CN107133448A - A kind of metabolism group data fusion optimized treatment method - Google Patents
A kind of metabolism group data fusion optimized treatment method Download PDFInfo
- Publication number
- CN107133448A CN107133448A CN201710227598.1A CN201710227598A CN107133448A CN 107133448 A CN107133448 A CN 107133448A CN 201710227598 A CN201710227598 A CN 201710227598A CN 107133448 A CN107133448 A CN 107133448A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- metabolism group
- classification
- group data
- 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
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
- G16C20/70—Machine learning, data mining or chemometrics
-
- 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
- G16C20/10—Analysis or design of chemical reactions, syntheses or processes
Landscapes
- Chemical & Material Sciences (AREA)
- Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Theoretical Computer Science (AREA)
- Crystallography & Structural Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a kind of metabolism group data fusion optimized treatment method, comprise the following steps:1) the metabolism group data on the separate sources of multiple patients are converted into metabolite data matrix respectively;2) the optimal pre-treating method of metabolism group data of separate sources is separately optimized using experimental design method, by the optimal pre-treating method process step 1 of data after optimization) in metabolite data matrix, combining classification model, which is found out, contributes larger metabolin data;3) by step 2) find out and contribute larger metabolin data conversion into important metabolite data matrix;4) the optimal pre-treating method of data of important metabolite data matrix is optimized using experimental design method, by the optimal pre-treating method process step 3 of data after optimization) in important metabolite data matrix, combining classification model carries out parting and classification to different patient.The present invention improves the parting and classification accuracy to disease, can aid in doctor and more accurately formulate personalized therapy program for patient.
Description
Technical field
The application is related to medical data information excavating field, relates more specifically at a kind of metabolism group data fusion optimization
Reason method.
Background technology
With the arrival of accurate Medical Era, the Accurate classification of disease is for formulating the personalized therapeutic scheme with precision
It is particularly important.Metabolism group is after a kind of relatively new omics technology after genomics and protein science, the technical purpose
It is that the small molecule metabolites in biological sample must be detected more as far as possible, so as to reflect organism (such as disease under various circumstances
Evolution, medicine/diet intervention etc. occur for disease) metabolic alterations situation.Metabonomic technology can be anti-in metabolin aspect
The personalized difference of organism is reflected, therefore, the technology can realize parting and the classification of clinical disease.
Metabolism group can produce substantial amounts of data message, and these data can derive from different biological samples,
Different analysis platforms, but the data in our very important any sources, because wherein comprising our institutes can also be derived from
Need patient information.Data fusion technique can be integrated the data of separate sources, so as to realize more accurate disease
Parting and classification.
The data of separate sources take on a different character, therefore, selection data pre-processing method on also can not without exception and
By.But, presently, there are many different data pre-processing methods, how according to the optimal data of different pieces of information feature selecting before
Processing method is a hot issue in data analysis field.In addition, data fusion be not simply by data investigation together,
Important information will reduce data volume in fusion separate sources data, improve data processing speed, still, how from separate sources
It is also the problem of one of data analysis field receives much concern that important information is selected in data.
The content of the invention
To solve the above mentioned problem that there is currently, present applicant proposes a kind of metabolism group data fusion optimized treatment method
The technical solution used in the present invention is as follows:A kind of metabolism group data fusion optimized treatment method, including it is following
Step:
1) the metabolism group data on the separate sources of multiple patients are converted into multiple metabolite data matrixes respectively;
2) the optimal pre-treating method of data of the metabolism group of separate sources is separately optimized using experimental design method, passes through
The optimal pre-treating method alignment processing step 1 of data after optimization) in metabolite data matrix, combining classification model finds out contribution
Larger metabolin data;
3) by step 2) find out and contribute larger metabolin data fusion to be converted into important metabolite data matrix;
4) the optimal pre-treating method of data of important metabolite data matrix is optimized using experimental design method, after optimization
The optimal pre-treating method process step 3 of data) in important metabolite data matrix, combining classification model enters to different patients
Row parting and classification.
Preferably, step 1) in metabolism group data source be blood, urine, excrement, sweat, heart tissue, kidney
One or more in tissue, liver organization, stomach intestinal tissue, the metabolism group data pass through nuclear magnetic resonance chemical analyser, liquid
One or more in matter combined instrument, gas chromatograph-mass spectrometer, infrared spectrometer, ultraviolet spectrometer, Raman spectrometer are obtained.
