CN110853714A - Drug relocation model based on pathogenic contribution network analysis - Google Patents
Drug relocation model based on pathogenic contribution network analysis Download PDFInfo
- Publication number
- CN110853714A CN110853714A CN201910997755.6A CN201910997755A CN110853714A CN 110853714 A CN110853714 A CN 110853714A CN 201910997755 A CN201910997755 A CN 201910997755A CN 110853714 A CN110853714 A CN 110853714A
- Authority
- CN
- China
- Prior art keywords
- pathogenic
- network
- data
- contribution
- node
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Classifications
-
- 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
- G16C20/50—Molecular design, e.g. of drugs
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a medicine relocation model based on pathogenic contribution network analysis, which comprises a data preparation module, a pathogenic contribution network construction and calculation module and a medicine embedding and sequencing module; the data preparation module is used for fully preparing data required by the construction and calculation part of the pathogenic contribution network, including data cleaning and entity disambiguation; the pathogenic contribution network construction and calculation module is used for completing pathogenic network construction and node contribution calculation and establishing a complete pathogenic contribution network, wherein the pathogenic contribution network construction and node contribution calculation comprises the pathogenic network construction and the node contribution calculation; the drug embedding and ordering module task is to list all the drugs that become effective drugs for the target disease and how likely each drug will become effective drugs for the target disease, including drug embedding and result ordering. The invention combines the biomedical data and the problem solving angle with the solution of the computer technology, better solves the defects existing in the development of new drugs and predicts the new drug-disease relationship.
Description
Technical Field
The invention relates to the fields of computer technology and biomedicine, in particular to a medicine relocation model based on pathogenic contribution network analysis.
Background
The development of new drugs has been the main approach for obtaining effective drugs for diseases. However, the research and development of new drugs have many defects, one new drug generally needs 13-15 years to be put into the market, and the average research and development fund needs 20-30 hundred million dollars. Moreover, the input-output ratio of new drug development is higher and higher. With the accumulation of biomedical data and the development of computer technology, strategies for drug relocation research to assist or replace new drug development are gradually developed. To some extent, this is a result of technological advances, mainly benefitting from the advent and development of large-scale screening systems (rapid testing of compound function in different cell lines), computer models (predicting likely drugs against similar pathogenic mechanisms), and large-scale screening systems (revealing similar molecular mechanisms among different diseases). Under such a large environment, over 3000 drugs that have been put on the market in the last decades have become virgins where relocated drugs are to be developed. Most of the drugs or active compounds can directly skip the clinical I phase, thereby greatly reducing the development cost and having lower risk of side effects at the later stage. It is estimated that the average development cost of such a relocated drug is only about 3 billion dollars, and the development cycle can be reduced even to half (6.5 years).
Drug relocation techniques have also been developed in recent years. At first, a new drug-disease relationship is mainly found through a relationship transmission mode of A- > B- > C, and then technologies such as neural networks, machine learning and the like are also applied to drug relocation research. But with the explosive increase in biomedical data volumes, network structures with inherent advantages in complex relationship processing have become a trend for drug relocation research. In network-based drug relocation studies, the focus is on the prediction of relationship pathways based on drug targets. That is, drug-other disease associations are established by finding and predicting associations of target substances of drugs with other diseases. This strategy is a strategy that is thought to be centered on drug treatment capacity, and has also achieved some cases of success.
Drug relocation studies based on pathogenic contribution network analysis do not focus on drug treatment capacity, but rather focus on the causative factors of the disease. The pathogenic contribution network expresses all causative factors of the disease by using a network structure, which is an innovative strategy for drug relocation research and is a new network model for the drug relocation research centered on the disease.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, introduce the network analysis technology in the computer field into the drug relocation research in the biomedical field, provide a drug relocation model based on pathogenic contribution network analysis aiming at most diseases, combine the biomedical data and the problem solving angle with the solution of the computer technology, better solve the defects in the development of new drugs and predict the new drug-disease relationship.
The purpose of the invention is realized by the following technical scheme.
