WO2021098256A1 - 一种预测方法及装置、电子设备和存储介质 - Google Patents

一种预测方法及装置、电子设备和存储介质 Download PDF

Info

Publication number
WO2021098256A1
WO2021098256A1 PCT/CN2020/103633 CN2020103633W WO2021098256A1 WO 2021098256 A1 WO2021098256 A1 WO 2021098256A1 CN 2020103633 W CN2020103633 W CN 2020103633W WO 2021098256 A1 WO2021098256 A1 WO 2021098256A1
Authority
WO
WIPO (PCT)
Prior art keywords
matrix
feature
substance
cell
tested
Prior art date
Application number
PCT/CN2020/103633
Other languages
English (en)
French (fr)
Inventor
刘桥
胡志强
Original Assignee
北京市商汤科技开发有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京市商汤科技开发有限公司 filed Critical 北京市商汤科技开发有限公司
Priority to JP2021543171A priority Critical patent/JP2022518283A/ja
Publication of WO2021098256A1 publication Critical patent/WO2021098256A1/zh
Priority to US17/739,541 priority patent/US20220285038A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/40—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for data related to laboratory analysis, e.g. patient specimen analysis
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B25/00—ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression
    • G16B25/10—Gene or protein expression profiling; Expression-ratio estimation or normalisation
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
    • G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
    • G16C20/30—Prediction of properties of chemical compounds, compositions or mixtures
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/10—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
    • G16H40/63—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices for local operation
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G16H70/40—ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
    • G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
    • G16C20/70—Machine learning, data mining or chemometrics

