mapping task to obtain a multi-scale reaction representation model. This invention can introduce reaction structure change constraints during model training and model the mapping relationship between local structural change information and the overall structure at different structural scales, while simultaneously achieving effective synergy between the two, thereby improving the model's performance in organic synthesis reaction modeling tasks.

The invention discloses a molecular property prediction-oriented hierarchical guidance collaborative attention method, and relates to the crossing field of artificial intelligence and chemoinformatics. The method comprises the following steps: S1, constructing atomic-scale graph representation and group-scale graph representation of molecules; s2, providing top-down context guidance for atomic-scale attention calculation through the group-level semantic features; s3, dynamically adjusting the weight of cross-scale information fusion by using a context gating mechanism; s4, breaking the independence of each attention head in the traditional multi-head attention by adopting a multi-head collaborative attention mechanism; and S5, generating a molecular representation fusing the coarse-grained functional semantics and the fine-grained local structure for property prediction. According to the method, a hierarchical guidance collaborative fusion mechanism is introduced, the problem of information conflict caused by simple splicing or summation in a traditional multi-scale fusion method is solved, collaborative enhancement of atomic-level details and group-level functional semantics is achieved, the accuracy and robustness of molecular property prediction are remarkably improved, and the method has good application prospects. And an efficient multi-scale molecular representation learning tool is provided for drug discovery and material design.

This application designs an artificial intelligence-based molecular characterization and soft sub-scaffold-constrained drug discovery system, comprising: a data collection module, a sub-scaffold analysis module, a 3D simulated molecular docking module, a Bayesian ridge regressor training module, a VAE generative model pre-training module, a reinforcement learning module, and a sampling evaluation module. By using sub-scaffold analysis and 3D simulated molecular docking information, this application can better understand and utilize the complexity and diversity of chemical space. Furthermore, by using a Bayesian ridge regressor and a VAE generative model, this application can generate novel, rationally structured, and potentially active drug molecules. Therefore, this artificial intelligence-based molecular characterization and soft sub-scaffold-constrained drug discovery system, by combining deep learning technology, cheminformatics, drug design, and computer-aided drug design methods, can significantly improve the efficiency and success rate of drug discovery.

The application discloses a molecular odor prediction method based on cross-modal attention fusion, relates to the fields of chemical informatics and deep learning, and comprises the following steps: acquiring a molecular dataset, extracting basic molecular property features, molecular fingerprint features and SMILES sequence features for each molecule in parallel, performing molecular odor prediction model training, wherein the molecular odor prediction model comprises a modal encoder, a cross-modal attention fusion module and a classification head connected in sequence, inputting the corresponding basic molecular property features, molecular fingerprint features and SMILES sequence features of a to-be-predicted molecule into the trained molecular odor prediction model to obtain prediction probabilities of corresponding odor categories, and determining an odor category prediction result of the to-be-predicted molecule based on the prediction probabilities. Through multi-modal feature fusion, an interpretable cross-modal attention and a gating mechanism, the method realizes more reasonable representation modeling of molecular odor, and effectively improves the accuracy of odor label prediction.

The invention discloses a screening method of skin fibroblast marker protein targeting components. The screening method comprises the following steps: (1) constructing a human skin fibroblast cell membrane positioning marker protein data list; (2) obtaining a membrane localization marker protein targeting component; (3) predicting chemical similarity; and (4) molecular docking verification. Database mining, chemoinformatics and molecular docking technologies are comprehensively applied, the HSF membrane localization marker protein beneficial to targeted research is accurately locked, a targeted component set of the HSF membrane localization marker protein is obtained, and an accurate and effective scientific basis is provided for research and development of anti-aging cosmetics.

The invention relates to the technical field of chemical informatics, in particular to a small molecular structure prediction method based on equal variable flow matching modeling. The method comprises the following steps: acquiring a data set and real 3D molecular map information in the data set; constructing a flow matching model, and performing flow matching to obtain flow matching time, a probability path of real molecular coordinates and a real vector field; the method comprises the following steps of: constructing a modified Equform model and training the modified Equform model; initial molecular coordinates which are identical in size and meet Gaussian distribution are randomly initialized based on the number of atoms in real 3D molecular map information, the initial molecular coordinates, flow matching time and the trained Modified Equform model are input into an ODE solver of a black box to generate predicted 3D molecular map information, and a predicted molecular structure map is obtained. According to the invention, the accuracy of small molecule structure information is improved, and the structure prediction speed and precision are greatly improved.