The invention relates to a medicine relocation model based on pathogenic contribution network analysis, which comprises a data preparation module, a pathogenic contribution network construction and calculation module and a medicine embedding and sequencing module;
the task of the data preparation module is to fully prepare all data required by the construction and calculation part of the pathogenic contribution network, and the data preparation module comprises two parts, namely data cleaning and entity disambiguation; the data cleaning part aims to process the data content and format into the condition required by constructing a network, including impurity filtering of the data and type conversion of the data; the entity disambiguation part aims to eliminate different names of the same concept in the biomedical field, and unifies different names of objects describing the same biomedical field in data into UMLS language by taking UMLS codes as a reference;
the main task of the pathogenic contribution network construction and calculation module is to complete the construction of a pathogenic network and the calculation of node contribution, and establish a complete pathogenic contribution network, including the construction of the pathogenic network and the calculation of the node contribution; the pathogenic network construction aims at unifying all pathogenic elements of the target disease in the network, including node introduction and edge introduction; the node contribution calculation is to calculate and distinguish the importance of each node (pathogenic element) in the network, firstly calculate the node centrality, and then calculate the node contribution value on the basis of the node centrality;
the task of the drug embedding and sorting module is to list all the drugs becoming effective drugs of the target diseases and the possibility of each drug becoming effective drugs of the target diseases, and the drug embedding and result sorting module comprises two parts; the purpose of the medicine embedding is to calculate the potential treatment capacity of the medicine on the pathogenic contribution network in a mode of medicine node and relation embedding; the purpose of the result ranking is to screen out candidate drugs with greater network treatment capacity, including gene tendency ranking and protein tendency ranking.
The impurity filtering of the data comprises filtering out meaningless data included in the desired data; the type conversion of the data comprises the steps of firstly unifying multi-source heterogeneous data into triple data in a main-predicate-guest form, and then converting the triple data into a CSV format.
The node introduction is to introduce all genes and proteins causing target diseases in CSV data after data cleaning into a graph database after the target diseases to be researched are selected; the side introduction is that the pathogenic gene and the pathogenic protein are taken as network nodes in the graph, and four types of relations of gene-gene, gene-protein, protein-gene and protein-protein are taken as the side of the network.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
the present invention can predict its potential drugs for any disease. The existing medicines for treating other diseases but medicines with the effect on the target diseases are screened out. Not only can relieve the research and development difficulty of medicines for certain complex diseases, but also can endow the existing medicines with new values. The resulting drug list can also be used as an alternative to clinical trials, greatly narrowing the scope of trials.
Drawings
FIG. 1 is a schematic diagram of the structure and module division of a drug relocation model based on pathogenic contribution network analysis.
FIG. 2 is a pathogenic contribution network construction flow.
Fig. 3 is a schematic diagram of a data storage form after data cleaning.
FIG. 4 shows data amounts related to PD (GNGM is a gene, and AAPP is a protein).
FIG. 5 is a predicted drug list resulting from two ordering strategies.
Figure 6 is the effective drug ratio of the different intercept criteria of the two strategies.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The invention relates to a medicine relocation model based on pathogenic contribution network analysis, which mainly comprises a data preparation module, a pathogenic contribution network construction and calculation module and a medicine embedding and sequencing module as shown in figure 1.
The task of the data preparation module is to fully prepare all data required by the construction and calculation part of the pathogenic contribution network, and mainly comprises two parts of data cleaning and entity disambiguation.
The data cleansing part aims to process the data content and format to the conditions required for building the network. The part mainly comprises two parts: impurity filtering of data and type conversion of data. Impurity filtering of the data includes filtering out some meaningless data that is entrained in the desired data; the type conversion of the data comprises the steps of unifying multi-source heterogeneous data into triple data in a main-predicate-guest form, and then converting the triple data into a CSV format.
The entity disambiguation section is intended to eliminate different nomenclature of the same concept within the biomedical field. The part is to unify all different names describing the same biomedical field in the data into UMLS Language based on the UMLS (the Unified Medical Language System) code.
The main task of the pathogenic contribution network construction and calculation module is to complete the construction of the pathogenic network and the calculation of the node contribution, and establish a complete pathogenic contribution network, including the construction of the pathogenic network and the calculation of the node contribution. Fig. 2 shows a flow of the pathogenic contribution network construction.
Before the technical content of this part is described in detail, the meaning of the pathogenic contribution network model needs to be introduced, which is mainly described from two perspectives of "pathogenic" (pathogenic network construction) and "contribution" (node contribution calculation).