Preferably, step 2) in specifically include following steps:
Step 2-1. selects the combination of different pieces of information pre-treating method by experimental design method;
Step 2-2. is respectively to step 1) obtained metabolite data matrix carried out before data by the combination in step 2-1
Processing;
The data input disaggregated model obtained after data pre-processing is set up and classified by step 2-3. by experimental design method
Relation between model performance parameter and different pieces of information pre-treating method, classification of assessment model performance analyzes different pre-treatments pair
The influence of disaggregated model performance parameter;
Step 2-4. maximizes disaggregated model performance parameter by experimental design method, selects optimum data pre-treating method
Combination;
Step 2-5. is using the optimum data pre-treating method combination obtained by step 2-4, to step 1) obtained metabolism number
Data pre-processing is carried out according to square, disaggregated model is inputted, the maximum metabolin data of contribution of classifying to disease parting are filtered out.
Preferably, step 4) in specifically include following steps:
Step 4-1. selects the combination of different pieces of information pre-treatment by experimental design method;
Step 4-2. is by step 3) obtained metabolite data matrix carries out locating before data by the combination in step 4-1
Reason;
The data input disaggregated model obtained after data pre-processing is set up and classified by step 4-3. by experimental design method
Relation between model performance parameter and different pieces of information pre-treating method, classification of assessment model performance analyzes different pre-treatments pair
The influence of disaggregated model performance parameter;
Step 4-4. is by step 3) obtained metabolite data matrix enters line number by the step 4-3 optimal pre-treatment schemes optimized
According to pre-treatment, disaggregated model is inputted, patient classification's model based on metabolism group data is set up.
Preferably, the experimental design method is response surface analysis, Mixed Design, D optimization designs, Latin―Square design, friendship
Pitch the one or more in design, matched-pair design, Factorial Design.
Preferably, the data pre-processing method be normalization, standardization, data transposition, data zooming, data smoothing,
One or more in data integration.
Preferably, disaggregated model is linear discriminant analysis model, partial least squares discriminant analysis model, artificial neural network
One or more in model, supporting vector machine model, Random Forest model, decision-tree model, Model of Fuzzy Clustering Analysis.
Preferably, classification of assessment model performance passes through fitting coefficient, estimated performance, classification accuracy, P values, subject's work
Make the one or more in feature (ROC) curve.
Preferably, each metabolin that the maximum metabolin data of contribution of classifying to disease parting are exported according to disaggregated model
Weight or contribution degree, self-defined setting threshold value screened.
Beneficial effects of the present invention are as follows:The metabolism group data of separate sources are merged, data message amount can be increased, carried
High parting and classification accuracy to disease, can aid in doctor more accurately to formulate personalized therapy program for patient.This
Invention compared with the conventional method, with advantages below:(1) data pre-processing is most important for metabolism group data analysis, but
It is to face diversified pre-treating method, attempts to waste time and energy one by one, and also there is friendship between Different front processing method
Mutual effect, therefore, is difficult to determine optimal data pre-processing scheme with traditional method.The present invention utilizes experimental design method
Different pre-treatments assembled scheme is designed, on the one hand the optimization time is saved, the interaction effect between distinct methods is on the other hand further contemplated
Should.
(2) present invention is set up and different pieces of information pre-treatment side using disaggregated model performance to be oriented to using experimental design method
Relation between method, optimum data pre-treatment scheme is determined by maximizing disaggregated model performance parameter.Compared with conventional method, this
Inventing the method proposed more accurate can more purposefully optimize metabolism group data pre-processing scheme.
(3) the pre-treatment scheme of the present invention metabolism group data of optimization separate sources first, and being sieved by disaggregated model
Important metabolin is selected, the important metabolin filtered out is finally merged.With simply by the method for separate sources data investigation compared with,
Method fusion critical data information proposed by the present invention simultaneously filters off garbage, so as to reduce data dimension, improves disaggregated model
Arithmetic speed.
(4) present invention also optimizes the optimal pre-treatment scheme of metabolism group data after fusion, further improves classification
The performance of model.
(5) simultaneously it can develop corresponding software to implement according to the present invention, edit routine.
Brief description of the drawings
Fig. 1 is the schematic flow diagram of metabolism group data fusion optimized treatment method proposed by the invention.