Definitions

  • the embodiments of the present disclosure relate to the field of computer technology, and relate to a prediction method and device, electronic equipment, and storage medium.
  • the embodiments of the present disclosure propose a prediction method and device, electronic equipment, and storage medium.
  • a prediction method including:
  • the result of the reaction of the test substance against the diseased cells is predicted.
  • the determining the material characteristics of the test substance according to the molecular structure of the test substance includes:
  • a structural feature map of the substance to be tested is constructed.
  • the structural feature map includes at least two nodes and a line between each node.
  • the nodes represent the components in the molecular structure.
  • Atoms, the connecting lines represent atomic bonds in the molecular structure;
  • the substance feature of the substance to be tested is determined.
  • the substance feature of the substance to be tested can be extracted, and the extracted substance feature is denser.
  • the accuracy of the test result and the efficiency of obtaining the test result can be improved.
  • the determining the material characteristic of the substance to be tested according to the structural characteristic diagram includes:
  • the first adjacency matrix and the first feature matrix of the substance to be tested are obtained according to the structural feature map, the first adjacency matrix represents the neighbor relationship between the atoms of the substance to be tested, and the first feature matrix Represents the attribute data of each atom of the substance to be tested;
  • the substance feature of the substance to be tested is obtained.
  • the structural characteristics of the substance to be tested can be represented by the first adjacency matrix and the first feature matrix, and the material characteristics of the substance to be tested can be extracted by performing graph convolution processing on the first adjacency matrix and the first feature matrix.
  • the obtaining the material characteristics of the substance to be measured according to the first adjacency matrix and the first characteristic matrix includes:
  • the first adjacency matrix and the complementary matrix of the first adjacency matrix are spliced to obtain a second adjacency matrix whose dimension is the preset input dimension, and the first feature matrix and the first feature Performing splicing processing on the supplementary matrix of the matrix to obtain a second feature matrix whose dimension is the preset input dimension;
  • test method provided by the embodiment of the present disclosure can be suitable for reaction test for substances of any size and structure and target types of diseased cells, and has a strong expansion capability.
  • the first adjacency matrix and the supplementary matrix of the first adjacency matrix do not have an adjacency relationship. Since the atoms of the substance to be tested do not have any adjacency relationship with the atoms in the supplementary matrix, the molecular structure of the substance to be tested will not be affected, and thus the test results of the substance to be tested will not be affected.
  • the first adjacency matrix and the complementary matrix of the first adjacency matrix are spliced to obtain a second adjacency matrix whose dimension is the preset input dimension
  • the The splicing process of the first feature matrix and the supplementary matrix of the first feature matrix to obtain the second feature matrix whose dimension is the preset input dimension includes:
  • the first feature matrix and the supplementary matrix of the first feature matrix are connected to obtain the second feature matrix whose dimension is the preset input dimension.
  • the material characteristics of the substance to be tested can be constructed into input data that meets the test requirements, and the molecular structure of the substance to be tested will not be affected, and thus the test result of the substance to be tested will not be affected.
  • the extraction of at least one cell feature of the diseased cell of the target category is extracted to obtain at least one cell feature of the diseased cell, including at least one of the following:
  • Feature extraction is performed on the deoxyribonucleic acid (Deoxyribo Nucleic Acid, DNA) methylation data of the diseased cell to obtain the epigenetic features of the diseased cell.
  • deoxyribonucleic acid Deoxyribo Nucleic Acid, DNA
  • the predicting the response result of the test substance against the diseased cell based on the substance characteristic and the at least one cell characteristic includes:
  • Convolution processing is performed on the combined features to obtain a predicted response result of the test substance against the diseased cell.
  • the cell characteristics include genomic characteristics, transcriptome characteristics, and epigenetic characteristics, and the material characteristics and the at least one cell characteristic are feature-connected to obtain the connected Combination features, including;
  • the material feature is feature-connected with at least one feature of the genome feature, the transcriptome feature, and the epigenetic feature to obtain a combined feature after connection.
  • the method is implemented by a neural network, and the method further includes: training the neural network through a preset training set, the training set includes multiple sets of sample data, and each set of sample data includes The structural feature map of the sample material, the gene table mutation of the sample diseased cell, the gene expression of the sample diseased cell, the DNA methylation data of the sample diseased cell, and the annotation reaction result of the sample material against the sample diseased cell.
  • the neural network includes a first feature extraction network, a second feature extraction network, and a prediction network.
  • the training of the neural network through a preset training set includes:
  • the second feature extraction network respectively extract the sample genome features corresponding to the gene table mutations of the sample diseased cells, the sample transcriptome features corresponding to the gene expression of the sample diseased cells, and the DNA of the sample diseased cells.
  • the neural network is trained.
  • the neural network used to implement the above prediction method can be trained to extract the material characteristics of the material to be tested based on the structural feature map of the material to be tested, and the extracted material characteristics are more dense, so that when the material characteristics are used for prediction , Can improve the accuracy of test results and the efficiency of obtaining test results.
  • a prediction device including:
  • the first determining part is configured to determine the material characteristics of the test substance according to the molecular structure of the test substance
  • the extraction part is configured to extract at least one cell feature of the diseased cell of the target category to obtain at least one cell feature of the diseased cell;
  • the second determining part is configured to determine the response prediction result of the test substance against the diseased cell based on the substance characteristic and the at least one cell characteristic.
  • an electronic device including: a processor; a memory configured to store executable instructions of the processor; wherein the processor is configured to call instructions stored in the memory, To perform the above method.
  • a computer-readable storage medium having computer program instructions stored thereon, and the computer program instructions implement the foregoing method when executed by a processor.
  • a structural feature map of the substance to be tested can be constructed, and then the material feature of the substance to be tested can be extracted based on the structural feature map, and after extracting at least one cell feature of the target type of diseased cell According to the material characteristics of the test substance and at least one cell characteristic of the diseased cell, the response result of the test substance against the diseased cell can be predicted.
  • the material characteristics of the substance to be tested can be extracted based on the structural feature map of the substance to be tested. Compared with manual extraction of the material characteristics, the extracted material characteristics are denser , Which can improve the accuracy of the reaction test results and the efficiency of obtaining the test results.
  • FIG. 1 shows a schematic flowchart of a prediction method provided by an embodiment of the present disclosure
  • FIG. 2 shows a schematic diagram of a matrix provided by an embodiment of the present disclosure
  • FIG. 3 shows a schematic flowchart of a prediction method provided by an embodiment of the present disclosure
  • FIG. 4 shows a schematic structural diagram of a prediction device provided by an embodiment of the present disclosure
  • FIG. 5 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure
  • FIG. 6 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • FIG. 1 shows a schematic flow chart of a prediction method provided by an embodiment of the present disclosure.
  • the prediction method may be executed by a terminal device or other processing device.
  • the terminal device may be a user equipment (User Equipment, UE), a mobile device, a user terminal, Terminals, cellular phones, cordless phones, personal digital assistants (PDAs), handheld devices, computing devices, in-vehicle devices, wearable devices, etc.
  • Other processing devices can be servers or cloud servers.
  • the prediction method may be implemented by a processor invoking computer-readable instructions stored in a memory.
  • the prediction method may include:
  • the substance to be tested may be a substance with a molecular structure, such as a drug.
  • the molecular structure of the substance to be tested is composed of multiple atoms and atomic bonds between multiple atoms, and the substance characteristics of the substance to be tested can be extracted according to the molecular structure of the substance to be tested.
  • the foregoing determination of the material characteristics of the test substance according to the molecular structure of the test substance may include:
  • the structure feature map including at least two nodes and the connection between each node, the nodes representing atoms in the molecular structure,
  • the connecting lines represent the atomic bonds in the molecular structure
  • the material feature of the substance to be tested is determined.
  • a structural feature map of the substance to be tested can be constructed.
  • the molecular structure of the substance to be tested is composed of at least two atoms and atomic bonds between at least two atoms, then the structure of the substance to be tested
  • the feature graph can include at least two nodes and connections between each node, where nodes can represent atoms in a molecular structure, and connections between nodes can represent atomic bonds between atoms.
  • the feature extraction can be performed through the structure feature map of the substance to be tested to obtain the material feature of the substance to be tested.
  • the convolutional neural network for feature extraction from the structure feature map can be pre-trained, and the convolutional neural network can be used to Perform feature extraction on the structure feature map of the substance to obtain the substance feature of the substance to be tested.
  • the substance feature of the substance to be tested can be extracted based on the structure feature map of the substance to be tested.
  • the material characteristics are also denser, and further predictions based on the material characteristics can improve the accuracy of the test results and the efficiency of obtaining the test results.
  • At least one cell feature of the diseased cell of the target category is extracted to obtain at least one cell feature of the diseased cell.
  • the target category may be a certain cancer or any other types of lesions, which is not limited in the present disclosure.
  • a therapeutic drug B for type A cancer is currently developed, and the response of drug B to cancer cells of type A cancer needs to be tested, then drug B is the substance to be tested, and cancer cells of type A cancer are the target type of diseased cells .
  • a convolutional neural network for feature extraction of diseased cells can be pre-trained, and cell feature extraction can be performed on diseased cells through the convolutional neural network to obtain at least one cell feature of the diseased cell, for example: extracting diseased cells At least one of the genomic characteristics, transcriptome characteristics, and epigenome characteristics of.
  • the response result of the test substance against the diseased cell is predicted based on the substance characteristic and at least one cell characteristic.
  • the prediction operation can be performed based on the material characteristics of the test substance and at least one cell characteristic of the diseased cell to obtain the predicted test substance for the disease.
  • the result of the cell's response can be performed based on the material characteristics of the test substance and at least one cell characteristic of the diseased cell to obtain the predicted test substance for the disease.
  • a convolutional neural network that performs reaction prediction based on material characteristics and at least one cell characteristic may be pre-trained, and the material characteristics of the substance to be tested and at least one cell characteristic of diseased cells can be predicted through the convolutional neural network. Obtain the predicted response result of the test substance against the diseased cells.
  • the foregoing prediction of the response result of the test substance against the diseased cell based on the substance characteristic and at least one cell characteristic may include:
  • Convolution processing is performed on the combined features to obtain the predicted response result of the test substance against the diseased cells.
  • the material feature of the substance to be tested and at least one cell feature can be directly connected to obtain a combined feature, and the combined feature can be expressed as: material feature + cell feature.
  • convolution processing is performed on the combined features.
  • the output of the convolutional neural network can be a probability value between 0 and 1, and the probability value indicates that the substance to be tested has a probability value between 0 and 1.
  • the substance characteristics of the substance to be tested can be determined, and after at least one cell characteristic of the diseased cells of the target category is extracted, the substance characteristics of the substance to be tested and at least one of the diseased cells can be extracted.
  • Cell characteristics predict the response of the test substance to diseased cells.
  • the material characteristics of the test substance can be extracted based on the molecular structure of the test substance.
  • the extracted substance characteristics are also denser.
  • the foregoing determination of the material characteristics of the substance to be tested according to the structural characteristic diagram may include:
  • the first adjacency matrix represents the neighbor relationship between each atom of the substance to be tested, and the first feature matrix represents the attribute data of each atom of the substance to be tested. ;
  • the substance characteristic of the substance to be measured is obtained.
  • the neighboring atoms of each atom of the substance to be tested can be extracted according to the structural feature map, and a first adjacency matrix can be formed according to the neighboring atoms of each atom, and each row of the first adjacency matrix represents the The neighbor relationship between each atom and other atoms, where the neighbor relationship refers to the connection relationship, for example, the first row of the first adjacency matrix indicates whether the first atom of the substance to be tested has a connection relationship with other atoms, If it is, it is represented as 1 in the first adjacency matrix, otherwise it is represented as 0 in the first adjacency matrix.
  • Each atom of the substance to be tested can be extracted according to the structural feature map, and the attribute data of each atom can be obtained.
  • the attribute data of each atom can be queried from the database.
  • the attribute data can include, but is not limited to, the atom type and the heterogeneity of the atom.
  • the first characteristic matrix can be formed according to the attribute data of each atom, and each row of the first characteristic matrix represents the attribute data of each atom of the substance to be tested.
  • the graph convolution processing of the first adjacency matrix and the first feature matrix can be implemented by the following formula (1-1) and formula (1-2):
  • the degree matrix of H represents the convolution result of the first layer graph convolution
  • the diagonal of the degree matrix D represents the number of neighboring atoms of each atom (the neighboring atoms are bonded to the atom)
  • X represents the first feature matrix
  • ⁇ represents the filter parameter of the first layer image convolution.
  • H (l+1) represents the convolution result of the l+1 layer image convolution