The first is "pathogenic". From a medical point of view, the disease of a disease usually has many pathogenic factors, and the factors are usually genes and proteins only considering the intrinsic factors. The pathogenic network takes the genes and the proteins as network nodes, and takes four relations of gene-protein, gene-gene, protein-gene and protein-protein as edges in the network. The construction of a pathogenic network is equivalent to aggregating all the factors that can cause the disease in the same network.
The second is the "contribution". The contribution of the pathogenic network is calculated to distinguish these factors that may cause disease by their importance. For example, if a gene A is present in a disease causing gene that positively stimulates most of the other genes in the disease causing gene, and a gene B is also present in a disease causing gene that does not stimulate other genes in the disease causing gene, it can be hypothesized that gene A plays a larger role than gene B in the overall disease process. The contribution calculation expresses the pathological process by digital calculation and structural simulation of a network.
The pathogenic network construction and the node contribution calculation respectively correspond to the two sections of descriptions, and after the pathogenic network construction is completed, the contribution value of the network node is calculated, so that the final pathogenic contribution network model is constructed.
The pathogenic network is constructed to unify all pathogenic elements of the target disease in the network, including node introduction and edge introduction. The target disease needs to be determined first, i.e. which disease the network model is directed to. After the target disease to be studied is selected, all genes and proteins that can cause the target disease in the data-washed CSV data are imported into the graph database (node import). In the figure, pathogenic genes and pathogenic proteins are used as network nodes, and four types of relations of gene-gene, gene-protein, protein-gene and protein-protein are used as sides (introduced) of the network.
The node contribution calculation is to calculate and distinguish the importance of each node (pathogenic element) in the network. Firstly, calculating the node centrality, and then calculating the node contribution value on the basis of the node centrality.
The basis of the node centrality calculation mainly comprises the relationship density degree around the node, and the higher the relationship density degree around the node is, the more important the node is in the pathogenic process, and the higher the calculation result is. The other basis is whether the neighbor of the node is important, whether the neighbor node influenced by the node is an important node, and the more important the neighbor node is, the more important the node is, and the higher the calculation result is. The last criterion is the number of other nodes that the node can influence, i.e. the more other nodes the node can influence on the network, the more important the node is, and the higher the calculation result is. And calculating the centrality of the nodes according to the three factors, wherein each node has a node centrality result.
The calculation of the node contribution value is to substitute the confidence condition of the nodes and the relations (the confidence condition represents the reliability of the data sources of the nodes or the relations) into the centrality on the basis of the centrality of the nodes, and the final calculation result is called 'contribution', which represents the importance degree of the substance in the pathogenic process of the target disease.
The task of the drug embedding and ranking module is to list all drugs that may become active agents for the target disease and how likely each drug will become an active agent for the target disease, including both drug embedding and result ranking.
The purpose of the medicine embedding is to calculate the treatment capacity of the medicine on the pathogenic contribution network in a medicine node and relationship embedding mode. The part is to embed the drug entities in the data into the network as new nodes, and if these newly embedded drugs have therapeutic relationships with the original gene and protein nodes in the network, then these therapeutic relationships are added to the network as new edges. And according to the newly added treatment relationship, the treatment capacity of the medicine to the whole network is calculated, which can also be called the possibility of the medicine becoming the effective medicine of the target disease. To express the above meaning, a variable of drug score is defined, which is divided into two sub-variables of drug gene score and drug protein score. The drug score was calculated as follows: assuming that the embedded drug D1 has a therapeutic relationship with pathogenic genes G1, G2 and G3 in the network and also has a therapeutic relationship with pathogenic proteins P1 and P2 in the network, the drug gene score of the drug D1 is equal to the sum of the contributions of the three nodes G1, G2 and G3; the pharmatein score is equal to the sum of the contributions from the two nodes P1, P2.