Fig. 2 for separate sources metabolism group data conversion into metabolite data matrix unify form.
Fig. 3 is the selection of the important metabolin based on disaggregated model, wherein metabolin of the contribution angle value more than 2.0 is used as weight
Want metabolin.
Embodiment
Embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.
Fig. 1 shows a kind of schematic flow diagram of metabolism group data fusion optimized treatment method proposed by the invention.Ginseng
Fig. 1 is examined, this method includes:
Step 1:Metabolism group data on the separate sources of multiple patients are converted into matrix respectively.Separate sources
Metabolism group data can be derived from different biological samples, such as blood, urine, excrement, sweat, heart tissue, kidney
Tissue, liver organization, stomach intestinal tissue or from different analytical technologies, such as nuclear magnetic resonance chemical analyser, liquid matter
Combined instrument, gas chromatograph-mass spectrometer, infrared spectrometer, ultraviolet spectrometer, Raman spectrometer.
But, these data will be converted into unified matrix format respectively, as shown in Fig. 2 being a patient per a line
Relevant information, wherein first is classified as patient number, second is classified as patient's packet, is afterwards metabolin data.
Step 2:The optimal pre-treating method of metabolism group data of separate sources is separately optimized using experimental design method, ties
Close disaggregated model and find out the larger metabolin data of contribution.Experimental design method can select not Tongfang according to specific actual conditions
Method, such as response surface analysis, Mixed Design, D optimization designs, Latin―Square design, cross-over design, matched-pair design, Factorial Design;Separately
Outside, data pre-processing method can also be used according to actual conditions selection, such as normalization, standardization, data transposition, data contracting
Put, data smoothing, data integration.
Step 2 specifically includes following steps:
Step 2-1:The combination of different pieces of information pre-treatment is selected by experimental design method.For example, 3 kinds of differences of optimization
Data pre-processing method assembled scheme, every kind of pre-treating method considers 3 kinds of different calculations, that is, normalizes (A1, B1
And C1), data transposition (A2, B2 and C2) and data zooming (A3, B3 and C3).Using traditional orthogonal experiment, 27 kinds of differences
Assembled scheme need to be attempted, if but from experimental design method, such as D optimization designs, need to only attempt 7 times it is different
Data pre-processing assembled scheme, as shown in table 1.
The combination of the different pieces of information pre-treatment of table 1
Pre-treatment is combined | Normalize (F1) | Data transposition (F2) | Data zooming (F3) | Model performance parameter (P) |
1 | C1 | B2 | A3 | N1 |
2 | C1 | A2 | B3 | N2 |
3 | B1 | C2 | C3 | N3 |
4 | B1 | B2 | B3 | N4 |
5 | B1 | A2 | A3 | N5 |
6 | A1 | C2 | B3 | N6 |
7 | A1 | A2 | C3 | N7 |
Step 2-2:The metabolism group data to separate sources carry out data pre-processing in the way of in step 2-1 respectively.
Step 2-3:By the data input disaggregated model after pre-treatment, analysis different pre-treatments are to disaggregated model performance parameter
Influence.As shown in table 1, the data treated through different pieces of information pre-treatment assembled scheme can produce different disaggregated model performances
Parameter.For example, by experimental design method set up disaggregated model performance parameter (P) and different pieces of information pre-treating method (F1, F2 and
F3 the relation between), as shown in Equation 1, wherein α, β, γ, δ, θ and μ represent model coefficient, and ε represents model residual error, F1F2,
F1F3 and F2F3 represent the interaction between different pieces of information pre-treating method, and the data pre-processing method in the formula can
To be the multiple combinations in normalization, standardization, data transposition, data zooming, data smoothing, data integration.Disaggregated model
It can be used according to actual conditions selection, such as linear discriminant analysis model, partial least squares discriminant analysis model, artificial neuron
Network model, supporting vector machine model, Random Forest model, decision-tree model, Model of Fuzzy Clustering Analysis;In addition, model
Can evaluate can select fitting coefficient, estimated performance, classification accuracy, P values, Receiver Operating Characteristics (ROC) curve.