This invention belongs to the interdisciplinary field of artificial intelligence and cheminformatics, and provides a method, inference method, and apparatus for constructing a molecular multimodal large language model based on parameter space alignment. The method includes: constructing an instance set; constructing a molecular multimodal large language model, including an adaptive weight generator, a graph neural network encoder, and a large language model; training the molecular multimodal large language model using the instance set to update the parameters of the adaptive weight generator; the graph neural network encoder extracting node-level feature representations of the molecular structure diagram; the adaptive weight generator obtaining the low-rank weight update amount of the target weight component in the large language model based on the node-level feature representation, and injecting the low-rank weight update amount into the original weights of the target weight component; and the large language model after injecting the low-rank weight update amount generating an actual text response based on the input training text command. By integrating the molecular structure into the parameter layer of the large language model, deep perception of the molecular structure is achieved without changing the length of the input sequence.

The invention discloses a method for comparing chemical components of different parts of abelmoschus manihot by integrating high-resolution mass spectrometry and chemoinformatics, which comprises the following steps: firstly, optimizing the extraction conditions of different parts of abelmoschus manihot, carrying out ultrasonic extraction, collecting primary and secondary mass spectrometry data by using ultra-high performance liquid chromatography-quadrupole-time-of-flight mass spectrometry, and comparing the primary and secondary mass spectrometry data with the chemical components of different parts of abelmoschus manihot. The method comprises the following steps: performing preliminary identification on compounds through database comparison and standard substance verification, converting mass spectrum data into a standard format, performing peak extraction, alignment and denoising, uploading the mass spectrum data to a GNPS platform, setting a spectrogram similarity threshold value and a minimum matching peak number, constructing a molecular network, realizing clustering and visual analysis of the compounds with similar structures, and finally, determining the quality of the compounds. And carrying out standardization processing on the identified compound list and relative abundance data, generating a heat map by adopting hierarchical clustering analysis, and revealing the similarity and difference of chemical compositions of different parts. According to the method, high-throughput and accurate comparison of chemical components of the abelmoschus manihot is realized, and technical support is provided for quality control of medicinal materials and comprehensive utilization of resources.

The invention belongs to the field of computer-aided drug design, and discloses a metal binding group extraction method based on chemical informatics analysis and modeling. The method comprises the following steps: constructing a metal binding compound data set; training and optimizing a metal combined compound prediction model by using a decision tree-based model and a graph neural network algorithm; using an SHAP interpretable method to analyze a relationship between the key features and a prediction result, and deriving a structure fragment with potential metal binding capacity; according to the attention score of the Attentive FP model, an atomic weight heat map of molecules in the test set is drawn, and key metal binding fragments are determined; molecules in the training set are fragmented, and representative substructures are extracted; a metal binding fragment chemistry library is compiled. According to the method, the problem that metal binding group extraction is limited to a compound structure data set is solved, efficient extraction of any metal ion binding group is realized, and discovery and research of metalloenzyme-related drugs are accelerated.

The application relates to the technical field of computational chemistry and cheminformatics, and discloses a FeNiCrCoCu high-entropy alloy hardness molecular dynamics optimization method, which comprises the following steps: constructing a lattice simulation model based on static configuration data of initial component proportions; outputting atomic coordinate data by simulating heat treatment on the model; performing molecular dynamics nanoindentation simulation on the lattice model to calculate hardness parameters of the initial component proportion high-entropy alloy material; combining collinearity feature elimination calculation and a recursive feature elimination algorithm to generate a hardness prediction descriptor sequence; inputting the sequence into a regression model to perform hardness prediction and update a historical hardness data set to continuously improve prediction accuracy until high-entropy alloy element components with high hardness are output; and the application modulates static features by using element component changes, converts basic data to a dynamic microstructure response level to make up for information faults, and further establishes a high-precision and light-weight data processing architecture.