The purpose of the result ranking is to screen out the candidate drugs with larger network treatment capacity. Because the pathological mechanisms of different diseases are different, genes account for the dominant part in the pathological mechanisms of some diseases, and some genes account for the dominant part, the model is divided into two sequencing strategies, namely gene tendency sequencing and protein tendency sequencing. Because the difference between the number of gene nodes and the number of protein nodes in the pathogenic contribution network is large, adding the drug gene score and the drug protein score in a ratio of 1:1 cannot scientifically show the treatment capability of the drug in the pathogenic network. For the genes in the network, the contribution values of the genes comprise the gene-protein and gene-gene relations, so when the genes are sorted by only depending on the variable of the drug gene score, the influence of the genes and the proteins on diseases is still considered. Of course this ordering strategy is more inclined to the ability of the gene to affect the disease. The rank of the pharmatein score is the ability of the protein of interest to influence the disease. Therefore, it was decided to rank using two ranking strategies, pharmacogenomic score and pharmacogrotein score, respectively. Therefore, we rank the disease network model results with more proteins in the pathogenic factors according to protein tendency, and rank the disease network model results with more proteins in the pathogenic factors according to gene tendency. In any sort, the more top-ranked drugs prove the greater the ability to treat the network, and the greater the likelihood of becoming potential drugs for the target disease.
Example 1:
the invention relates to a medicine relocation model based on pathogenic contribution network analysis.
The realization technology comprises the following steps: in the process of model construction and system development, java programming language is used as background language, Mysql relational database and Neo4j graphic database are used as databases, and SQL + CQL (Cypher) is used as query. The display portion was processed and rendered using the neo4j Bloom tool. The computation part uses the neo4j APOC storage procedure and the graph algorithm package.
The model can be used for carrying out drug relocation research on any disease, selecting Parkinson disease (PD for short) as a target disease, and selecting a SemmedDB database as a data source for case display.
A data preparation module:
selecting a SemmedDB as a data source, extracting all genes and proteins related to PD and relations among the genes and the proteins from the SemMedDB to unify into a triple data format in a main predicate form, and converting into a data format of CSV. And performing entity disambiguation on the triple data through UMLS unified specification language. Fig. 3 shows the storage form of data in the source database, and fig. 4 shows the data volume details related to the PD.
The pathogenic contribution network construction and calculation module comprises:
and (4) importing the processed three CSV data format in the data preparation module into a Neo4j graph database, wherein the gene protein entity is used as a node, and the relationship is used as an edge. After the network construction is completed, the contribution value of each node is calculated. The contribution value is calculated as follows:
con (g) in the formula1) Representative Gene g1D is a constant in the interval (0.1) and is used for adjusting the iteration stability. n isiIs g1Other nodes of influence, Od (n)i) Representing a node niOut of date in the network. When the number of iterations in the PD experiment was set at 20, the results tended to be stable. Therefore, the iteration number 20, d-0.85, was selected as the parameter for calculating the initial settings of the PD drug relocation experiment.
A drug embedding and ordering module:
the existing medicine nodes and treatment relations are embedded into a network, the sum of the contribution values of the nodes which can be treated by each medicine is counted, and the contribution values of all elements which can be treated by the same medicine are summed up to obtain the medicine score. Then, the medicines embedded into the network are sequenced to obtain 2 medicine sets under 2 sequencing strategies. These drugs were pooled as the final result of the model. Since the goal of drug relocation studies is to discover new therapeutically-capable drugs, the known effective drugs for the disease should not be in the outcome of the model. Figure 5 shows a PD drug relocation model ordered drug list that excludes known PD active drugs. P2ASRM and P2GRRM are abbreviations for the 2 ordering strategy, respectively. The list in fig. 5 was taken from the top25 of the 2 ranking strategies, but there were 10 drugs in the P2ASRM column and 6 in the P2GRRM column of the list, since 15 of the top25 drugs in the P2ASRM ranking prediction were known (documented) as potent PD drugs, and 19 of the same P2GRRM drugs were known as potent PD drugs. The drug relocation model based on pathogenic contribution network analysis is a potential effective drug for predicting diseases from pathogenic factors of the diseases, and the obtained result contains a large proportion of PD effective drugs, on the contrary, the quantity of the PD effective drugs comprises 675, while the drugs and pharmacological substances embedded in the network exceed 5 ten thousand, almost only 1% of the PD effective substances of the whole drugs are contained in TOP25 of the model prediction result, but the proportion is very high, which also proves that the drug relocation model based on the pathogenic contribution network analysis has strong prediction capability. The prediction set is divided into TOP10, TOP25, TOP50 and to100, and the effective drug ratio under different TOP selection conditions is shown in figure 6.