Formula 1
P=α F1+ β F2+ γ F3+ δ (F1F2)+θ (F1F3)+μ (F2F3)+ε
Step 2-4:Disaggregated model performance parameter is maximized by experimental design method, optimum data pre-treating method is selected
Combination, as shown in Equation 2, wherein α, β, γ, δ, θ and μ represent model coefficient, and ε represents model residual error, F1F2, F1F3 and
F2F3 represents the interaction between different pieces of information pre-treating method, and the data pre-processing method in the formula can be normalizing
Multiple combinations in change, standardization, data transposition, data zooming, data smoothing, data integration, Max represents that maximization operation is ordered
Order.
Formula 2
Max (P)=α P1+ β P2+ γ P3+ δ (P1P2)+θ (P1P3)+μ (P2P3)+ε
Step 2-5:Metabolism group data are entered by the optimum data pre-treating method combination obtained by being optimized using step 2-4
Row pre-treatment, inputs disaggregated model, filters out the maximum metabolin data of contribution of classifying to disease parting.Contribute maximum metabolism
The weight or contribution degree for each metabolin that thing data can be exported according to disaggregated model, self-defined setting threshold value are sieved
Choosing.For example, as shown in figure 3, contribution angle value be more than 2.0 metabolin be screened as important metabolin, carry out next step number
According to fusion.
Step 3:By the important metabolin data conversion filtered out from the metabolism group data of separate sources into same
Matrix.
Step 4:The optimal pre-treating method of metabolism group data after experimental design method optimization fusion is reused, with reference to
Disaggregated model carries out parting and classification to different patients, and formulating personalized therapeutic strategy for doctor provides reference, specific bag
Include following steps:
Step 4-1:The combination of different pieces of information pre-treatment is selected by experimental design method, with step 2-1.
Step 4-2:Metabolism group data after fusion are subjected to data pre-processing in the way of step 4-1.
Step 4-3:By the data input disaggregated model after pre-treatment, and determined by experimental design method before optimum data
Processing scheme, with step 2-3.
Step 4-4:Metabolism group data after fusion are carried out before data by the step 4-3 optimal pre-treatment schemes optimized
Processing, inputs disaggregated model, sets up patient classification's model based on metabolism group data.
Those skilled in the art simultaneously can develop corresponding software to implement according to the above method, edit routine.
The metabolism group data of separate sources are merged, data message amount can be increased, the parting to disease and classification is improved
Accuracy, can aid in doctor and more accurately formulate personalized therapy program for patient.Embodiments of the present invention with it is existing
Method is compared, with advantages below:
Advantage 1:Data pre-processing is most important for metabolism group data analysis, but in face of diversified pre-treatment
Method, attempts to waste time and energy one by one, and also there is interaction between Different front processing method, therefore, with traditional
Method is difficult to determine optimal data pre-processing scheme.The present invention utilizes experimental design method design different pre-treatments combination side
Case, on the one hand saves the optimization time, on the other hand further contemplates the interaction between distinct methods.
Advantage 2:The present invention is set up and different pieces of information pre-treatment using disaggregated model performance to be oriented to using experimental design method
Relation between method, optimum data pre-treatment scheme is determined by maximizing disaggregated model performance parameter.Compared with conventional method,
Method proposed by the present invention more accurate can more purposefully optimize metabolism group data pre-processing scheme.
Advantage 3:The pre-treatment scheme of the present invention metabolism group data of optimization separate sources first, and pass through disaggregated model
Important metabolin is screened, the important metabolin filtered out is finally merged.With simply by the method phase of separate sources data investigation
Than method fusion critical data information proposed by the present invention simultaneously filters off garbage, so as to reduce data dimension, improves classification mould
The arithmetic speed of type.
Advantage 4:The present invention also optimizes the optimal pre-treatment scheme of metabolism group data after fusion, further improves and divides
The performance of class model.
Although it will be understood to those skilled in the art that the application is to illustrate this hair by describing embodiment
It is bright, but the specific method of some in embodiment be it is nonrestrictive, such as experimental design method, pre-treating method, classification
Model, model performance evaluation criterion, metabolin screening criteria etc., can modify and replace according to real needs and condition.