  • H (l) represents the convolution result of the l layer image convolution
  • ⁇ (l) represents the filter of the l layer image convolution
  • the parameter, ⁇ () represents the nonlinear activation function.
  • the first adjacency matrix and the first characteristic matrix can be used to express the structural characteristics of the substance to be tested, and the first adjacency matrix and the first characteristic matrix can be subjected to graph convolution processing to extract the substance characteristics of the substance to be tested.
  • obtaining the material characteristics of the substance to be measured according to the first adjacency matrix and the first characteristic matrix may include:
  • the first adjacency matrix and the complementary matrix of the first adjacency matrix are spliced to obtain a second adjacency matrix whose dimension is a preset input dimension
  • the first feature matrix and the first feature matrix are The supplementary matrix is spliced to obtain a second feature matrix whose dimension is the preset input dimension
  • Graph convolution processing is performed on the second adjacency matrix and the second feature matrix to obtain the material feature of the substance to be tested.
  • the aforementioned preset input dimension may be a preset dimension of input data.
  • the preset input dimension may be set to 100.
  • the difference between the dimensions of an adjacency matrix is the dimension of the supplementary matrix of the first adjacency matrix.
  • the dimension of the first adjacency matrix is 20*20, and the dimension of the first feature matrix is 20*75
  • the dimension of the supplementary matrix of the first adjacency matrix can be determined It is 80*80, and the dimension of the supplementary matrix of the first feature matrix is 80*25.
  • the supplementary matrix of the first adjacency matrix can be set as a zero matrix or randomly sampled as an adjacency matrix with any nearest neighbor relationship. After obtaining the first feature matrix, it is necessary to determine the dimension of the supplementary matrix of the first feature matrix according to the dimension of the first feature matrix, and then construct the supplementary matrix of the first feature matrix of this dimension, for example: determine the preset input dimension and the first feature matrix.
  • the difference of the dimension of a characteristic matrix is the dimension of the supplementary matrix of the first characteristic matrix, and the common atoms in the first characteristic matrix are randomly selected, and the supplementary matrix of the first characteristic matrix is constructed by the selected atoms.
  • the dimension of the second adjacency matrix is the preset input dimension*preset Set the input dimension.
  • the first characteristic matrix and the supplementary matrix of the first characteristic matrix can be spliced to obtain the second characteristic matrix.
  • the dimension of the second characteristic matrix is the preset input dimension*atom Feature dimensions. Exemplarily, when the preset input dimension is set to 100 and the atomic feature dimension is 75, it can be determined that the dimension of the second adjacency matrix is 100*100, and the dimension of the second feature matrix is 100*75.
  • the graph convolution processing of the second adjacency matrix and the second feature matrix can be implemented by the following formula (1-3), formula (1-4) and formula (1-5):
  • H (l, ⁇ ) represents the first n (the number of atoms of the substance to be tested ) in the convolution result of the first layer
  • H (l, ⁇ ) represents the convolution result of the first layer divided by H ( l, ⁇ )
  • B represents the first connection matrix
  • D B and Respectively represent the two degree matrices of the row and column of the first connection matrix B
  • X represents the first feature matrix
  • X C represents the supplementary matrix of the first feature matrix
  • ⁇ () represents the nonlinear activation function
  • ⁇ represents the filter parameter of the first layer graph convolution
  • ⁇ (l) represents the lth layer graph convolution
  • the filter parameters When the first connection matrix is zero, that is, the first adjacency matrix and the supplementary matrix of the first adjacency matrix do not have an adja
  • test method provided by the embodiment of the present disclosure can be suitable for reaction test for substances of any size and structure and target types of diseased cells, and has a strong expansion capability.
  • the first adjacency matrix and the supplementary matrix of the first adjacency matrix do not have an adjacency relationship. There is no adjacency relationship between the matrices, which means that the atoms contained in one matrix do not have any connection relationship with the atoms contained in the other matrix.
  • the first adjacency matrix and the supplementary matrix of the first adjacency matrix do not have an adjacency relationship, that is, the atoms of the substance to be measured and the supplementary matrix
  • the atoms do not have any connection relationship, so that the supplementary matrix of the first adjacency matrix can construct the second adjacency matrix of the preset input dimension with the first adjacency matrix, and the supplementary matrix of the first feature matrix can construct the preset second adjacency matrix with the first feature matrix.
  • the second feature matrix of input dimensions because the atoms of the substance to be tested do not have any adjacency with the atoms in the supplementary matrix, it will not affect the molecular structure of the substance to be tested, and thus will not affect the test results of the substance to be tested.
  • the first adjacency matrix and the complementary matrix of the first adjacency matrix are spliced to obtain a second adjacency matrix whose dimension is a preset input dimension
  • the first adjacency matrix is The feature matrix and the supplementary matrix of the first feature matrix are spliced to obtain the second feature matrix whose dimension is the preset input dimension, which may include:
  • a first connection matrix with all 0 elements can be constructed, and the first connection matrix, the first adjacency matrix, and the supplementary matrix of the first adjacency matrix form a second adjacency matrix.
  • the first adjacency matrix The connection matrix connects the first adjacency matrix and the supplementary matrix of the first adjacency matrix, so that the first adjacency matrix and the supplementary matrix of the first adjacency matrix do not have an adjacency relationship.
  • FIG. 2 shows a schematic diagram of a matrix provided by an embodiment of the present disclosure. In the second adjacency matrix with a dimension of 100*100 as shown in FIG.
  • the first adjacency matrix with a dimension of 20*20 is located in the second adjacency matrix.
  • the upper left position of the matrix, the supplementary matrix of the first adjacency matrix with a dimension of 80*80 is located at the lower right position of the second adjacency matrix, the dimension below the first adjacency matrix and the left position of the supplementary matrix of the first adjacency matrix is
  • the first connection matrix of 20*80, located on the right side of the first adjacency matrix and above the supplementary matrix of the first adjacency matrix, is the first connection matrix with a dimension of 80*20.
  • any supplementary matrix that makes the first adjacency matrix and the first adjacency matrix The matrix does not have an adjacency relationship.
  • the first adjacency matrix with a dimension of 20*20 is located in the lower right position of the second adjacency matrix
  • the supplementary matrix of the first adjacency matrix with a dimension of 80*80 is located in the second adjacency matrix.
  • the upper left position of the adjacency matrix is located above the first adjacency matrix and the right position of the supplementary matrix of the first adjacency matrix is the first connection matrix with a dimension of 80*20, which is located at the left position of the first adjacency matrix and the position of the first adjacency matrix
  • the position below the supplementary matrix is the first connection matrix with a dimension of 20*80.
  • the present disclosure does not specifically limit the manner in which the first connection matrix connects the first adjacency matrix and the supplementary matrix of the first adjacency matrix.
  • connection mode of the first feature matrix and the complementary matrix of the first feature matrix can be determined according to the connection mode of the first adjacency matrix and the complementary matrix of the first adjacency matrix, for example: refer to the first adjacency matrix and the first adjacency matrix in FIG. 2
  • a connection manner of the supplementary matrix of the adjacency matrix may be that the first feature matrix is located at the upper position and the supplementary matrix of the first feature matrix is located at the lower position.
  • connection mode of the first adjacency matrix and the supplementary matrix of the first adjacency matrix is that the first adjacency matrix is located at the lower right position of the second adjacency matrix, and the supplementary matrix of the first adjacency matrix is located at the upper left of the second adjacency matrix In the case of position, the first feature matrix in the second feature matrix is located at the lower position, and the supplementary matrix of the first feature matrix is located at the upper position.
  • the material characteristics of the substance to be tested can be configured to meet the requirements of the input data of the reaction test, and the molecular structure of the substance to be tested will not be affected, and thus the reaction test result of the substance to be tested will not be affected.
  • At least one cell feature extraction is performed on the diseased cells of the target category to obtain at least one cell feature of the diseased cells, including at least one of the following:
  • Feature extraction is performed on the deoxyribonucleic acid (Deoxyribo Nucleic Acid, DNA) methylation data of the diseased cell to obtain the epigenetic features of the diseased cell.
  • deoxyribonucleic acid Deoxyribo Nucleic Acid, DNA
  • the gene table mutation, gene expression, and DNA methylation data of the diseased cell can be obtained.
  • the acquisition process can be extracted by using related technologies, or directly from a database. Inquiries, this disclosure will not repeat the process here.
  • the gene table mutation, gene expression, and DNA methylation data of the diseased cell can be preprocessed into a fixed-dimensional vector.
  • the gene table mutation of the diseased cell can be preprocessed into a 34673-dimensional vector
  • the diseased cell’s gene table mutation can be preprocessed into a 34673-dimensional vector.
  • the gene expression is preprocessed into a 697-dimensional vector
  • the DNA methylation data of diseased cells is preprocessed into a 808-dimensional vector.
  • the convolutional neural network for extracting genomic features is pre-trained, and the convolutional neural network is used to preprocess the
  • the gene table mutations of the diseased cells are feature extracted to obtain the genomic features of the diseased cells; the convolutional neural network for extracting transcriptome features can be pre-trained, and the gene expression of the preprocessed diseased cells can be performed through the convolutional neural network.
  • the convolutional neural network for extracting epigenetic features can be pre-trained, and the preprocessed DNA methylation data can be extracted through the convolutional neural network to obtain the
  • the epigenetic characteristics of diseased cells where the dimensions of genomic characteristics, transcriptome characteristics, and epigenetic characteristics are the same as the dimensions of material characteristics.
  • the convolutional neural network used to extract cell features is a multi-modal sub-neural network.
  • the aforementioned cell characteristics may include genomic characteristics, transcriptome characteristics, and epigenetic characteristics. After the material characteristics and the at least one cell characteristic are feature-connected, the connection is obtained.
  • the following combination features include:
  • the material feature and at least one of the genomic feature, the transcriptome feature, and the epigenetic feature are feature-connected, a combined feature after the connection is obtained.
  • the combined feature can be obtained by connecting the material feature of the substance to be tested with the genome feature, the transcriptome feature, and the epigenetic feature, and the combined feature can be expressed as: material feature+genomic feature +Transcriptome features+Epigenetic features.
  • material feature+genomic feature +Transcriptome features+Epigenetic features By performing convolution processing on the combined features, the response prediction result of the test substance against the diseased cells can be obtained.
  • Fig. 3 shows a schematic flow chart of the prediction method provided by an embodiment of the present disclosure.
  • the test substance is a drug and the diseased cell is a cancer cell.
  • the cell feature network includes: genome feature extraction network, transcriptome feature extraction network, and genetic group feature extraction network ,
  • the gene table mutations can be feature extracted through the genome feature extraction network to obtain the genome features of cancer cells, and the gene expression can be feature extracted through the transcriptome feature extraction network to obtain the transcriptome features of the cancer cells, and the epigenetic features can be extracted through epigenetic feature extraction
  • the network performs feature extraction on DNA methylation data to obtain the epigenetic features of cancer cells. After the material characteristics of the drug to be tested are pooled, the pooled material characteristics are connected with the genome characteristics, transcriptome characteristics, and epigenetic characteristics to obtain the combined characteristics, and the combined characteristics are convolved. , Obtain the predicted response result of the test drug to the cancer cell (the response result indicates whether the test drug is sensitive or inhibited to the cancer cell).
  • the above method is implemented by a neural network, and the method further includes: training the neural network through a preset training set, the training set includes multiple sets of sample data, and each set of sample data includes samples The structural feature map of the substance, the mutation of the gene table of the sample diseased cell, the gene expression of the sample diseased cell, the DNA methylation data of the sample diseased cell, and the labeling reaction result of the sample material against the sample diseased cell.
  • the neural network is a consensus graph convolutional neural network.
  • the neural network may include a first feature extraction network, a second feature extraction network, and a prediction network.
  • the method for training the neural network through a preset training set may include:
  • the second feature extraction network respectively extract the sample genome features corresponding to the gene table mutations of the sample diseased cells, the sample transcriptome features corresponding to the gene expression of the sample diseased cells, and the DNA methyl groups of the sample diseased cells.
  • the neural network is trained.
  • the first feature extraction network may be used to perform feature extraction on the structural feature map of the sample substance to obtain the sample substance characteristics of the sample substance.
  • the second feature extraction network may include a first sub-network, a second sub-network, and a third sub-network.
  • the first sub-network can perform feature extraction on the gene table mutations of the sample diseased cells to obtain the sample genome features
  • the second sub-network Feature extraction is performed on the gene expression of the sample diseased cells to obtain the sample transcriptome feature
  • the DNA methylation data of the sample diseased cell is feature extracted through the third sub-network to obtain the sample epigenetic group feature.
  • embodiments of the present disclosure also provide prediction devices, electronic equipment, computer-readable storage media, and programs, all of which can be used to implement any of the prediction methods provided in the embodiments of the present disclosure.
  • prediction devices electronic equipment, computer-readable storage media, and programs, all of which can be used to implement any of the prediction methods provided in the embodiments of the present disclosure.
  • FIG. 4 shows a schematic structural diagram of a prediction device provided by an embodiment of the present disclosure.
  • the prediction device may include:
  • the first determining part 401 may be configured to determine the material characteristics of the test substance according to the molecular structure of the test substance;
  • the extraction part 402 may be configured to extract at least one cell feature of the diseased cell of the target category to obtain at least one cell feature of the diseased cell;
  • the second determining part 403 may be configured to predict the response result of the test substance against the diseased cell based on the substance characteristic and the at least one cell characteristic.