While the present invention has been described in terms of its functions and operations with reference to the accompanying drawings, it is to be understood that the invention is not limited to the precise functions and operations described above, and that the above-described embodiments are illustrative rather than restrictive, and that various changes may be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (3)
1. A medicine relocation model based on pathogenic contribution network analysis is characterized by comprising a data preparation module, a pathogenic contribution network construction and calculation module and a medicine embedding and sequencing module;
the task of the data preparation module is to fully prepare all data required by the construction and calculation part of the pathogenic contribution network, including two parts of data cleaning and entity disambiguation; the data cleaning part aims to process the data content and format into the condition required by constructing a network, including impurity filtering of the data and type conversion of the data; the entity disambiguation part aims to eliminate different names of the same concept in the biomedical field, and unifies different names of objects describing the same biomedical field in data into UMLS language by taking UMLS codes as a reference;
the main task of the pathogenic contribution network construction and calculation module is to complete the construction of a pathogenic network and the calculation of node contribution, and establish a complete pathogenic contribution network, including the construction of the pathogenic network and the calculation of the node contribution; the pathogenic network construction aims at unifying all pathogenic elements of the target disease in the network, including node introduction and edge introduction; the node contribution calculation is to calculate and distinguish the importance of each node (pathogenic element) in the network, firstly calculate the node centrality, and then calculate the node contribution value on the basis of the node centrality;
the task of the drug embedding and sorting module is to list all the drugs becoming effective drugs of the target diseases and the possibility of each drug becoming effective drugs of the target diseases, and the drug embedding and result sorting module comprises two parts; the purpose of the medicine embedding is to calculate the potential treatment capacity of the medicine on a pathogenic contribution network in a medicine node and relationship embedding mode; the purpose of the result ranking is to screen out candidate drugs with larger network treatment capacity, including gene tendency ranking and protein tendency ranking.
2. The pathogenic contribution network analysis-based drug relocation model according to claim 1, wherein the impurity filtering of the data includes filtering out meaningless data entrained in the desired data; the type conversion of the data comprises the steps of firstly unifying multi-source heterogeneous data into triple data in a main-predicate-guest form, and then converting the triple data into a CSV format.
3. The drug relocation model according to claim 1, wherein the node importing step is to import all target disease-causing genes and proteins in the data-washed CSV data into the graph database after the target disease to be studied is selected; the side introduction is that the pathogenic gene and the pathogenic protein are taken as network nodes in the graph, and four types of relations of gene-gene, gene-protein, protein-gene and protein-protein are taken as the side of the network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910997755.6A CN110853714B (en) | 2019-10-21 | 2019-10-21 | Drug repositioning system based on pathogenic contribution network analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910997755.6A CN110853714B (en) | 2019-10-21 | 2019-10-21 | Drug repositioning system based on pathogenic contribution network analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110853714A true CN110853714A (en) | 2020-02-28 |
| CN110853714B CN110853714B (en) | 2023-04-21 |
Family
ID=69596836
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910997755.6A Active CN110853714B (en) | 2019-10-21 | 2019-10-21 | Drug repositioning system based on pathogenic contribution network analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110853714B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111768869A (en) * | 2020-09-03 | 2020-10-13 | 成都索贝数码科技股份有限公司 | Medical guide mapping construction search system and method for intelligent question-answering system |
| CN112133380A (en) * | 2020-09-24 | 2020-12-25 | 南京中爱人工智能与生命科学研究院有限公司 | Medicine research and development analysis application method based on intelligent data model and platform thereof |
| CN112331275A (en) * | 2020-10-28 | 2021-02-05 | 东南大学 | Memory network and attention-based drug relocation calculation method |