Claims (9)
1. a kind of metabolism group data fusion optimized treatment method, it is characterised in that comprise the following steps:
1) the metabolism group data on the separate sources of multiple patients are converted into multiple metabolite data matrixes respectively;
2) the optimal pre-treating method of metabolism group data of separate sources is separately optimized using experimental design method, after optimization
The optimal pre-treating method alignment processing step 1 of data) in metabolite data matrix, combining classification model find out contribution it is larger
Metabolin data;
3) by step 2) find out and contribute larger metabolin data fusion to be converted into important metabolite data matrix;
4) optimize the optimal pre-treating method of data of important metabolite data matrix using experimental design method, pass through the number after optimization
According to optimal pre-treating method process step 3) in important metabolite data matrix, combining classification model divided different patients
Type and classification.
2. metabolism group data fusion optimized treatment method according to claim 1, it is characterised in that:Step 1) middle metabolism
The source that group learns data can be blood, urine, excrement, sweat, heart tissue, renal tissue, liver organization, stomach intestinal tissue
In one or more, the metabolism group data can also be by nuclear magnetic resonance chemical analyser, LC-MS instrument, gas chromatography mass spectrometry
One or more in instrument, infrared spectrometer, ultraviolet spectrometer, Raman spectrometer are obtained.
3. metabolism group data fusion optimized treatment method according to claim 1, it is characterised in that step 2) in it is specific
Comprise the following steps:
Step 2-1. selects the combination of different pieces of information pre-treating method by experimental design method;
Step 2-2. is respectively to step 1) obtained metabolite data matrix carries out locating before data by the combination in step 2-1
Reason;
The data input disaggregated model obtained after data pre-processing is set up disaggregated model by step 2-3. by experimental design method
Relation between performance parameter and different pieces of information pre-treating method, classification of assessment model performance, analysis different pre-treatments are to classification
The influence of model performance parameter;
Step 2-4. maximizes disaggregated model performance parameter by experimental design method, selects optimum data pre-treating method group
Close;
Step 2-5. is using the optimum data pre-treating method combination obtained by step 2-4, to step 1) obtained metabolite data square
Data pre-processing is carried out, disaggregated model is inputted, the maximum metabolin data of contribution of classifying to disease parting are filtered out.
4. metabolism group data fusion optimized treatment method according to claim 1, it is characterised in that step 4) in it is specific
Comprise the following steps:
Step 4-1. selects the combination of different pieces of information pre-treatment by experimental design method;
Step 4-2. is by step 3) obtained metabolite data matrix carries out data pre-processing by the combination in step 4-1;
The data input disaggregated model obtained after data pre-processing is set up disaggregated model by step 4-3. by experimental design method
Relation between performance parameter and different pieces of information pre-treating method, classification of assessment model performance, analysis different pre-treatments are to classification
The influence of model performance parameter;
Step 4-4. is by step 3) obtained metabolite data matrix carried out before data by the step 4-3 optimal pre-treatment schemes optimized
Processing, inputs disaggregated model, sets up patient classification's model based on metabolism group data.
5. the metabolism group data fusion optimized treatment method according to claim 3 or 4, it is characterised in that:The experiment
Design method is response surface analysis, Mixed Design, D optimization designs, Latin―Square design, cross-over design, matched-pair design, Factorial Design
In one or more.
6. the metabolism group data fusion optimized treatment method according to claim 3 or 4, it is characterised in that:The data
Pre-treating method is the one or more in normalization, standardization, data transposition, data zooming, data smoothing, data integration.
7. the metabolism group data fusion optimized treatment method according to claim 3 or 4, it is characterised in that:Disaggregated model
For linear discriminant analysis model, partial least squares discriminant analysis model, artificial nerve network model, supporting vector machine model, with
One or more in machine forest model, decision-tree model, Model of Fuzzy Clustering Analysis.
8. the metabolism group data fusion optimized treatment method according to claim 3 or 4, it is characterised in that:Classification of assessment
Model performance passes through one kind or many in fitting coefficient, estimated performance, classification accuracy, P values, Receiver operating curve
Kind.
9. metabolism group data fusion optimized treatment method according to claim 3, it is characterised in that:To disease parting point
The weight or contribution degree for each metabolin that the maximum metabolin data of class contribution are exported according to disaggregated model, self-defined setting
Threshold value is screened.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710227598.1A CN107133448B (en) | 2017-04-10 | 2017-04-10 | Metabonomics data fusion optimization processing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710227598.1A CN107133448B (en) | 2017-04-10 | 2017-04-10 | Metabonomics data fusion optimization processing method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107133448A true CN107133448A (en) | 2017-09-05 |
CN107133448B CN107133448B (en) | 2020-05-01 |
Family
ID=59716649
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710227598.1A Active CN107133448B (en) | 2017-04-10 | 2017-04-10 | Metabonomics data fusion optimization processing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107133448B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109187614A (en) * | 2018-09-27 | 2019-01-11 | 厦门大学 | Based on nuclear magnetic resonance and mass spectrographic metabolism group data fusion method and its application |
CN109300502A (en) * | 2018-10-10 | 2019-02-01 | 汕头大学医学院 | A kind of system and method for the analyzing and associating changing pattern from multiple groups data |
CN111768367A (en) * | 2020-05-20 | 2020-10-13 | 深圳迈瑞生物医疗电子股份有限公司 | Data processing method, device and storage medium |
CN111768366A (en) * | 2020-05-20 | 2020-10-13 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging system, BI-RADS classification method and model training method |
CN111933281A (en) * | 2020-09-30 | 2020-11-13 | 平安科技(深圳)有限公司 | Disease typing determination system, method, device and storage medium |
CN113376199A (en) * | 2021-05-12 | 2021-09-10 | 兰立生物科技(苏州)有限公司 | Biochemical analysis and detection system for cancer detection |
CN114664382A (en) * | 2022-04-28 | 2022-06-24 | 中国人民解放军总医院 | Multi-group association analysis method and device and computing equipment |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101727455A (en) * | 2008-10-24 | 2010-06-09 | 国际商业机器公司 | Method and device for optimizing configuration parameters of database consultant program |
CN101769910A (en) * | 2008-12-30 | 2010-07-07 | 中国科学院大连化学物理研究所 | Method for screening malignant ovarian tumor markers from blood serum metabolic profiling |
CN102103132A (en) * | 2009-12-18 | 2011-06-22 | 中国科学院大连化学物理研究所 | Method for screening diabetes markers from body fluid metabonome profile |
CN102100706A (en) * | 2009-12-18 | 2011-06-22 | 中国科学院大连化学物理研究所 | Method for evaluating quality of Chinese patent medicament by using metabonomics |
CN102323285A (en) * | 2010-11-15 | 2012-01-18 | 上海聚类生物科技有限公司 | Method for analyzing NMR (Nuclear Magnetic Resonance) metabonomics detection data |
US20140276173A1 (en) * | 2013-03-14 | 2014-09-18 | Convergent Engineering, Inc. | Methods and systems for monitoring resistance and work of breathing for ventilator-dependent patients |
CN104063631A (en) * | 2014-06-13 | 2014-09-24 | 周家锐 | Big data-oriented metabolome characteristic data analysis method and system thereof |
CN105718744A (en) * | 2016-01-25 | 2016-06-29 | 深圳大学 | Metabolic mass spectrum screening method and system based on the deep learning |
-
2017
- 2017-04-10 CN CN201710227598.1A patent/CN107133448B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101727455A (en) * | 2008-10-24 | 2010-06-09 | 国际商业机器公司 | Method and device for optimizing configuration parameters of database consultant program |
CN101769910A (en) * | 2008-12-30 | 2010-07-07 | 中国科学院大连化学物理研究所 | Method for screening malignant ovarian tumor markers from blood serum metabolic profiling |
CN102103132A (en) * | 2009-12-18 | 2011-06-22 | 中国科学院大连化学物理研究所 | Method for screening diabetes markers from body fluid metabonome profile |
CN102100706A (en) * | 2009-12-18 | 2011-06-22 | 中国科学院大连化学物理研究所 | Method for evaluating quality of Chinese patent medicament by using metabonomics |
CN102323285A (en) * | 2010-11-15 | 2012-01-18 | 上海聚类生物科技有限公司 | Method for analyzing NMR (Nuclear Magnetic Resonance) metabonomics detection data |
US20140276173A1 (en) * | 2013-03-14 | 2014-09-18 | Convergent Engineering, Inc. | Methods and systems for monitoring resistance and work of breathing for ventilator-dependent patients |
CN104063631A (en) * | 2014-06-13 | 2014-09-24 | 周家锐 | Big data-oriented metabolome characteristic data analysis method and system thereof |