  • a structural feature map of the substance to be tested can be constructed, and then the material feature of the substance to be tested can be extracted based on the structural feature map, and after extracting at least one cell feature of the target type of diseased cell
  • the response result of the test substance against the diseased cell can be predicted.
  • the material characteristics of the test substance can be extracted based on the structure feature map of the test substance. Compared with manual extraction of the substance characteristics, the extracted material characteristics are denser, thereby improving the accuracy of the test results. And the efficiency of obtaining test results.
  • the first determining part 401 is configured to:
  • the structure feature map including at least two nodes and the connection between each node, the nodes representing atoms in the molecular structure,
  • the connecting lines represent the atomic bonds in the molecular structure
  • the material feature of the substance to be tested is determined.
  • the first determining part 401 is further configured to:
  • the first adjacency matrix and the first feature matrix of the substance to be tested are obtained according to the structural feature map, the first adjacency matrix represents the neighbor relationship of each atom of the substance to be tested, and the first feature matrix represents the State the attribute data of each atom of the substance to be tested;
  • the material feature of the substance to be tested is obtained.
  • the first determining part 401 is further configured to:
  • the first adjacency matrix and the complementary matrix of the first adjacency matrix are spliced to obtain a second adjacency matrix whose dimension is the preset input dimension, and the first feature matrix and the first feature Performing splicing processing on the supplementary matrix of the matrix to obtain a second feature matrix whose dimension is the preset input dimension;
  • the first adjacency matrix and the supplementary matrix of the first adjacency matrix do not have an adjacency relationship.
  • the first determining part 401 is further configured to:
  • the first feature matrix and the supplementary matrix of the first feature matrix are connected to obtain a second feature matrix whose dimension is the preset input dimension.
  • the extraction part 402 is configured as at least one of the following:
  • Feature extraction is performed on the DNA methylation data of the diseased cell to obtain the epigenetic group feature of the diseased cell.
  • the second determining part 403 is configured to:
  • Convolution processing is performed on the combined features to obtain a response result of the test substance against the diseased cell.
  • the cell characteristics include genomic characteristics, transcriptome characteristics, and epigenetic characteristics
  • the second determining part 403 is further configured to:
  • the material feature is feature-connected with at least one feature of the genome feature, the transcriptome feature, and the epigenetic feature to obtain a combined feature after connection.
  • the device is implemented through a neural network, and the device further includes:
  • the training part is configured to train the neural network through a preset training set, the training set includes multiple sets of sample data, each set of sample data includes a structural feature map of the sample substance, the gene table mutation of the sample diseased cell, and the sample disease The gene expression of the cell, the DNA methylation data of the sample diseased cell, and the labeling reaction result of the sample material against the sample diseased cell.
  • the neural network includes a first feature extraction network, a second feature extraction network, and a prediction network
  • the training part is further configured to:
  • the second feature extraction network respectively extract the sample genome features corresponding to the gene table mutations of the sample diseased cells, the sample transcriptome features corresponding to the gene expression of the sample diseased cells, and the DNA of the sample diseased cells.
  • the neural network is trained.
  • the functions or parts included in the device provided in the embodiments of the present disclosure may be configured to execute the methods described in the above method embodiments.
  • the functions or parts included in the device provided in the embodiments of the present disclosure may be configured to execute the methods described in the above method embodiments.
  • parts may be parts of circuits, parts of processors, parts of programs or software, etc., of course, may also be units, modules, or non-modular.
  • the embodiments of the present disclosure also provide a computer-readable storage medium on which computer program instructions are stored, and the computer program instructions implement the above-mentioned method when executed by a processor.
  • the computer-readable storage medium may be a non-volatile computer-readable storage medium.
  • An embodiment of the present disclosure also proposes an electronic device, including: a processor; a memory configured to store executable instructions of the processor; wherein the processor is configured to call instructions stored in the memory to execute the above method.
  • the embodiments of the present disclosure also provide a computer program product, which includes computer-readable code.
  • the processor in the device executes a computer program that is configured to implement the prediction method provided by any of the above embodiments. instruction.
  • the embodiments of the present disclosure also provide another computer program product, which is configured to store computer-readable instructions, and when the instructions are executed, the computer executes the operation of the prediction method provided in any of the foregoing embodiments.
  • the electronic device can be provided as a terminal, server or other form of device.
  • FIG. 5 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • the electronic device 800 may be a mobile phone, a computer, a digital broadcasting terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and other terminals.
  • the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, and a sensor component 814 , And communication component 816.
  • the processing component 802 generally controls the overall operations of the electronic device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations.
  • the processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the foregoing method.
  • the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components.
  • the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.
  • the memory 804 is configured to store various types of data to support operations in the electronic device 800. Examples of such data include instructions for any application or method operating on the electronic device 800, contact data, phone book data, messages, pictures, videos, and the like.
  • the memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable and Programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read-only memory
  • EPROM erasable and Programmable read only memory
  • PROM programmable read only memory
  • ROM read only memory
  • magnetic memory flash memory
  • flash memory magnetic disk or optical disk.
  • the power supply component 806 provides power for various components of the electronic device 800.
  • the power supply component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the electronic device 800.
  • the multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user.
  • the touch panel includes one or more touch sensors to sense touch, sliding, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure related to the touch or slide operation.
  • the multimedia component 808 includes a front camera and/or a rear camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capabilities.
  • the audio component 810 is configured to output and/or input audio signals.
  • the audio component 810 includes a microphone (MIC), and when the electronic device 800 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal.
  • the received audio signal may be further stored in the memory 804 or transmitted via the communication component 816.
  • the audio component 810 further includes a speaker for outputting audio signals.
  • the I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module.
  • the above-mentioned peripheral interface module may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: home button, volume button, start button, and lock button.
  • the sensor component 814 includes one or more sensors for providing the electronic device 800 with various aspects of state evaluation.
  • the sensor component 814 can detect the on/off status of the electronic device 800 and the relative positioning of the components.
  • the component is the display and the keypad of the electronic device 800.
  • the sensor component 814 can also detect the electronic device 800 or the electronic device 800.
  • the position of the component changes, the presence or absence of contact between the user and the electronic device 800, the orientation or acceleration/deceleration of the electronic device 800, and the temperature change of the electronic device 800.
  • the sensor component 814 may include a proximity sensor configured to detect the presence of nearby objects when there is no physical contact.
  • the sensor component 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications.
  • the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
  • the communication component 816 is configured to facilitate wired or wireless communication between the electronic device 800 and other devices.
  • the electronic device 800 can access a wireless network based on a communication standard, such as WiFi, 2G, or 3G, or a combination thereof.
  • the communication component 816 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communication component 816 further includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field-available A programmable gate array (FPGA), controller, microcontroller, microprocessor, or other electronic component implementation is configured to perform the above-mentioned methods.
  • ASIC application-specific integrated circuits
  • DSP digital signal processors
  • DSPD digital signal processing devices
  • PLD programmable logic devices
  • FPGA field-available A programmable gate array
  • controller microcontroller, microprocessor, or other electronic component implementation is configured to perform the above-mentioned methods.
  • a non-volatile computer-readable storage medium such as the memory 804 including computer program instructions, which can be executed by the processor 820 of the electronic device 800 to complete the foregoing method.
  • FIG. 6 shows a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • the electronic device 1900 may be provided as a server. 6
  • the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by the memory 1932, configured to store instructions executable by the processing component 1922, such as application programs.
  • the application program stored in the memory 1932 may include one or more parts each corresponding to a set of instructions.
  • the processing component 1922 is configured to execute instructions to perform the aforementioned prediction method.
  • the electronic device 1900 may also include a power supply component 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to the network, and an input output (I/O) interface 1958 .
  • the electronic device 1900 can operate based on an operating system stored in the memory 1932, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or the like.
  • a non-volatile computer-readable storage medium is also provided, such as the memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the electronic device 1900 to complete the foregoing method.
  • the present disclosure may be a system, method and/or computer program product.
  • the computer program product may include a computer-readable storage medium loaded with computer-readable program instructions for enabling a processor to implement various aspects of the present disclosure.
  • the computer-readable storage medium may be a tangible device that can hold and store instructions used by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Non-exhaustive list of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as a printer with instructions stored thereon
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory flash memory
  • SRAM static random access memory
  • CD-ROM compact disk read-only memory
  • DVD digital versatile disk
  • memory stick floppy disk
  • mechanical encoding device such as a printer with instructions stored thereon
  • the computer-readable storage medium used here is not interpreted as the instantaneous signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (for example, light pulses through fiber optic cables), or through wires Transmission of electrical signals.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing/processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device .
  • the computer program instructions used to perform the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or in one or more programming languages.
  • Source code or object code written in any combination, the programming language includes object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as "C" language or similar programming languages.
  • Computer-readable program instructions can be executed entirely on the user's computer, partly on the user's computer, executed as a stand-alone software package, partly on the user's computer and partly executed on a remote computer, or entirely on the remote computer or server carried out.
  • the remote computer can be connected to the user's computer through any kind of network-including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect to the user's computer) connection).
  • LAN local area network
  • WAN wide area network
  • an electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), can be customized by using the status information of the computer-readable program instructions.
  • FPGA field programmable gate array
  • PDA programmable logic array
  • the computer-readable program instructions are executed to realize various aspects of the present disclosure.
  • These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine that makes these instructions when executed by the processor of the computer or other programmable data processing device , A device that implements the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams is produced. It is also possible to store these computer-readable program instructions in a computer-readable storage medium. These instructions make computers, programmable data processing apparatuses, and/or other devices work in a specific manner. Thus, the computer-readable medium storing the instructions includes An article of manufacture, which includes instructions for implementing various aspects of the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams.
  • each block in the flowchart or block diagram may represent a module, program segment, or part of an instruction, and the module, program segment, or part of an instruction contains one or more components for realizing the specified logical function.
  • Executable instructions may also occur in a different order than the order marked in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, or they can sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart can be implemented by a dedicated hardware-based system that performs the specified functions or actions Or it can be realized by a combination of dedicated hardware and computer instructions.
  • the computer program product can be specifically implemented by hardware, software, or a combination thereof.
  • the computer program product is specifically embodied as a computer storage medium.
  • the computer program product is specifically embodied as a software product, such as a software development kit (SDK), etc. Wait.
  • SDK software development kit
  • the embodiments of the present disclosure determine the material characteristics of the test substance according to the molecular structure of the test substance, and after extracting at least one cell characteristic of the diseased cells of the target type, according to the substance characteristics of the test substance and at least one of the diseased cells Cell characteristics, predict the response of the test substance to diseased cells.
  • the material characteristics of the substance to be tested can be extracted based on the structural feature map of the substance to be tested. Compared with manual extraction of the material characteristics, the extracted material characteristics are denser , Which can further improve the accuracy of test results and the efficiency of obtaining test results.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Medicinal Chemistry (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Genetics & Genomics (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Pharmacology & Pharmacy (AREA)
  • Toxicology (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Biotechnology (AREA)
  • Analytical Chemistry (AREA)
  • Business, Economics & Management (AREA)
  • General Business, Economics & Management (AREA)
  • Crystallography & Structural Chemistry (AREA)
  • Proteomics, Peptides & Aminoacids (AREA)
  • Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)

Abstract

一种预测方法及装置、电子设备和存储介质，所述方法包括：根据待测物质的分子结构，确定待测物质的物质特征(S11)；提取目标类别的病变细胞的至少一项细胞特征，得到所述病变细胞的至少一项细胞特征(S12)；根据所述物质特征以及所述至少一项细胞特征，预测所述待测物质针对所述病变细胞的反应结果(S13)。

Description

一种预测方法及装置、电子设备和存储介质
相关申请的交叉引用
本申请基于申请号为201911125921.X、申请日为2019年11月18日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本公开实施例涉及计算机技术领域，涉及一种预测方法及装置、电子设备和存储介质。
背景技术
由于药物疗效的不确定性和癌症患者的异质性，精准测试药物对癌细胞是否有抑制作用至关重要。
相关技术中通常基于人工手动提取得到的药物特征(例如：分子指纹)与癌细胞的单种组学数据提取的癌细胞特征进行机器学习，得到药物对于该种癌细胞的抑制效果，由于手动提取的药物特征往往比较稀疏，故最终得到的抑制效果精准度较低且计算过程较为低效。
发明内容
本公开实施例提出了一种预测方法及装置、电子设备和存储介质。
根据本公开实施例的一方面，提供了一种预测方法，包括：
根据待测物质的分子结构，确定待测物质的物质特征；
提取目标类别的病变细胞的至少一项细胞特征，得到所述病变细胞的至少一项细胞特征；
根据所述物质特征和所述至少一项细胞特征，预测所述待测物质针对所述病变细胞的反应结果。
在一种可能的实现方式中，所述根据待测物质的分子结构，确定待测物质的物质特征，包括：
根据所述待测物质的分子结构，构建所述待测物质的结构特征图，所述结构特征图包括至少两个节点及各节点之间的连线，所述节点表示所述分子结构中的原子，所述连线表示所述分子结构中的原子键；
根据所述结构特征图，确定所述待测物质的所述物质特征。
这样，基于待测物质的结构特征图可以提取待测物质的物质特征，提取的物质特征更为稠密，进一步的通过该物质特征进行预测时，可以提高测试结果的精度及获得测试结果的效率。
在一种可能的实现方式中，所述根据所述结构特征图，确定所述待测物质的所述物质特征，包括：
根据所述结构特征图得到所述待测物质的第一邻接矩阵及第一特征矩阵，所述第一邻接矩阵表示所述待测物质的各原子之间的近邻关系，所述第一特征矩阵表示所述待测物质的各原子的属性数据；
根据所述第一邻接矩阵及所述第一特征矩阵，得到所述待测物质的所述物质特征。
这样，可以通过第一邻接矩阵及第一特征矩阵来表示待测物质的结构特征，进而可以通过对第一邻接矩阵及第一特征矩阵进行图卷积处理，可以提取到待测物质的物质特征。
在一种可能的实现方式中，所述根据所述第一邻接矩阵及所述第一特征矩阵，得到所述待测物质的物质特征，包括：
根据预设输入维度及所述第一邻接矩阵的维度，构建所述第一邻接矩阵的补充矩阵，及根据所述预设输入维度及所述第一特征矩阵的维度，构建所述第一特征矩阵的补充矩阵；
将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二特征矩阵；
对所述第二邻接矩阵及所述第二特征矩阵进行图卷积处理，得到所述待测物质的所述物质特征。
这样一来，本公开实施例提供的测试方法可以适用于针对任意大小、结构的物质和目标类别的病变细胞进行反应测试，有较强的扩展能力。
在一种可能的实现方式中，在所述第二邻接矩阵中，所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵不具有邻接关系。由于待测物质的原子与补充矩阵中的原子不具有任何邻接关系，故不会对待测物质的分子结构产生影响，进而不会对待测物质的测试结果产生影响。
在一种可能的实现方式中，所述将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二特征矩阵，包括：
根据所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵构建第一连接矩阵；
通过所述第一连接矩阵，将所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵进行连接，得到维度为所述预设输入维度的所述第二邻接矩阵；
将所述第一特征矩阵与所述第一特征矩阵的补充矩阵进行连接，得到维度为所述预设输入维度的所述第二特征矩阵。
这样一来，即可以将待测物质的物质特征构造成满足测试要求的输入数据，且又不会对待测物质的分子结构产生影响，进而不会对待测物质的测试结果产生影响。
在一种可能的实现方式中，所述提取目标类别的病变细胞的至少一项细 胞特征提取，得到所述病变细胞的至少一项细胞特征，包括以下至少一项：
对所述病变细胞的基因表突变进行特征提取，得到所述病变细胞的基因组特征；
对所述病变细胞的基因表达进行特征提取，得到所述病变细胞的转录组特征；
对所述病变细胞的脱氧核糖核酸(DeoxyriboNucleic Acid，DNA)甲基化数据进行特征提取，得到所述病变细胞的表观遗传组特征。
这样一来，可以多模态的学习病变细胞的多种细胞特征，根据充分的细胞特征进行反应预测，可以提高预测结果的精准度。
在一种可能的实现方式中，所述根据所述物质特征以及所述至少一项细胞特征，预测所述待测物质针对所述病变细胞的反应结果，包括：
将所述物质特征及所述至少一项细胞特征进行特征连接，得到连接后的组合特征；
对所述组合特征进行卷积处理，得到预测的所述待测物质针对所述病变细胞的反应结果。
这样，基于待测物质的分子结构提取待测物质更为稠密的物质特征，而且连接至少一项细胞特征，可以提高测试结果的精度及获得测试结果的计算效率。
在一种可能的实现方式中，所述细胞特征包括基因组特征、转录组特征和表观遗传组特征，所述将所述物质特征及所述至少一项细胞特征进行特征连接，得到连接后的组合特征，包括；
将所述物质特征与所述基因组特征、所述转录组特征和所述表观遗传组特征中的至少一项特征进行特征连接，得到连接后的组合特征。
这样一来，可以多模态的学习病变细胞的多种细胞特征，根据充分的细胞特征进行反应预测，可以提高预测结果的精准度。
在一种可能的实现方式中，所述方法通过神经网络实现，所述方法还包括：通过预设的训练集训练所述神经网络，所述训练集包括多组样本数据，每组样本数据包括样本物质的结构特征图、样本病变细胞的基因表突变、样本病变细胞的基因表达、样本病变细胞的DNA甲基化数据、及样本物质针对所述样本病变细胞的标注反应结果。
在一种可能的实现方式中，所述神经网络包括第一特征提取网络、第二特征提取网络，及预测网络，所述通过预设的训练集训练所述神经网络，包括：
通过所述第一特征提取网络，对所述样本物质的结构特征图进行特征提取，得到所述样本物质的样本物质特征；
通过所述第二特征提取网络，分别提取所述样本病变细胞的基因表突变对应的样本基因组特征、所述样本病变细胞的基因表达对应的样本转录组特征、及所述样本病变细胞的DNA甲基化数据对应的样本表观遗传组特征；
通过所述预测网络，对连接后的样本物质特征、样本基因组特征、样本 转录组特征及样本表观遗传组特征进行卷积处理，得到样本物质对所述样本病变细胞的反应结果；
根据所述反应结果及所述标注反应结果，确定所述神经网络的预测损失；
根据所述预测损失，训练所述神经网络。
这样一来，可以训练用于实现上述预测方法的神经网络，以基于待测物质的结构特征图可以提取待测物质的物质特征，提取的物质特征更为稠密，从而通过该物质特征进行预测时，可以提高测试结果的精度及获得测试结果的效率。
根据本公开的一方面，提供了一种预测装置，包括：
第一确定部分，被配置为根据待测物质的分子结构，确定待测物质的物质特征；
提取部分，被配置为对目标类别的病变细胞进行至少一项细胞特征提取，得到所述病变细胞的至少一项细胞特征；
第二确定部分，被配置为根据所述物质特征以及所述至少一项细胞特征，确定所述待测物质针对所述病变细胞的反应预测结果。
根据本公开实施例的一方面，提供了一种电子设备，包括：处理器；被配置为存储处理器可执行指令的存储器；其中，所述处理器被配置为调用所述存储器存储的指令，以执行上述方法。
根据本公开实施例的一方面，提供了一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。
这样，根据待测物质的分子结构，可以构建待测物质的结构特征图，进而基于该结构特征图可以提取待测物质的物质特征，并在提取目标类别的病变细胞的至少一项细胞特征后，可以根据待测物质的物质特征及病变细胞的至少一项细胞特征，预测待测物质针对病变细胞的反应结果。根据本公开实施例提供的预测方法及装置、电子设备和存储介质，可以基于待测物质的结构特征图提取待测物质的物质特征，相比于人工提取物质特征，提取的物质特征更为稠密，从而可以提高反应测试结果的精度及获得测试结果的效率。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，而非限制本公开。根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分，这些附图示出了符合本公开的实施例，并与说明书一起用于说明本公开的技术方案。
图1示出本公开实施例提供的预测方法的流程示意图；
图2示出本公开实施例提供的矩阵示意图；
图3示出本公开实施例提供的预测方法的流程示意图；
图4示出本公开实施例提供的预测装置的结构示意图；
图5示出本公开实施例提供的一种电子设备的结构示意图；
图6示出本公开实施例提供的一种电子设备的结构示意图。
具体实施方式
以下将参考附图详细说明本公开的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。
本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中术语“至少一种”表示多种中的任意一种或多种中的至少两种的任意组合，例如，包括A、B、C中的至少一种，可以表示包括从A、B和C构成的集合中选择的任意一个或多个元素。
另外，为了更好地说明本公开，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本公开同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本公开的主旨。
图1示出本公开实施例提供的预测方法的流程示意图，该预测方法可以由终端设备或其它处理设备执行，其中，终端设备可以为用户设备(User Equipment，UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字处理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、可穿戴设备等。其它处理设备可为服务器或云端服务器等。在一些可能的实现方式中，该预测方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。
如图1所示，所述预测方法可以包括：
在S11中，根据待测物质的分子结构，确定待测物质的物质特征。
举例来说，待测物质可以为具有分子结构的物质，例如：药物。待测试物质的分子结构由多个原子及多个原子间的原子键构成，根据该待测物质的分子结构可以提取待测物质的物质特征。
在一种可能的实现方式中，上述根据待测物质的分子结构，确定待测物质的物质特征，可以包括：
根据待测物质的分子结构，构建所述待测物质的结构特征图，所述结构特征图包括至少两个节点及各节点之间的连线，所述节点表示所述分子结构中的原子，所述连线表示所述分子结构中的原子键；
根据所述结构特征图，确定所述待测物质的物质特征。
举例来说，根据待测物质的分子结构，可以构建待测物质的结构特征图，待测物质的分子结构由至少两个原子及至少两个原子间的原子键构成，则待测物质的结构特征图中可以包括至少两个节点及各节点之间的连线，其中， 节点可以表示分子结构中的原子，节点之间的连线可以表示原子之间的原子键。
可以通过待测物质的结构特征图进行特征提取，得到待测物质的物质特征，示例性的，可以预训练对结构特征图进行特征提取的卷积神经网络，通过该卷积神经网络可以对待测物质的结构特征图进行特征提取，得到待测物质的物质特征，这样一来，基于待测物质的结构特征图可以提取待测物质的物质特征，同时，相比于人工提取物质特征，提取的物质特征也更为稠密，进一步的通过该物质特征进行预测时，可以提高测试结果的精度和获得测试结果的效率。
在S12中，提取目标类别的病变细胞的至少一项细胞特征，得到病变细胞的至少一项细胞特征。
举例来说，目标类别可以为某种癌或者任意其他类别的病变，本公开对此不作限定。示例性的，目前研发了针对A类型癌症的治疗药物B，需要测试药物B对A类型癌症的癌细胞的反应，则药物B为待测物质，A类型癌症的癌细胞为目标类别的病变细胞。
示例性的，可以预训练对病变细胞进行特征提取的卷积神经网络，通过该卷积神经网络可以对病变细胞进行细胞特征提取，得到该病变细胞的至少一项细胞特征，例如：提取病变细胞的基因组特征、转录组特征和表观基因组特征中的至少一项特征。
在S13中，根据物质特征以及至少一项细胞特征，预测待测物质针对病变细胞的反应结果。
在得到待测物质的物质特征及病变细胞的至少一项细胞特征后，可以根据待测物质的物质特征及病变细胞的至少一项细胞特征进行预测操作，得到预测的待测物质针对所述病变细胞的反应结果。
示例性的，可以预训练根据物质特征及至少一项细胞特征进行反应预测的卷积神经网络，通过该卷积神经网络对待测物质的物质特征及病变细胞的至少一项细胞特征进行预测操作，得到预测的待测物质针对病变细胞的反应结果。
在一种可能的实现方式中，上述根据物质特征以及至少一项细胞特征，预测待测物质针对病变细胞的反应结果，可以包括：
将物质特征及至少一项细胞特征进行特征连接，得到组合特征；
对组合特征进行卷积处理，得到预测的待测物质针对病变细胞的反应结果。
举例来说，可以将待测物质的物质特征及至少一项细胞特征直接相连后，得到组合特征，该组合特征可以表示为：物质特征+细胞特征。通过预训练的进行反应测试的卷积神经网络，对该组合特征进行卷积处理，该卷积神经网络的输出可以为0～1之间的概率值，该概率值表示该待测物质对该病变细胞起抑制作用的概率。
这样，根据待测物质的分子结构，可以确定待测物质的物质特征，并在 提取目标类别的病变细胞的至少一项细胞特征后，可以根据待测物质的物质特征及病变细胞的至少一项细胞特征，预测待测物质针对病变细胞的反应结果。根据本公开实施例提供的预测方法，可以基于待测物质的分子结构提取待测物质的物质特征，同时，相比于人工提取物质特征，提取的物质特征也更为稠密，当采用提取的物质特征进行反应结果的预测时，可以提高反应结果的测试精度及获得测试结果的效率。
在一种可能的实现方式中，上述根据结构特征图，确定待测物质的物质特征，可以包括：
根据结构特征图得到待测物质的第一邻接矩阵及第一特征矩阵，第一邻接矩阵表示待测物质的各原子之间的近邻关系，第一特征矩阵表示待测物质的各原子的属性数据；
根据待测物质的第一邻接矩阵及第一特征矩阵，得到待测物质的物质特征。
举例来说，可以根据结构特征图提取待测物质的每个原子的相邻原子，并根据每个原子的相邻原子组成第一邻接矩阵，该第一邻接矩阵的每一行表示待测物质的每个原子与其他原子之间的近邻关系，其中，该近邻关系指连接关系，例如，第一邻接矩阵的第一行，表示该待测物质的第一个原子与其他原子是否存在连接关系，若是，则在第一邻接矩阵中表示为1，否则在第一邻接矩阵中表示为0。可以根据结构特征图提取待测物质的每个原子，并获取每个原子的属性数据，例如：从数据库中查询每个原子的属性数据，该属性数据可以包括但不限于原子类型、原子的杂化程度等化学性质，根据每个原子的属性数据可以组成第一特征矩阵，该第一特征矩阵的每一行表示待测物质的每个原子的属性数据。通过对第一邻接矩阵与第一特征矩阵进行图卷积处理，可以提取到待测物质的物质特征。
第一邻接矩阵与第一特征矩阵的图卷积处理可以通过以下公式(1-1)和公式(1-2)实现：
Figure PCTCN2020103633-appb-000001
Figure PCTCN2020103633-appb-000002
其中，
Figure PCTCN2020103633-appb-000003
表示
Figure PCTCN2020103633-appb-000004
的度矩阵，H表示第一层图卷积的卷积结果，
Figure PCTCN2020103633-appb-000005
表示归一化后的度矩阵D，度矩阵D的对角线表示每一个原子的相邻原子的个数(与该原子存在键连接的即为相邻原子)，
Figure PCTCN2020103633-appb-000006
表示归一化后的第一邻接矩阵，X表示第一特征矩阵，Θ表示第一层图卷积的滤波器参数。H (l+1)表示第l+1层图卷积的卷积结果，H (l)表示第l层图卷积的卷积结果，Θ (l)表示第l层图卷积的滤波器参数，σ()表示非线性激活函数。
这样，可以通过第一邻接矩阵及第一特征矩阵来表示待测物质的结构特 征，进而可以通过对第一邻接矩阵及第一特征矩阵进行图卷积处理，提取到待测物质的物质特征。
在一种可能的实现方式中，上述根据所述第一邻接矩阵及所述第一特征矩阵，得到所述待测物质的物质特征，可以包括：
根据预设的输入维度及所述第一邻接矩阵的维度，构建所述第一邻接矩阵的补充矩阵，及根据预设的输入维度及所述第一特征矩阵的维度，构建所述第一特征矩阵的补充矩阵；
将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为预设输入维度的第二特征矩阵；
对所述第二邻接矩阵及所述第二特征矩阵进行图卷积处理，得到所述待测物质的物质特征。
举例来说，上述预设的输入维度可以为预设的输入数据的维度大小，例如：预设的输入维度可以设置为100。在获得第一邻接矩阵后，需要根据第一邻接矩阵的维度确定第一邻接矩阵的补充矩阵的维度，进而构建该维度的第一邻接矩阵的补充矩阵，例如：确定预设的输入维度与第一邻接矩阵的维度的差值为第一邻接矩阵的补充矩阵的维度。示例的，在预设的输入维度设置为100，第一邻接矩阵的维度为20*20，第一特征矩阵的维度为20*75的情况下，则可以确定第一邻接矩阵的补充矩阵的维度为80*80，第一特征矩阵的补充矩阵的维度为80*25。
第一邻接矩阵的补充矩阵可以设置为零矩阵或者随机采样为具有任意近邻关系的邻接矩阵。在获得第一特征矩阵后，需要根据第一特征矩阵的维度确定第一特征矩阵的补充矩阵的维度，进而构建该维度的第一特征矩阵的补充矩阵，例如：确定预设的输入维度与第一特征矩阵的维度的差值为第一特征矩阵的补充矩阵的维度，随机选取第一特征矩阵中常见的原子，通过选取的原子构建第一特征矩阵的补充矩阵。
在构建第一邻接矩阵的补充矩阵后，可以将第一邻接矩阵与第一邻接矩阵的补充矩阵进行拼接处理，得到第二邻接矩阵，该第二邻接矩阵的维度为预设的输入维度*预设的输入维度。在构建第一特征矩阵的补充矩阵后，可以将第一特征矩阵与第一特征矩阵的补充矩阵进行拼接处理，得到第二特征矩阵，该第二特征矩阵的维度为预设的输入维度*原子特征维度。示例性的，在预设的输入维度设置为100，原子特征维度为75的情况下，可以确定第二邻接矩阵的维度为100*100，第二特征矩阵的维度为100*75。
对第二邻接矩阵及第二特征矩阵的图卷积处理可以通过以下公式(1-3)、公式(1-4)及公式(1-5)实现：
Figure PCTCN2020103633-appb-000007
Figure PCTCN2020103633-appb-000008
Figure PCTCN2020103633-appb-000009
其中，
Figure PCTCN2020103633-appb-000010
表示
Figure PCTCN2020103633-appb-000011
的度矩阵，
Figure PCTCN2020103633-appb-000012
表示
Figure PCTCN2020103633-appb-000013
的度矩阵，H (l,α)表示第一层的卷积结果中的前n(待测物质的原子数)行，H (l,β)表示第一层的卷积结果中除H (l,α)以外的行，B表示第一连接矩阵，D B和
Figure PCTCN2020103633-appb-000014
分别表示第一连接矩阵B的行和列的两个度矩阵，X表示第一特征矩阵，X C表示第一特征矩阵的补充矩阵，
Figure PCTCN2020103633-appb-000015
表示归一化后的第一邻接矩阵的补充矩阵，
Figure PCTCN2020103633-appb-000016
表示归一化后的第一邻接矩阵的补充矩阵的度矩阵，σ()表示非线性激活函数，Θ表示第一层图卷积的滤波器参数，Θ (l)表示第l层图卷积的滤波器参数。在第一连接矩阵为零，即第一邻接矩阵与所述第一邻接矩阵的补充矩阵不具有邻接关系的情况下，由公式(1-3)、(1-4)简化可得到公式(1-5)。
这样一来，本公开实施例提供的测试方法可以适用于针对任意大小、结构的物质和目标类别的病变细胞进行反应测试，有较强的扩展能力。
在一种可能的实现方式中，在所述第二邻接矩阵中，所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵不具有邻接关系。其中矩阵之间不具有邻接关系，是指其中一个矩阵所包含的原子与另一个矩阵所包含的原子不具有任何相连关系。
在第一邻接矩阵与第一邻接矩阵的补充矩阵拼接得到的第二邻接矩阵中，第一邻接矩阵与第一邻接矩阵的补充矩阵不具有邻接关系，即待测物质的原子与补充矩阵中的原子不具有任何连接关系，使得第一邻接矩阵的补充矩阵可以与第一邻接矩阵构造预设的输入维度的第二邻接矩阵，第一特征矩阵的补充矩阵可以与第一特征矩阵构造预设的输入维度的第二特征矩阵，由于待测物质的原子与补充矩阵中的原子不具有任何邻接关系，故不会对待测物质的分子结构产生影响，进而不会对待测物质的测试结果产生影响。
在一种可能的实现方式中，上述将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为预设输入维度的第二特征矩阵，可以包括：
根据所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵构建第一连接矩阵，其中，所述第一连接矩阵中的元素均为预设值；
通过所述第一连接矩阵，将所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵进行连接，得到维度为预设输入维度的第二邻接矩阵；
将所述第一特征矩阵与所述第一特征矩阵的补充矩阵进行连接，得到维度为预设输入维度的第二特征矩阵。
举例来说，可以构造元素均为0的第一连接矩阵，该第一连接矩阵、第一邻接矩阵及第一邻接矩阵的补充矩阵组成第二邻接矩阵，在第二邻接矩阵中，该第一连接矩阵连接第一邻接矩阵及第一邻接矩阵的补充矩阵，使得第一邻接矩阵与第一邻接矩阵的补充矩阵不具有邻接关系。示例性的，图2示出本公开实施例提供的矩阵示意图，如图2所示的维度为100*100的第二邻接矩阵中，维度为20*20的第一邻接矩阵位于该第二邻接矩阵的左上位置，维度为80*80的第一邻接矩阵的补充矩阵位于该第二邻接矩阵的右下位置，位于第一邻接矩阵下方及第一邻接矩阵的补充矩阵的左边位置的为维度为20*80的第一连接矩阵，位于第一邻接矩阵右侧位置及第一邻接矩阵的补充矩阵的上方位置为维度为80*20的第一连接矩阵。
需要说明的是，上述图2示意的仅作为第一连接矩阵连接第一邻接矩阵与第一邻接矩阵的补充矩阵的一种示例，实际上，任何使得第一邻接矩阵与第一邻接矩阵的补充矩阵不具有邻接关系的连接方式均可以，例如：维度为20*20的第一邻接矩阵位于第二邻接矩阵的右下位置，维度为80*80的第一邻接矩阵的补充矩阵位于该第二邻接矩阵的左上位置，位于第一邻接矩阵上方及第一邻接矩阵的补充矩阵的右边位置的为维度为80*20的第一连接矩阵，位于第一邻接矩阵左侧位置及第一邻接矩阵的补充矩阵的下方位置为维度为20*80的第一连接矩阵，本公开对于第一连接矩阵连接第一邻接矩阵及第一邻接矩阵的补充矩阵的方式不做具体限定。
对应的，可以按照第一邻接矩阵及第一邻接矩阵的补充矩阵的连接方式，确定第一特征矩阵与第一特征矩阵的补充矩阵的连接方式，例如：参照图2中第一邻接矩阵及第一邻接矩阵的补充矩阵的连接方式，第一特征矩阵及第一特征矩阵的补充矩阵的连接方式可以为第一特征矩阵位于上方位置，第一特征矩阵的补充矩阵位于下方位置。
需要说明的是，在第一邻接矩阵及第一邻接矩阵的补充矩阵的连接方式为第一邻接矩阵位于第二邻接矩阵的右下位置，第一邻接矩阵的补充矩阵位于第二邻接矩阵的左上位置的情况下，第二特征矩阵中第一特征矩阵位于下方位置，第一特征矩阵的补充矩阵位于上方位置。
这样一来，即可以将待测物质的物质特征构造成满足反应测试要求的输入数据，且又不会对待测物质的分子结构产生影响，进而不会对待测物质的反应测试结果产生影响。
在一种可能的实现方式中，上述对目标类别的病变细胞进行至少一项细胞特征提取，得到所述病变细胞的至少一项细胞特征，包括以下至少一项：
对所述病变细胞的基因表突变进行特征提取，得到所述病变细胞的基因组特征；
对所述病变细胞的基因表达进行特征提取，得到所述病变细胞的转录组特征；
对所述病变细胞的脱氧核糖核酸(DeoxyriboNucleic Acid，DNA)甲基化数据进行特征提取，得到所述病变细胞的表观遗传组特征。
举例来说，在确定目标类别的病变细胞后，可以获取该病变细胞的基因表突变、基因表达及DNA甲基化数据，该获取过程可以为采用相关技术进行提取，或者是从数据库中直接进行查询，本公开在此对该过程不再赘述。
示例性的，可以预先将病变细胞的基因表突变、基因表达及DNA甲基化数据预处理为固定维度的向量，例如：将病变细胞的基因表突变预处理成为34673维的向量、将病变细胞的基因表达预处理成为697维的向量、将病变细胞的DNA甲基化数据预处理成为808维的向量，预训练提取基因组特征的卷积神经网络，并通过该卷积神经网络对预处理后的病变细胞的基因表突变进行特征提取，得到该病变细胞的基因组特征；可以预训练提取转录组特征的卷积神经网络，并通过该卷积神经网络对预处理后的病变细胞的基因表达进行特征提取，得到该病变细胞的转录组特征；可以预训练提取表观遗传组特征的卷积神经网络，并通过该卷积神经网络对预处理后的DNA甲基化数据进行特征提取，得到该病变细胞的表观遗传组特征，其中，基因组特征的维度、转录组特征的维度及表观遗传组特征的维度与物质特征的维度相同。在一种可能的实现方式中，用于提取细胞特征的卷积神经网络为多模态子神经网络。
在一种可能的实现方式中，上述所述细胞特征可以包括基因组特征、转录组特征、表观遗传组特征，上述将所述物质特征及所述至少一项细胞特征进行特征连接后，得到连接后的组合特征，包括：
将所述物质特征和所述基因组特征、所述转录组特征、所述表观遗传组特征中的至少一项进行特征连接后，得到连接后的组合特征。
示例性的，可以通过将待测物质的物质特征与基因组特征、所述转录组特征、所述表观遗传组特征进行特征连接，得到组合特征，该组合特征可以表示为：物质特征+基因组特征+转录组特征+表观遗传组特征。通过对该组合特征进行卷积处理，可以得到待测物质针对病变细胞的反应预测结果。
这样一来，可以多模态的学习病变细胞的多种细胞特征，根据充分的细胞特征进行反应结果的预测，可以提高预测结果的精准度。
为使本领域技术人员更好的理解本申请实施例，以下通过图3所示示例对本公开实施例加以说明。
图3示出本公开实施例提供的预测方法的流程示意图，如图3所示，待测物质为药物，病变细胞为癌细胞。根据待测药物的分子结构构建待测药物的结构特征图，通过物质特征提取网络对该结构特征图进行特征提取，得到待测药物的物质特征。获取癌细胞的基因表突变、基因表达及DNA甲基化数据，通过细胞特征提取网络进行细胞特征提取，其中，细胞特征网络包括：基因组特征提取网络、转录组特征提取网络及遗传组特征提取网络，可以通过基因组特征提取网络对基因表突变进行特征提取，得到癌细胞的基因组特征，通过转录组特征提取网络对基因表达进行特征提取，得到癌细胞的转录组特征，通过表观遗传组特征提取网络对DNA甲基化数据进行特征提取，得 到癌细胞的表观遗传组特征。在对待测药物的物质特征进行池化处理后，将池化处理后的物质特征与基因组特征、转录组特征及表观遗传组特征进行连接处理，得到组合特征，并对组合特征进行卷积处理，得到待测药物对该癌细胞的预测的反应结果(该反应结果表示该待测药物对该癌细胞敏感还是抑制)。
在一种可能的实现方式中，上述方法通过神经网络实现，所述方法还包括：通过预设的训练集训练所述神经网络，所述训练集包括多组样本数据，每组样本数据包括样本物质的结构特征图、样本病变细胞的基因表突变、样本病变细胞的基因表达、及样本病变细胞的DNA甲基化数据、及样本物质针对所述样本病变细胞的标注反应结果。
在一种可能的实现方式中，所述神经网络为一致性图卷积神经网络。
在一种可能的实现方式中，所述神经网络可以包括第一特征提取网络、第二特征提取网络及预测网络，所述方法通过预设的训练集训练所述神经网络，可以包括：
通过所述第一特征提取网络对所述样本物质的结构特征图进行特征提取，得到所述样本物质的样本物质特征；
通过所述第二特征提取网络分别提取所述样本病变细胞的基因表突变对应的样本基因组特征、所述样本病变细胞的基因表达对应的样本转录组特征、及所述样本病变细胞的DNA甲基化数据对应的样本表观遗传组特征；
通过所述预测网络对连接后的样本物质特征、样本基因组特征、样本转录组特征及样本表观遗传组特征，进行卷积处理，预测样本物质对所述样本病变细胞的反应结果；
根据所述反应预测结果及所述标注反应结果，确定所述神经网络的预测损失；
根据所述预测损失，训练所述神经网络。
举例来说，可以通过第一特征提取网络对样本物质的结构特征图进行特征提取，得到样本物质的样本物质特征。第二特征提取网络可以包括第一子网络、第二子网络及第三子网络，可以通过第一子网络对样本病变细胞的基因表突变进行特征提取，得到样本基因组特征，通过第二子网络对样本病变细胞的基因表达进行特征提取，得到样本转录组特征，通过第三子网络对样本病变细胞的DNA甲基化数据进行特征提取，得到样本表观遗传组特征。将样本物质特征、样本基因组特征、样本转录组特征及样本表观遗传组特征进行连接，得到组合样本特征；通过预测网络对组合样本特征进行卷积处理，得到样本物质对样本病变细胞的反应结果。根据反应结果及标注反应结果，确定神经网络的预测损失，并根据该预测损失调整神经网络的网络参数，以使神经网络的预测损失满足训练要求，例如：使神经网络的预测损失小于训练阈值。
可以理解，本公开实施例提供的上述各个方法实施例，在不违背原理逻辑的情况下，均可以彼此相互结合形成结合后的实施例，限于篇幅，本公开 不再赘述。本领域技术人员可以理解，在具体实施方式的上述方法中，各步骤的具体执行顺序应当以其功能和可能的内在逻辑确定。
此外，本公开实施例还提供了预测装置、电子设备、计算机可读存储介质和程序，上述均可用来实现本公开实施例提供的任一种预测方法，相应技术方案和描述和参见方法部分的相应记载，不再赘述。
图4示出本公开实施例提供的预测装置的结构示意图，如图4所示，所述预测装置可以包括：
第一确定部分401，可以被配置为根据待测物质的分子结构，确定待测物质的物质特征；
提取部分402，可以被配置为提取目标类别的病变细胞的至少一项细胞特征，得到所述病变细胞的至少一项细胞特征；
第二确定部分403，可以被配置为根据所述物质特征以及所述至少一项细胞特征，预测所述待测物质针对所述病变细胞的反应结果。
这样，根据待测物质的分子结构，可以构建待测物质的结构特征图，进而基于该结构特征图可以提取待测物质的物质特征，并在提取目标类别的病变细胞的至少一项细胞特征后，可以根据待测物质的物质特征及病变细胞的至少一项细胞特征，预测待测物质针对病变细胞的反应结果。根据本公开实施例提供的预测装置，可以基于待测物质的结构特征图提取待测物质的物质特征，相比于人工提取物质特征，提取的物质特征更为稠密，从而可以提高测试结果的精度及获得测试结果的效率。
在一种可能的实现方式中，所述第一确定部分401，被配置为：
根据待测物质的分子结构，构建所述待测物质的结构特征图，所述结构特征图包括至少两个节点及各节点之间的连线，所述节点表示所述分子结构中的原子，所述连线表示所述分子结构中的原子键；
根据所述结构特征图，确定所述待测物质的物质特征。
在一种可能的实现方式中，所述第一确定部分401，还被配置为：
根据所述结构特征图得到所述待测物质的第一邻接矩阵及第一特征矩阵，所述第一邻接矩阵表示所述待测物质的各原子的近邻关系，所述第一特征矩阵表示所述待测物质的各原子的属性数据；
根据所述第一邻接矩阵及所述第一特征矩阵，得到所述待测物质的物质特征。
在一种可能的实现方式中，所述第一确定部分401，还被配置为：
根据预设输入维度及所述第一邻接矩阵的维度，构建所述第一邻接矩阵的补充矩阵，及根据所述预设的输入维度及所述第一特征矩阵的维度，构建所述第一特征矩阵的补充矩阵；
将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二特征矩阵；
对所述第二邻接矩阵及所述第二特征矩阵进行图卷积处理，得到所述待测物质的所述物质特征。
在一种可能的实现方式中，在所述第二邻接矩阵中，所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵不具有邻接关系。
在一种可能的实现方式中，所述第一确定部分401，还被配置为：
根据所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵构建第一连接矩阵；
通过所述第一连接矩阵，将所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵进行连接，得到维度为所述预设输入维度的第二邻接矩阵；
将所述第一特征矩阵与所述第一特征矩阵的补充矩阵进行连接，得到维度为所述预设输入维度的第二特征矩阵。
在一种可能的实现方式中，所述提取部分402，被配置为以下至少一项：
对所述病变细胞的基因表突变进行特征提取，得到所述病变细胞的基因组特征；
对所述病变细胞的基因表达进行特征提取，得到所述病变细胞的转录组特征；
对所述病变细胞的DNA甲基化数据进行特征提取，得到所述病变细胞的表观遗传组特征。
在一种可能的实现方式中，所述第二确定部分403，被配置为：
将所述物质特征及所述至少一项细胞特征进行特征连接，得到连接后的组合特征；
对所述组合特征进行卷积处理，得到所述待测物质针对所述病变细胞的反应结果。
在一种可能的实现方式中，所述细胞特征包括基因组特征、转录组特征和表观遗传组特征，所述第二确定部分403，还被配置为：
将所述物质特征与所述基因组特征、所述转录组特征和所述表观遗传组特征中的至少一项特征进行特征连接，得到连接后的组合特征。
在一种可能的实现方式中，所述装置通过神经网络实现，所述装置还包括：
训练部分，被配置为通过预设的训练集训练所述神经网络，所述训练集包括多组样本数据，每组样本数据包括样本物质的结构特征图、样本病变细胞的基因表突变、样本病变细胞的基因表达、样本病变细胞的DNA甲基化数据、及样本物质针对所述样本病变细胞的标注反应结果。
在一种可能的实现方式中，所述神经网络包括第一特征提取网络、第二特征提取网络及预测网络，所述训练部分，还被配置为：
通过所述第一特征提取网络，对所述样本物质的结构特征图进行特征提取，得到所述样本物质的样本物质特征；
通过所述第二特征提取网络，分别提取所述样本病变细胞的基因表突变对应的样本基因组特征、所述样本病变细胞的基因表达对应的样本转录组特 征、及所述样本病变细胞的DNA甲基化数据对应的样本表观遗传组特征；
通过所述预测网络，对连接后的样本物质特征、样本基因组特征、样本转录组特征及样本表观遗传组特征进行卷积处理，得到样本物质对所述样本病变细胞的反应结果；
根据所述反应结果及所述标注反应结果，确定所述神经网络的预测损失；
根据所述预测损失，训练所述神经网络。
在一些实施例中，本公开实施例提供的装置具有的功能或包含的部分可以被配置为执行上文方法实施例描述的方法，其具体实现可以参照上文方法实施例的描述，为了简洁，这里不再赘述。
在本公开实施例以及其他的实施例中，“部分”可以是部分电路、部分处理器、部分程序或软件等等，当然也可以是单元，还可以是模块也可以是非模块化的。
本公开实施例还提出一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。计算机可读存储介质可以是非易失性计算机可读存储介质。
本公开实施例还提出一种电子设备，包括：处理器；被配置为存储处理器可执行指令的存储器；其中，所述处理器被配置为调用所述存储器存储的指令，以执行上述方法。
本公开实施例还提供了一种计算机程序产品，包括计算机可读代码，当计算机可读代码在设备上运行时，设备中的处理器执行被配置为实现如上任一实施例提供的预测方法的指令。
本公开实施例还提供了另一种计算机程序产品，被配置为存储计算机可读指令，指令被执行时使得计算机执行上述任一实施例提供的预测方法的操作。
电子设备可以被提供为终端、服务器或其它形态的设备。
图5示出本公开实施例提供的一种电子设备的结构示意图。例如，电子设备800可以是移动电话，计算机，数字广播终端，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等终端。
参照图5，电子设备800可以包括以下一个或多个组件：处理组件802，存储器804，电源组件806，多媒体组件808，音频组件810，输入/输出(I/O)的接口812，传感器组件814，以及通信组件816。
处理组件802通常控制电子设备800的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理组件802可以包括一个或多个处理器820来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件802可以包括一个或多个模块，便于处理组件802和其他组件之间的交互。例如，处理组件802可以包括多媒体模块，以方便多媒体组件808和处理组件802之间的交互。
存储器804被配置为存储各种类型的数据以支持在电子设备800的操作。这些数据的示例包括用于在电子设备800上操作的任何应用程序或方法的指 令，联系人数据，电话簿数据，消息，图片，视频等。存储器804可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器(EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。
电源组件806为电子设备800的各种组件提供电力。电源组件806可以包括电源管理系统，一个或多个电源，及其他与为电子设备800生成、管理和分配电力相关联的组件。
多媒体组件808包括在所述电子设备800和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(LCD)和触摸面板(TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。所述触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与所述触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件808包括一个前置摄像头和/或后置摄像头。当电子设备800处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。
音频组件810被配置为输出和/或输入音频信号。例如，音频组件810包括一个麦克风(MIC)，当电子设备800处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器804或经由通信组件816发送。在一些实施例中，音频组件810还包括一个扬声器，用于输出音频信号。
I/O接口812为处理组件802和外围接口模块之间提供接口，上述外围接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。
传感器组件814包括一个或多个传感器，用于为电子设备800提供各个方面的状态评估。例如，传感器组件814可以检测到电子设备800的打开/关闭状态，组件的相对定位，例如所述组件为电子设备800的显示器和小键盘，传感器组件814还可以检测电子设备800或电子设备800一个组件的位置改变，用户与电子设备800接触的存在或不存在，电子设备800方位或加速/减速和电子设备800的温度变化。传感器组件814可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件814还可以包括光传感器，如CMOS或CCD图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件814还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。
通信组件816被配置为便于电子设备800和其他设备之间有线或无线方式的通信。电子设备800可以接入基于通信标准的无线网络，如WiFi，2G或3G，或它们的组合。在一个示例性实施例中，通信组件816经由广播信道接 收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，所述通信组件816还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT)技术和其他技术来实现。
在示例性实施例中，电子设备800可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现，被配置为执行上述方法。
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器804，上述计算机程序指令可由电子设备800的处理器820执行以完成上述方法。
图6示出本公开实施例提供的一种电子设备的结构示意图。例如，电子设备1900可以被提供为一服务器。参照图6，电子设备1900包括处理组件1922，其进一步包括一个或多个处理器，以及由存储器1932所代表的存储器资源，被配置为存储可由处理组件1922的执行的指令，例如应用程序。存储器1932中存储的应用程序可以包括一个或一个以上的每一个对应于一组指令的部分。此外，处理组件1922被配置为执行指令，以执行上述预测方法。
电子设备1900还可以包括一个电源组件1926被配置为执行电子设备1900的电源管理，一个有线或无线网络接口1950被配置为将电子设备1900连接到网络，和一个输入输出(I/O)接口1958。电子设备1900可以操作基于存储在存储器1932的操作系统，例如Windows ServerTM，Mac OS XTM，UnixTM,LinuxTM，FreeBSDTM或类似。
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器1932，上述计算机程序指令可由电子设备1900的处理组件1922执行以完成上述方法。
本公开可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本公开的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电 信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本公开操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
该计算机程序产品可以具体通过硬件、软件或其结合的方式实现。在一个可选实施例中，所述计算机程序产品具体体现为计算机存储介质，在另一个可选实施例中，计算机程序产品具体体现为软件产品，例如软件开发包(Software Development Kit，SDK)等等。
以上已经描述了本公开的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。
工业实用性
本公开实施例根据待测物质的分子结构，确定待测物质的物质特征，并在提取目标类别的病变细胞的至少一项细胞特征后，根据待测物质的物质特征及病变细胞的至少一项细胞特征，预测待测物质针对病变细胞的反应结果。根据本公开实施例提供的预测方法及装置、电子设备和存储介质，可以基于待测物质的结构特征图提取待测物质的物质特征，相比于人工提取物质特征，提取的物质特征更为稠密，进一步的可以提高测试结果的精度及获得测试结果的效率。

Claims (11)

  1. 一种预测方法，包括：
    根据待测物质的分子结构，确定待测物质的物质特征；
    提取目标类别的病变细胞的至少一项细胞特征，得到所述病变细胞的至少一项细胞特征；
    根据所述物质特征和所述至少一项细胞特征，预测所述待测物质针对所述病变细胞的反应结果。
  2. 根据权利要求1所述的方法，其中，所述根据待测物质的分子结构，确定待测物质的物质特征，包括：
    根据所述待测物质的分子结构，构建所述待测物质的结构特征图，所述结构特征图包括至少两个节点及各节点之间的连线，所述节点表示所述分子结构中的原子，所述连线表示所述分子结构中的原子键；
    根据所述结构特征图，确定所述待测物质的所述物质特征。
  3. 根据权利要求2所述的方法，其中，所述根据所述结构特征图，确定所述待测物质的所述物质特征，包括：
    根据所述结构特征图得到所述待测物质的第一邻接矩阵及第一特征矩阵，所述第一邻接矩阵表示所述待测物质的各原子之间的近邻关系，所述第一特征矩阵表示所述待测物质的各原子的属性数据；
    根据所述第一邻接矩阵及所述第一特征矩阵，得到所述待测物质的所述物质特征。
  4. 根据权利要求3所述的方法，其中，所述根据所述第一邻接矩阵及所述第一特征矩阵，得到所述待测物质的所述物质特征，包括：
    根据预设输入维度及所述第一邻接矩阵的维度，构建所述第一邻接矩阵的补充矩阵，及根据所述预设输入维度及所述第一特征矩阵的维度，构建所述第一特征矩阵的补充矩阵；
    将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二特征矩阵；
    对所述第二邻接矩阵及所述第二特征矩阵进行图卷积处理，得到所述待测物质的所述物质特征。
  5. 根据权利要求4所述的方法，其中，在所述第二邻接矩阵中，所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵不具有邻接关系。
  6. 根据权利要求4或5所述的方法，其中，所述将所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二邻接矩阵，及将所述第一特征矩阵及所述第一特征矩阵的补充矩阵进行拼接处理，得到维度为所述预设输入维度的第二特征矩阵，包括：
    根据所述第一邻接矩阵及所述第一邻接矩阵的补充矩阵构建第一连接矩阵，其中，所述第一连接矩阵中的元素均为预设值。
    通过所述第一连接矩阵，将所述第一邻接矩阵与所述第一邻接矩阵的补充矩阵进行连接，得到维度为所述预设输入维度的所述第二邻接矩阵；
    将所述第一特征矩阵与所述第一特征矩阵的补充矩阵进行连接，得到维度为所述预设输入维度的所述第二特征矩阵。
  7. 根据权利要求1至5任一项所述的方法，其中，所述提取目标类别的病变细胞的至少一项细胞特征，得到所述病变细胞的至少一项细胞特征，包括以下至少一项：
    对所述病变细胞的基因表突变进行特征提取，得到所述病变细胞的基因组特征；
    对所述病变细胞的基因表达进行特征提取，得到所述病变细胞的转录组特征；
    对所述病变细胞的脱氧核糖核酸(DeoxyriboNucleic Acid，DNA)甲基化数据进行特征提取，得到所述病变细胞的表观遗传组特征。
  8. 一种预测装置，包括：
    第一确定部分，被配置为根据待测物质的分子结构，确定待测物质的物质特征；
    提取部分，被配置为提取目标类别的病变细胞的至少一项细胞特征，得到所述病变细胞的至少一项细胞特征；
    第二确定部分，被配置为根据所述物质特征和所述至少一项细胞特征，预测所述待测物质针对所述病变细胞的反应结果。
  9. 一种电子设备，包括：
    处理器；
    被配置为存储处理器可执行指令的存储器；
    其中，所述处理器被配置为调用所述存储器存储的指令，以执行权利要求1至7中任意一项所述的方法。
  10. 一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现权利要求1至7中任意一项所述的方法。
  11. 一种计算机程序，包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行用于实现权利要求1至7任一项所述的预测方法。
PCT/CN2020/103633 2019-11-18 2020-07-22 一种预测方法及装置、电子设备和存储介质 WO2021098256A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2021543171A JP2022518283A (ja) 2019-11-18 2020-07-22 予測方法及び装置、電子機器並びに記憶媒体
US17/739,541 US20220285038A1 (en) 2019-11-18 2022-05-09 Prediction method, electronic device, and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911125921.X 2019-11-18
CN201911125921.XA CN110867254A (zh) 2019-11-18 2019-11-18 预测方法及装置、电子设备和存储介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/739,541 Continuation US20220285038A1 (en) 2019-11-18 2022-05-09 Prediction method, electronic device, and storage medium

Publications (1)

Publication Number Publication Date
WO2021098256A1 true WO2021098256A1 (zh) 2021-05-27

Family

ID=69654942

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/103633 WO2021098256A1 (zh) 2019-11-18 2020-07-22 一种预测方法及装置、电子设备和存储介质

Country Status (5)

Country Link
US (1) US20220285038A1 (zh)
JP (1) JP2022518283A (zh)
CN (1) CN110867254A (zh)
TW (1) TWI771803B (zh)
WO (1) WO2021098256A1 (zh)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110867254A (zh) * 2019-11-18 2020-03-06 北京市商汤科技开发有限公司 预测方法及装置、电子设备和存储介质
US20210287137A1 (en) * 2020-03-13 2021-09-16 Korea University Research And Business Foundation System for predicting optical properties of molecules based on machine learning and method thereof
CN111986740B (zh) * 2020-09-03 2024-05-14 深圳赛安特技术服务有限公司 化合物分类方法及相关设备
CN112768089B (zh) * 2021-04-09 2021-06-22 至本医疗科技（上海）有限公司 用于预测药物敏感状态的方法、设备和存储介质
CN113241130B (zh) * 2021-06-08 2022-04-22 西南交通大学 一种基于图卷积网络的分子结构预测方法
CN113707214B (zh) * 2021-08-31 2024-05-24 平安科技（深圳）有限公司 代谢物标记方法、装置、计算机设备及存储介质
CN117524346A (zh) * 2023-11-20 2024-02-06 东北林业大学 一种多视图癌症药物反应预测系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160267397A1 (en) * 2015-03-11 2016-09-15 Ayasdi, Inc. Systems and methods for predicting outcomes using a prediction learning model
CN108959841A (zh) * 2018-04-16 2018-12-07 华南农业大学 一种基于dbn算法的药物靶向蛋白作用预测方法
CN109033738A (zh) * 2018-07-09 2018-12-18 湖南大学 一种基于深度学习的药物活性预测方法
CN110289050A (zh) * 2019-05-30 2019-09-27 湖南大学 一种基于图卷积和词向量的药物－靶标相互作用预测方法
CN110867254A (zh) * 2019-11-18 2020-03-06 北京市商汤科技开发有限公司 预测方法及装置、电子设备和存储介质

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010003020A1 (en) * 2008-07-01 2010-01-07 The Board Of Trustees Of The Leland Stanford Junior University Methods and systems for assessment of clinical infertility
JP6671348B2 (ja) * 2014-05-05 2020-03-25 アトムワイズ，インコーポレイテッド 結合親和性予測システム及び方法
US9984201B2 (en) * 2015-01-18 2018-05-29 Youhealth Biotech, Limited Method and system for determining cancer status
WO2016209999A1 (en) * 2015-06-22 2016-12-29 Counsyl, Inc. Methods of predicting pathogenicity of genetic sequence variants
US20190050530A1 (en) * 2016-02-09 2019-02-14 Toma Biosciences, Inc. Systems and Methods for Analyzing Nucleic Acids
EP3598128A4 (en) * 2016-12-28 2020-12-30 National Institute of Biomedical Innovation, Healty and Nutrition CHARACTERISTICS ANALYSIS PROCEDURES AND CLASSIFICATION OF PHARMACEUTICAL COMPONENTS USING TRANSCRIPTOMS
CN110232978B (zh) * 2019-06-14 2022-05-17 西安电子科技大学 基于多维网络的癌症细胞系治疗药物预测方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160267397A1 (en) * 2015-03-11 2016-09-15 Ayasdi, Inc. Systems and methods for predicting outcomes using a prediction learning model
CN108959841A (zh) * 2018-04-16 2018-12-07 华南农业大学 一种基于dbn算法的药物靶向蛋白作用预测方法
CN109033738A (zh) * 2018-07-09 2018-12-18 湖南大学 一种基于深度学习的药物活性预测方法
CN110289050A (zh) * 2019-05-30 2019-09-27 湖南大学 一种基于图卷积和词向量的药物－靶标相互作用预测方法
CN110867254A (zh) * 2019-11-18 2020-03-06 北京市商汤科技开发有限公司 预测方法及装置、电子设备和存储介质

Also Published As

Publication number Publication date
TWI771803B (zh) 2022-07-21
US20220285038A1 (en) 2022-09-08
CN110867254A (zh) 2020-03-06
JP2022518283A (ja) 2022-03-14
TW202133183A (zh) 2021-09-01

Similar Documents

Publication Publication Date Title
WO2021098256A1 (zh) 一种预测方法及装置、电子设备和存储介质
KR102365890B1 (ko) 신경망의 훈련 방법, 장치 및 저장 매체
TWI781359B (zh) 人臉和人手關聯檢測方法及裝置、電子設備和電腦可讀儲存媒體
CN110009090B (zh) 神经网络训练与图像处理方法及装置
CN109359056B (zh) 一种应用程序测试方法及装置
WO2021036382A1 (zh) 图像处理方法及装置、电子设备和存储介质
WO2016011741A1 (zh) 自动连接无线局域网的方法及装置
TWI785638B (zh) 目標檢測方法、電子設備和電腦可讀儲存介質
CN110659690B (zh) 神经网络的构建方法及装置、电子设备和存储介质
KR20150117202A (ko) 클러스터링 방법, 관련 장치, 프로그램 및 기록매체
CN109522937B (zh) 图像处理方法及装置、电子设备和存储介质
TWI779449B (zh) 對象計數方法、電子設備、電腦可讀儲存介質
CN111523599B (zh) 目标检测方法及装置、电子设备和存储介质
JP2022510660A (ja) データ処理方法及びその装置、電子機器、並びに記憶媒体
CN113470735A (zh) 预测方法及装置、电子设备和存储介质
CN115035947A (zh) 蛋白质结构建模方法及装置、电子设备和存储介质
WO2020164261A1 (zh) 液质检测结果的质量控制方法及装置
CN113053454A (zh) 分类方法及装置、电子设备和存储介质
US20150262033A1 (en) Method and terminal device for clustering
CN112381223A (zh) 神经网络训练与图像处理方法及装置
WO2022110411A1 (zh) 信息处理方法及装置、电子设备和计算机可读存储介质
CN110765943A (zh) 网络训练、识别方法及装置、电子设备和存储介质
JP2023507697A (ja) 情報処理方法及び装置、電子デバイス並びにコンピュータ可読記憶媒体
CN113312475B (zh) 一种文本相似度确定方法及装置
CN111081314A (zh) 基因变异的识别方法及装置、电子设备和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20888957

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021543171

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20888957

Country of ref document: EP

Kind code of ref document: A1