The invention discloses a compound fat solubility prediction method based on a multilevel feature fusion graph neural network, and relates to the field of computational pharmacology and chemical information.The method comprises the steps that an atomic-scale graph structure, molecular-scale global features and group-scale substructure features of a target molecular compound are obtained; performing feature extraction on the atomic-scale graph structure to obtain an embedded representation of each atom in the target molecular compound, and performing global average pooling on the embedded representation of each atom to obtain an atomic vector; normalizing the molecular-level global features to form molecular vectors; splicing the atom vector, the molecule vector and the substructure vector to obtain a global vector; predicting the fat solubility of the target molecular compound according to the global vector; according to the method, an atom-molecule-substructure multilevel characteristic system is constructed, and the characteristics with different granularities and different properties are organically combined for combined prediction, so that more comprehensive molecular characterization and more accurate fat solubility prediction of molecular compounds are realized.

A quantitative structure-activity relationship model construction method based on model robustness relates to a quantitative structure-activity relationship model construction method, in particular to a quantitative structure-activity relationship local model construction method based on model robustness and a consistency model construction method of the quantitative structure-activity relationship local model, taking a compound liver metabolic clearance rate prediction model as an object, belonging to the field of chemical information science and bioinformatics. In view of the fact that the current global modeling method is difficult to cope with the complex modeling compounds of the liver metabolic clearance rate prediction model, the model robustness is poor, and the prediction performance is difficult to further improve, the QSAR local model based on the activity mechanism is difficult to be applied in practice, and the QSAR local model based on the chemical structure similarity lacks the technical means to guarantee the model robustness, the application provides a QSAR local model construction method based on model robustness, and a support vector machine regression technology is used to establish a QSAR local model.

The invention discloses a neuron toxic compound screening method based on a brain organoid model, and belongs to the technical field of chemoinformatics. The method comprises the following steps: constructing stably transferred human embryonic stem cell lines and dividing the stably transferred human embryonic stem cell lines into two groups, namely a first cell line and a second cell line; inserting a first expression element into a genome of a cell of the first cell line by using a gene editing technology, wherein the first expression element expresses neurons for labeling activation; inserting the second expression element into a genome of a cell of a second cell line by using a gene editing technology, wherein the second expression element is used for realizing measurability of neuronal activity; performing differentiation culture on the edited first cell line and the edited second cell line to respectively obtain a first brain organ and a second brain organ; the first brain organoid and the second brain organoid are used as experimental carriers, a neuron toxic compound screening platform is constructed, and the neuron toxic compound screening platform is tested.

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.

The embodiment of the invention relates to a processing method and device for a molecular conformation generator with conditional constraints. The method comprises the following steps: constructing a molecular conformation generator based on a plurality of Un -Mo < l > models and a property prediction model of which model training is completed; initializing model parameters of all Un -Mo < l > models in the molecular conformation generator on the basis of preset base model parameters; receiving a first molecule sequence, a first molecule property vector and a first generated molecule number input by a user; performing three-dimensional conformation conversion processing on the first molecular sequence based on a preset chemoinformatics tool to obtain a corresponding first conversion conformation; according to the first conversion conformation, the first molecule property vector, the first generated molecule number and the molecule conformation generator, batch molecule generation task processing is carried out, and a corresponding first batch molecule task report is obtained and fed back to a user. According to the method, the molecular design efficiency can be improved, and the exploration capability of unknown molecular structures is improved.

The invention relates to the crossing field of artificial intelligence safety and chemical informatics, and particularly discloses a chemical scene-oriented small language model safety risk assessment method and device, and the method comprises the steps: determining a corresponding risk classification standard and a knowledge base range according to a use scene preset by a user, and selecting an adaptive test dimension combination according to the corresponding risk classification standard and knowledge base range; calling a preset template library, a rule combiner or a large language model to generate a plurality of evaluation instructions; performing a corresponding evaluation step according to the evaluation instruction category and a preset rule; outputting a comprehensive evaluation conclusion, and generating an evaluation report; the chemical scene-oriented small language model behavior risk and knowledge reliability double evaluation method realizes double evaluation of the chemical scene-oriented small language model behavior risk and knowledge reliability, and provides support for blocking dangerous information diffusion.

The invention relates to the technical field of chemical informatics, and discloses an oral absorption rate prediction and compound structure optimization method, device and equipment, and the method comprises the steps: obtaining a compound sample data set which comprises a multi-dimensional characteristic sample parameter corresponding to each compound sample and a reference oral absorption rate; the multi-dimensional characteristic sample parameters comprise at least one of the following parameters: a fat solubility parameter, a polarity parameter, an electronic structural parameter and an in-vitro penetration parameter, and the multi-dimensional characteristic sample parameters further comprise a molecular structural parameter; and by taking the multi-dimensional characteristic sample parameters as input and the reference oral absorption rate as an output label, training a pre-constructed initial model to obtain an oral absorption rate prediction model. According to the method, molecular structural parameters are introduced into oral absorption rate prediction model construction, the oral absorption rate prediction precision of the model on compounds with complex structures and high flexibility can be remarkably improved, and more accurate data support is provided for subsequent compound structure optimization.

This invention relates to the field of cheminformatics technology, and discloses a method, apparatus, and device for predicting oral absorption rate and optimizing compound structures. The method includes: acquiring a compound sample dataset, which includes: multi-dimensional feature sample parameters corresponding to each compound sample and a reference oral absorption rate; the multi-dimensional feature sample parameters include at least one of the following: lipophilicity parameters, polarity parameters, electronic structure parameters, and in vitro permeability parameters, and also include molecular structure parameters; using the multi-dimensional feature sample parameters as input and the reference oral absorption rate as the output label, training a pre-constructed initial model to obtain an oral absorption rate prediction model. This invention incorporates molecular structure parameters into the construction of the oral absorption rate prediction model, which can significantly improve the model's accuracy in predicting the oral absorption rate of compounds with complex structures and high flexibility, providing more accurate data support for subsequent compound structure optimization.

The application discloses a chemical experiment product intelligent characterization analysis system and method based on deep learning, and belongs to the technical field of chemical informatics and artificial intelligence. The system comprises: a data acquisition and standardization preprocessing module, which is used for automatically acquiring original spectrum data of an analytical instrument and generating standardized spectrum data; a deep learning analysis engine, which is used for analyzing the standardized spectrum data to obtain an analysis result comprising at least a product structure; and a reaction path automatic tracking and knowledge graph construction module, which is used for automatically associating the analysis result with synthesis records in an upstream experiment execution system, inferring a conversion path and storing the conversion path in a reaction knowledge graph. The application realizes the automation and standardization of spectrum analysis, can automatically associate the analysis result with synthesis conditions and conversion paths to construct a knowledge graph, supports mixture analysis and complex compound identification, and realizes model iteration optimization through human-computer interaction feedback, thereby significantly improving the efficiency and accuracy of chemical experiment product characterization analysis.

The invention discloses a molecular structure-activity relationship modeling method based on symbol regression KAN, and belongs to the field of chemoinformatics. The method comprises the following steps: constructing and cleaning molecular property data to form a molecular-chemical descriptor-target property data set; calculating and screening descriptors, and reducing redundancy and colinearity by adopting variance screening and correlation analysis; and training the KAN to obtain a target property prediction function and exporting an explicit expression. By taking the organic amine CO2 absorption cycle period as an example, the upper / lower limit auxiliary screening of the cycle period can be predicted. According to the modeling method, an interpretable basis is provided under a small sample condition, and the dependence on a large amount of experimental data is reduced.

An embodiment of the present invention relates to a processing method and device for a molecular conformation generator with conditional constraints. The method includes: constructing a molecular conformation generator based on multiple Uni-Mol models and a property prediction model that has completed model training; initializing the model parameters of all Uni-Mol models in the molecular conformation generator based on preset base model parameters; receiving a first molecular sequence, a first molecular property vector, and a first number of generated molecules input by a user; performing three-dimensional conformation conversion processing on the first molecular sequence based on a preset chemoinformatics tool to obtain a corresponding first converted conformation; performing batch molecular generation task processing according to the first converted conformation, the first molecular property vector, the first number of generated molecules, and the molecular conformation generator to obtain a corresponding first batch molecular task report and feedback it to the user. Through the present invention, the molecular design efficiency can be improved, and the exploration ability of unknown molecular structures can be enhanced.