CN105718744A (en) * | 2016-01-25 | 2016-06-29 | 深圳大学 | Metabolic mass spectrum screening method and system based on the deep learning |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109187614A (en) * | 2018-09-27 | 2019-01-11 | 厦门大学 | Based on nuclear magnetic resonance and mass spectrographic metabolism group data fusion method and its application |
CN109187614B (en) * | 2018-09-27 | 2020-03-06 | 厦门大学 | Metabonomics data fusion method based on nuclear magnetic resonance and mass spectrum and application thereof |
CN109300502A (en) * | 2018-10-10 | 2019-02-01 | 汕头大学医学院 | A kind of system and method for the analyzing and associating changing pattern from multiple groups data |
CN111768367A (en) * | 2020-05-20 | 2020-10-13 | 深圳迈瑞生物医疗电子股份有限公司 | Data processing method, device and storage medium |
CN111768366A (en) * | 2020-05-20 | 2020-10-13 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging system, BI-RADS classification method and model training method |
CN111768367B (en) * | 2020-05-20 | 2024-03-29 | 深圳迈瑞生物医疗电子股份有限公司 | Data processing method, device and storage medium |
CN111933281A (en) * | 2020-09-30 | 2020-11-13 | 平安科技(深圳)有限公司 | Disease typing determination system, method, device and storage medium |
CN113376199A (en) * | 2021-05-12 | 2021-09-10 | 兰立生物科技(苏州)有限公司 | Biochemical analysis and detection system for cancer detection |
CN114664382A (en) * | 2022-04-28 | 2022-06-24 | 中国人民解放军总医院 | Multi-group association analysis method and device and computing equipment |
CN114664382B (en) * | 2022-04-28 | 2023-01-31 | 中国人民解放军总医院 | Multi-group association analysis method and device and computing equipment |
Also Published As
Publication number | Publication date |
---|---|
CN107133448B (en) | 2020-05-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107133448A (en) | A kind of metabolism group data fusion optimized treatment method | |
Hutter et al. | The cancer genome atlas: creating lasting value beyond its data | |
Keppler et al. | Advances in the application of comprehensive two-dimensional gas chromatography in metabolomics | |
Khan et al. | Classification and diagnostic prediction of cancers using gene expression profiling and artificial neural networks | |
Mishra et al. | A precise analysis of deep learning for medical image processing | |
Fenn et al. | Raman spectroscopy utilizing Fisher‐based feature selection combined with Support Vector Machines for the characterization of breast cell lines | |
Ahmed et al. | Early Alzheimer's Disease Detection Using Different Techniques Based on Microarray Data: A Review. | |
Tang et al. | Research of insomnia on traditional Chinese medicine diagnosis and treatment based on machine learning | |
Matos et al. | Research techniques made simple: mass cytometry analysis tools for decrypting the complexity of biological systems | |
Basar et al. | Identification of drug and protein-protein interaction network among stress and depression: a bioinformatics approach | |
Qin et al. | MLP-based regression prediction model for compound bioactivity | |
Kim et al. | Controlling the false discovery rate for feature selection in high‐resolution NMR Spectra | |
Revanth Kumar et al. | A deep learning-based approach for an automated brain tumor segmentation in MR images | |
EP3891761A1 (en) | Integrated system and method for personalized stratification and prediction of neurodegenerative disease | |
Wang et al. | Predicting polypharmacy side effects based on an enhanced domain knowledge graph | |
CN115662504A (en) | Multi-angle fusion-based biological omics data analysis method | |
Li et al. | Adolescent fluid intelligence prediction from regional brain volumes and cortical curvatures using BlockPC-XGBoost | |
Wang et al. | AdaTiSS: a novel data-Ada ptive robust method for identifying Ti ssue S pecificity S cores | |
Hilario et al. | Data mining for mass-spectra based diagnosis and biomarker discovery | |
Wang et al. | Principal Component Regression Analysis for lncRNA-Disease Association Prediction Based on Pathological Stage Data | |
Kumar et al. | A Review on Cancer Dataset Classification using Data Mining Methods | |
Fang et al. | Rapid diagnosis of type II diabetes using fourier transform mid-infrared attenuated total reflection spectroscopy combined with support vector machine | |
Jabbari et al. | New conditional generative adversarial capsule network for imbalanced classification of human sperm head images | |
Reddy et al. | Real-time data mining-based cancer disease classification using KEGG gene dataset | |
Koleini et al. | Biomarker discovery in multi-omics datasets using tensor decompositions; A comprehensive review |
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 |