| CN114242186A (en) * | 2021-12-30 | 2022-03-25 | 湖南大学 | Chinese and western medicine relocation method and system fusing GHP and GCN and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090138251A1 (en) * | 2004-03-18 | 2009-05-28 | Andrej Bugrim | Bioinformatics research and analysis system and methods associated therewith |
| CN106529205A (en) * | 2016-11-03 | 2017-03-22 | 中南大学 | Drug target relation prediction method based on drug substructure and molecule character description information |
| CN109411033A (en) * | 2018-11-05 | 2019-03-01 | 杭州师范大学 | A kind of curative effect of medication screening technique based on complex network |
-
2019
- 2019-10-21 CN CN201910997755.6A patent/CN110853714B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090138251A1 (en) * | 2004-03-18 | 2009-05-28 | Andrej Bugrim | Bioinformatics research and analysis system and methods associated therewith |
| CN106529205A (en) * | 2016-11-03 | 2017-03-22 | 中南大学 | Drug target relation prediction method based on drug substructure and molecule character description information |
| CN109411033A (en) * | 2018-11-05 | 2019-03-01 | 杭州师范大学 | A kind of curative effect of medication screening technique based on complex network |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111768869A (en) * | 2020-09-03 | 2020-10-13 | 成都索贝数码科技股份有限公司 | Medical guide mapping construction search system and method for intelligent question-answering system |
| CN111768869B (en) * | 2020-09-03 | 2020-12-11 | 成都索贝数码科技股份有限公司 | Medical guide mapping construction search system and method for intelligent question-answering system |
| CN112133380A (en) * | 2020-09-24 | 2020-12-25 | 南京中爱人工智能与生命科学研究院有限公司 | Medicine research and development analysis application method based on intelligent data model and platform thereof |
| CN112133380B (en) * | 2020-09-24 | 2024-02-23 | 南京泛泰数字科技研究院有限公司 | Medicine research, development, analysis and application method based on intelligent data model and platform thereof |
| CN112331275A (en) * | 2020-10-28 | 2021-02-05 | 东南大学 | Memory network and attention-based drug relocation calculation method |
| CN114242186A (en) * | 2021-12-30 | 2022-03-25 | 湖南大学 | Chinese and western medicine relocation method and system fusing GHP and GCN and storage medium |
| CN114242186B (en) * | 2021-12-30 | 2022-08-12 | 湖南大学 | Chinese and western medicine relocation method and system fusing GHP and GCN and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110853714B (en) | 2023-04-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110853714A (en) | Drug relocation model based on pathogenic contribution network analysis | |
| Choobdar et al. | Assessment of network module identification across complex diseases | |
| Huang et al. | Hardware acceleration of the pair-HMM algorithm for DNA variant calling | |
| Ni et al. | Inside the atoms: ranking on a network of networks | |
| Peng et al. | Identification of protein complexes using weighted pagerank-nibble algorithm and core-attachment structure | |
| Doncheva et al. | Recent approaches to the prioritization of candidate disease genes | |
| CN109637579B (en) | Tensor random walk-based key protein identification method | |
| WO2001097076A2 (en) | Apparatus and method for providing sequence database comparison | |
| Choobdar et al. | Open community challenge reveals molecular network modules with key roles in diseases | |
| CN113012770B (en) | Multi-modal deep neural network based prediction of drug-drug interaction events | |
| Diao et al. | Building Highly-Optimized, Low-Latency Pipelines for Genomic Data Analysis. | |
| CN106503483B (en) | Myeloma signal path mechanism confirmation method based on modularization factor graph | |
| Gao et al. | Protein2vec: Aligning multiple ppi networks with representation learning | |
| CN113053457B (en) | Drug target prediction method based on multi-pass graph convolution neural network | |
| CN109712685B (en) | Prescription and medicament construction method and system based on multi-objective evolutionary algorithm | |
| Faes et al. | Scalable hardware accelerator for comparing DNA and protein sequences | |
| Taheri et al. | A high-performance memristor-based smith-waterman dna sequence alignment using fpni structure | |
| CN110534153B (en) | Target prediction system and method based on deep learning | |
| CN114882960A (en) | Medicine disease association prediction method and device, electronic equipment and readable storage medium | |
| Gossmann et al. | A sparse regression method for group-wise feature selection with false discovery rate control | |
| Guerrini et al. | Lossy compressor preserving variant calling through extended BWT | |
| Zhao et al. | Logistic regression augmented community detection for network data with application in identifying autism-related gene pathways | |
| CN109935341B (en) | Method and device for predicting new drug indication | |
| Boată et al. | Using digital twins in health care | |
| Syarofina et al. | Cluster analysis in prediction of biological activity and molecular structure relationship of dipeptidyl peptidase-4 inhibitors for the type two diabetes mellitus treatment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |