CN106934252A - A kind of triple net Resources Spread method - Google Patents
A kind of triple net Resources Spread method Download PDFInfo
- Publication number
- CN106934252A CN106934252A CN201710133230.9A CN201710133230A CN106934252A CN 106934252 A CN106934252 A CN 106934252A CN 201710133230 A CN201710133230 A CN 201710133230A CN 106934252 A CN106934252 A CN 106934252A
- Authority
- CN
- China
- Prior art keywords
- mirna
- node
- resource
- lncrna
- matrix
- 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.)
- Pending
Links
Classifications
-
- 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
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
-
- 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
- G16B5/00—ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Theoretical Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Biotechnology (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Bioethics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Epidemiology (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Physiology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of triple net Resources Spread method.It is applied in bioinformatics, in the research process of environmental factor and long non-coding RNA incidence relation.By combining the resource transfers in lncRNAs miRNAs and miRNAs EFs related networks and network, the present invention proposes a kind of network inference method to infer the potential association between lncRNAs and EFs.Experiment analysis results of the present invention show that the method proposed with other, the present invention can predict more reliable lncRNAs and EFs relations.These results demonstrate the ability that our method predicts biologically significant association, and this can cause to more fully understand molecular process.
Description
Technical field
The present invention relates to bioinformatics, more particularly to a kind of triple net Resources Spread method.LncRNA can be obtained
Incidence relation and environmental factor between.
Background technology
The variation of biological character is congenital or posteriori arguement, mainly due to variation be caused by hereditary difference or
The arguement that environmental difference causes.Current main science viewpoint thinks phenotypic difference, is not by single hereditary difference or ring
Border difference is produced, but is influenced each other by the two, is together decided on.It means that phenotype and disease are considered as inherent cause
(GFs) interaction relationship complicated and environmental factor (EFs) between is determined.To this day, it has been recognized that almost institute
Some diseases are all the results from interaction complicated between individual genomic constitution and each residing environment.Some are common
Human body diseases caused by interaction relationship complicated between GEs and EFs, such as cancer, heart disease, alzheimer '
Mo's disease, and diabetes.
Look back in the challenge for designing and run into researching and analysing, preferably to find mutual between genetic and environmental factor
Interactively, and to provide better method in public health and clinical practice.Muriel Koehl have studied in adult neural
In system, the dynamic interaction of genes and environmental factor.Su et al. has carried out case-control study to assess RAGE genes
Variation and environment in influence of the carcinogen to OSCC.Yang et al., manual have collected reliably can be with
The experimental data of miRNA-EF incidence relations is supported, and establishes miREnvironment databases.Qiu et al. is analyzed
MiRNA-EF incidence relations pair relevant with human diseases in miREnvironment databases, and by analyzing miRNA-EF phases
Mutual pattern obtains Main Conclusions.It is assumed that the disease given in miRNA-EFs incidence relations research based on disease,
Similar miRNAs (EFs) often produces interaction relationship with similar EFs (miRNAs).Chen be based on this it is assumed that
A new calculating is established using type Random Walk Algorithm (random walk with restart, RWR) is restarted
The research framework of miRNA-EFs incidence relations, miREFRWR[14].However, with gene compared with miRNA, lncRNA and environment because
The complete research system of neither one is gone back in the incidence relation research of element so far, it is desirable to be able to the method for predicting lncRNA-EFs relations
To instruct the experimental study direction of lncRNA-EFs.
With reference to genes/miRNA-EFs incidence relation Forecasting Methodologies, have by hypergeometric distribution, between calculating biological information
The discrete probabilistic of correlation degree determines incidence relation；Least square method is utilized, by calculating the minimum on biological information network
Change cost functions to obtain the optimal classification of incidence relation；Have using type random walk method is restarted, one is set up to biological information
The association migration network for magnifying, finds the incidence relation of each biological node by way of converged paths；Utilize machine
The sorting technique of study, such as SVM, decision tree and Spectral Clustering, it is classified together by the close node of incidence relation；Or
Person utilizes dissemination method, is iterated according to probability in network of personal connections, determines correlation degree.Other include neutral net, heat
Diffusion also has the various methods of text mining etc., all for the prediction to gene-EFs incidence relations in the middle of.
Due to being set up currently without a complete reliable lncRNA-EFs linked database, we can select
The data predicted as lncRNA-EFs are simultaneously few, limit the selection of Forecasting Methodology.
The pass between lncRNA and environmental factor is predicted by using the triple net in machine learning in this research
Connection relation.This research conclusion lncRNA that may related multipair to 10,000 and environmental factor incidence relation have carried out possibility row
Sequence, this can be directed to the larger lncRNA of possibility to playing a part of guidance during biologist's research ring afterwards
With environmental factor carry out experiment sequencing, it is to avoid the sequencing of blindness is compared, and reduces workload.
The content of the invention
Present invention aim at the deficiency for solving above-mentioned prior art presence, there is provided a kind of triple net Resources Spread side
Method, the present invention can be directed to the larger lncRNA of possibility and environmental factor and carry out experiment sequencing, it is to avoid the sequencing ratio of blindness
It is right, reduce workload.The present invention is, based on the basis of two subnetwork models, two two subnetwork figures to be associated together, group
Into a triple net transferring structure, the network of personal connections set up between node, based on this network of personal connections acquisition pass between points
Connection relation.
In real life, if two nodes do not have reliable direct correlation relation, it is necessary to we are common by both
Third party's node predict the degree of association of the two.In order to solve this problem, in the definition of bipartite graph, one can be defined
The triple net figure of individual resource transfer, this triple net figure is connected two two subnetworks by a common intermediate node
Pick up and.Define one and possess three undirected graph G=(V1, E12, V2, E23, V3) of node, wherein Eij is tie point
Side between Vi, Vj, form such as Fig. 2.Comprising two networks of level in triple net, for the bipartite graph of each level,
New interaction relationship can be extrapolated according to priori, for bipartite graph G1=(V1, E12, V2), it is believed that bag
Projection containing both of which, V1 projection and V2 projection, wherein V1 projection represent be for non-directed graph G1, node from V1,
The corresponding node in V2 is communicated to by side E12；V2 projections also have identical definition.The pattern of this projection, it is possible to regard knot as
Resource, transmission path between point, the weights that resource is finally superimposed out as judge the weight of relevance.
We have proposed the invention of this algorithm of triple net.In triple net, the process of whole Resources Spread can be seen
It is made in the transmittance process in two two subnetworks of level.In bipartite graph G1, resource by the projection on G1, by resource from
V2 is delivered to V1, then is passed back to V2, and the weights for obtaining are namely based on the associated weights of two subnetwork G1；In G2, resource is first
V2 is delivered to from V3 points, and the final weights as V2 points are merged with the associated weight value obtained in ground floor, this weight is led to again
The projection for crossing V2 in G2 figures passes to V3 points.By this process, the weights of the V3 points for obtaining, as V1 and V3 incidence relations
Weight.
The present invention is achieved through the following technical solutions：
A kind of triple net Resources Spread method, comprises the following steps：
Two two subnetworks are connected by common intermediate node, a triple net is formed, resource is in three nodes
Between constantly transmission superposition, finally obtain the weights relation of initial point and finish node predicting and be mutually related between node journey
Degree；
Triple net refers to that two networks are integrated by common intermediate node, and on new network, resource has direction
Transmission superposition；Being superimposed the process of transmission can regard the two two points processes propagated with last resource consolidation as.
The triple net Resources Spread method is achieved by the steps of：
It is in fact process that resource mutually shifts superposition in individual node in the TRANSFER MODEL of triple net；Handle well
Bipartite graph Nlm=(Vl,Vm,Elm) and Nme=(Vm,Ve,Eme) adjacency matrix is built respectivelyWith
Wherein in bipartite graph NlmIn, ifWithIt is interrelated, thenOtherwiseTwo
Component NmeIn also have similar definition, ifWith It is interrelated, thenOtherwise
The process of the resource transfers of triple net resource transfer method, really will be projected in tripartite on a bipartite graph
Weighting procedure on network of network, i.e., based on middle miRNA, the unilateral projection on lncRNA-miRNA and miRNA-EF networks
Weighting；Therefore three parts are divided into, i.e., calculate unilateral projections of the miRNA on two two subnetworks, and joint respectively
Two projections constitute final recommendation matrix.Process is as follows：
In lncRNA-miRNA related networks Nlm=(Vl,Vm,Elm) in, resource is first from node VmTurn to Vl, then resource
Node V is transferred back to againm, this process can obtain miRNA in NlmOn resource projection matrixIt is defined as
WhereinWithRepresent in bipartite graph NlmIn, miRNAs nodesWithIn vectorial VmDegree, it is sameRepresent in bipartite graph NlmIn, lncRNA nodesIn vectorial VlDegree, nlRepresent in NlmThe quantity of middle lncRNA, and join
Number λ1∈[0,1]；
The N on miRNA-EF related networksme=(Vm,Ve,Eme), the transmission of resource is from node VeStart to be delivered to node Vm,
Then transfer back to V againeProcess；Final weights transfer matrix is defined as
WhereinWithRepresent in bipartite graph NmeIn, EF nodesWithIn vectorial VeDegree, it is sameTable
Show in bipartite graph NmeIn, miRNA nodesIn vectorial VmDegree, nmRepresent in NmeThe quantity of middle miRNA, and parameter lambda2∈
[0,1]；
Wherein parameter lambda1And λ2Weight for adjusting Resources Spread matrix；When λ value is closer to 0, the resource quilt of node is represented
Be calculated as the average value of adjacent node, and closer to 1, represent resource be distributed in neck connects node it is more uniform；In prediction, parameter
Closer to 0, predicted value is more conservative, and is worth close to 1, and last predicting the outcome more tend to and integrally predict the outcome；
The weight matrix W that above-mentioned two step is obtainedm, WeWith adjacency matrix AmeJoint, can obtain weight matrix After joint weight matrix will have been obtained, with bipartite graph NlmNeighbour
Connect matrix AlmCombine again, final forecast power matrixCan be expressed as
Wherein ri,jRepresentWithAssociated weights；In the recommendation matrix for finally giving, each node generation
The correlation degree of table lncRNA and EF, weights are higher, and the two association possibility is bigger.
Relative to prior art, advantage of the present invention and effect are：
Predict that the association between lncRNA and environmental factor is closed by using the triple net in machine learning in invention
System.The present invention lncRNA that may related multipair to 10,000 and environmental factor incidence relation have carried out possibility sequence, this opposite
Play a part of guidance during thing scholar research ring afterwards, the larger lncRNA of possibility and environmental factor can be directed to
Carry out experiment sequencing, it is to avoid the sequencing of blindness is compared, and reduces workload.
In sum, present invention illustrates a new triple net resource transfer method, and it is applied to biology
In informatics, in the research process of environmental factor and long non-coding RNA incidence relation.By combine lncRNAs-miRNAs and
Resource transfers in miRNAs-EFs related networks and network, it is proposed that a kind of network inference method is inferred
Potential association between lncRNAs and EFs.
Triple net resource transfer method of the present invention can predict more reliable lncRNAs and EFs relations.
Brief description of the drawings
Fig. 1 is the algorithm flow chart of triple net resource transfer.
Fig. 2 is that algorithm is implemented after specific lncRNA-miRNA and miRNA-Efs associated datas, the lncRNA that obtains and
The topological network of Efs incidence relations.
Fig. 3 is that algorithm is implemented after specific lncRNA-miRNA and miRNA-Efs associated datas, the lncRNA that obtains and
The degree distribution of the topological network of Efs incidence relations, by spending distribution come the reasonability of test organisms network.
Specific embodiment
With reference to Fig. 1 to Fig. 3, the present invention is further described.
It is in fact process that resource mutually shifts superposition in individual node in the TRANSFER MODEL of triple net.To handle well
Bipartite graph Nlm=(Vl,Vm,Elm) and Nme=(Vm,Ve,Eme) adjacency matrix is built respectivelyWithWherein in bipartite graph NlmIn, ifWithIt is interrelated, thenOtherwiseIn bipartite graph NmeIn also have similar definition, ifWithIt is interrelated, thenOtherwise
The process of the resource transfers of triple net resource transfer algorithm, actual is exactly that will be projected in three on a bipartite graph
Weighting procedure on square network of network, i.e., based on middle miRNA, the unilateral projection on lncRNA-miRNA and miRNA-EF networks
Weighting.Therefore algorithm is divided into three parts, i.e., calculate unilateral projections of the miRNA on two two subnetworks respectively, with
And two projections of joint constitute final recommendation matrix.Process is as shown below：
In lncRNA-miRNA related networks Nlm=(Vl,Vm,Elm) in, resource is first from node VmTurn to Vl, then resource
Node V is transferred back to againm.This process can obtain miRNA in NlmOn resource projection matrixIt is defined as
WhereinWithRepresent in bipartite graph NlmIn, miRNAs nodesWithIn vectorial VmDegree, it is sameRepresent in bipartite graph NlmIn, lncRNA nodesIn vectorial VlDegree, nlRepresent in NlmThe quantity of middle lncRNA, and join
Number λ1∈[0,1]。
The N on miRNA-EF related networksme=(Vm,Ve,Eme), the transmission of resource is from node VeStart to be delivered to node Vm,
Then transfer back to V againeProcess.Final weights transfer matrix is defined as
WhereinWithRepresent in bipartite graph NmeIn, EF nodesWithIn vectorial VeDegree, it is sameTable
Show in bipartite graph NmeIn, miRNA nodesIn vectorial VmDegree, nmRepresent in NmeThe quantity of middle miRNA, and parameter lambda2∈
[0,1]。
Wherein parameter lambda1And λ2Weight for adjusting Resources Spread matrix.When λ value is closer to 0, the resource quilt of node is represented
Be calculated as the average value of adjacent node, and closer to 1, represent resource be distributed in neck connects node it is more uniform.In prediction, parameter
Closer to 0, predicted value is more conservative, and is worth close to 1, and last predicting the outcome more tend to and integrally predict the outcome.
By the weight matrix W that two steps are obtained abovem, WeWith adjacency matrix AmeJoint, can obtain weight matrix After joint weight matrix will have been obtained, with bipartite graph NlmNeighbour
Connect matrix AlmCombine again, final forecast power matrixCan be expressed as
Wherein ri,jRepresentWithAssociated weights.In the recommendation matrix for finally giving, each node is represented
The correlation degree of lncRNA and EF, weights are higher, and the two association possibility is bigger.
In the foundation of model, it is intended that obtain the incidence relation between lncRNA and EF.Hypothesis based on ceRNA and
Experiment support, and in view of the interaction incidence relation between miRNA and lncRNA and EF, it is proposed that a prediction
The New Algorithm Model of lncRNA-EF associations.
Accumulative research shows, almost all of bio-networks such as metabolism network, protein-protein interaction network,
The node degree of protein domain network, gene interaction, gene expression network etc. follows power law distribution, p (x)~cx-k。R2With it is equal
Square error (root mean squared error, RMSE) is used to weigh predicts the lncRNA-EF related networks for obtaining to power
Restrain the degree of agreement of distribution.
Algorithm is applied to based on the interaction contact between lncRNA and miRNA, using having studied what is obtained
MiRNA-EFs associated datas and lncRNA-miRNA related networks build tripartite's resource transfer topological diagram.Original
LncRNA-miRNA related networks can be downloaded from starbase V2.0 and obtained, and the database provides a large amount of by extensive
CLIP-Seq sequencing experiments are obtained, most comprehensive lncRNA-miRNA associated datas.MiRNA-EFs association network from
MiREnvironment databases obtain, the miRNA-EFs related informations in this database be by PUBMED
Experiments prove that document, text mining arranges and obtains.
In paper, there are three parameters to have an impact the result predicted, including matrix is recommended in two lambda parameters and last interception
The value α of weight.By combining different λ1=0.1,0.3 ..., 0.9, λ2=0.1,0.3 ..., 0.9 and α=100,150 ...,
500, the different lncRNA-EF prediction incidence matrix that will be obtained utilize R2Assessed with root-mean-square error.Result discovery, with
The increase of threshold alpha, lncRNA-EF prediction associations are to reducing, and R2Increased trend is presented.For example, when α=500, it is most of
R2Value is higher than 0.8, R in this value item more less than those α2Maximum is also big.For parameter lambda1Or λ2=0.9, it is most of
R20.5, folding to be less than it is meant that in these, lncRNA-EF prediction related networks do not follow the degree distribution of power rate.
Finally, combine lncRNA-EF associations that all parameters correspondence obtains to number and corresponding power-law curve intend
Conjunction value, it has been found that get 150, λ when the threshold value of matrix R is recommended1=0.3, λ2When=0.1, the lncRNA-EF networks of personal connections for obtaining
The most compound power-law curve of network.Now, we have obtained 1086 lncRNAs and 326 the 8148 of EFs lncRNA-EF association
It is right.
Claims (2)
1. a kind of triple net Resources Spread method, it is characterised in that comprise the following steps：
Two two subnetworks are connected by common intermediate node, a triple net is formed, resource between three nodes not
Superposition is passed in stealpass, finally obtains the weights relation of initial point and finish node to predict degree of be mutually related between node；
Triple net refers to that two networks are integrated by common intermediate node, on new network, the directive biography of resource
Pass superposition；Being superimposed the process of transmission can regard the two two points processes propagated with last resource consolidation as.
2. triple net Resources Spread method according to claim 1, it is characterised in that：The triple net Resources Spread side
Method is achieved by the steps of：
It is in fact process that resource mutually shifts superposition in individual node in the TRANSFER MODEL of triple net；For handle well two
Component Nlm=(Vl,Vm,Elm) and Nme=(Vm,Ve,Eme) adjacency matrix is built respectivelyWith
Wherein in bipartite graph NlmIn, if lncRNAAnd miRNAIt is interrelated, thenOtherwiseIn bipartite graph Nme
In also have similar definition, if miRNAWith EFIt is interrelated, thenOtherwise
The process of the resource transfers of triple net resource transfer method, really will be projected in triple net on a bipartite graph
Weighting procedure on network, i.e., based on middle miRNA, unilateral projection adds on lncRNA-miRNA and miRNA-EF networks
Power；Therefore three parts are divided into, i.e., calculate unilateral projections of the miRNA on two two subnetworks, and joint two respectively
Individual projection constitutes final recommendation matrix.Process is as follows：
In lncRNA-miRNA related networks Nlm=(Vl,Vm,Elm) in, resource is first from node VmTurn to Vl, then resource pass again
Return to node Vm, this process can obtain miRNA in NlmOn resource projection matrixIt is defined as
WhereinWithRepresent in bipartite graph NlmIn, miRNAs nodesWithIn vectorial VmDegree, it is same
Represent in bipartite graph NlmIn, lncRNA nodesIn vectorial VlDegree, nlRepresent in NlmThe quantity of middle lncRNA, and parameter lambda1
∈[0,1]；
The N on miRNA-EF related networksme=(Vm,Ve,Eme), the transmission of resource is from node VeStart to be delivered to node Vm, then
V is transferred back to againeProcess；Final weights transfer matrix is defined as
WhereinWithRepresent in bipartite graph NmeIn, EF nodesWithIn vectorial VeDegree, it is sameRepresent
Bipartite graph NmeIn, miRNA nodesIn vectorial VmDegree, nmRepresent in NmeThe quantity of middle miRNA, and parameter lambda2∈[0,1]；
Wherein parameter lambda1And λ2Weight for adjusting Resources Spread matrix；When λ value is closer to 0, the resource for representing node is calculated
Be the average value of adjacent node, and closer to 1, represent resource be distributed in neck connects node it is more uniform；In prediction, parameter is got over and is connect
Nearly 0, predicted value is more conservative, and is worth close to 1, and last predicting the outcome more tend to and integrally predict the outcome；
The weight matrix W that above-mentioned two step is obtainedm, WeWith adjacency matrix AmeJoint, can obtain weight matrixAfter joint weight matrix will have been obtained, with bipartite graph
NlmAdjacency matrix AlmCombine again, final forecast power matrixCan be expressed asWherein ri,jRepresent lncRNAAnd EFAssociated weights；In the recommendation matrix for finally giving, often
One node all represents the correlation degree of lncRNA and EF, and weights are higher, and the two association possibility is bigger.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710133230.9A CN106934252A (en) | 2017-03-08 | 2017-03-08 | A kind of triple net Resources Spread method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710133230.9A CN106934252A (en) | 2017-03-08 | 2017-03-08 | A kind of triple net Resources Spread method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106934252A true CN106934252A (en) | 2017-07-07 |
Family
ID=59423508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710133230.9A Pending CN106934252A (en) | 2017-03-08 | 2017-03-08 | A kind of triple net Resources Spread method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106934252A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108427865A (en) * | 2018-03-14 | 2018-08-21 | 华南理工大学 | A method of prediction LncRNA and environmental factor incidence relation |
| CN108712663A (en) * | 2018-05-03 | 2018-10-26 | 武汉斗鱼网络科技有限公司 | Direct broadcasting room based on bipartite graph recommends method, corresponding medium and equipment |
| CN109711653A (en) * | 2017-10-26 | 2019-05-03 | 厦门一品威客网络科技股份有限公司 | Prestige visitor's task recommendation method based on prestige visitor's-task-label tripartite's figure |
| CN112270958A (en) * | 2020-10-23 | 2021-01-26 | 大连民族大学 | Prediction method based on hierarchical deep learning miRNA-lncRNA interaction relation |
| CN112289373A (en) * | 2020-10-27 | 2021-01-29 | 齐齐哈尔大学 | lncRNA-miRNA-disease association method fusing similarity |
-
2017
- 2017-03-08 CN CN201710133230.9A patent/CN106934252A/en active Pending
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109711653A (en) * | 2017-10-26 | 2019-05-03 | 厦门一品威客网络科技股份有限公司 | Prestige visitor's task recommendation method based on prestige visitor's-task-label tripartite's figure |
| CN109711653B (en) * | 2017-10-26 | 2020-12-15 | 厦门一品威客网络科技股份有限公司 | Weike task recommendation method based on Weike-task-label three-square diagram |
| CN108427865A (en) * | 2018-03-14 | 2018-08-21 | 华南理工大学 | A method of prediction LncRNA and environmental factor incidence relation |
| CN108427865B (en) * | 2018-03-14 | 2022-04-22 | 华南理工大学 | Method for predicting correlation between LncRNA and environmental factors |
| CN108712663A (en) * | 2018-05-03 | 2018-10-26 | 武汉斗鱼网络科技有限公司 | Direct broadcasting room based on bipartite graph recommends method, corresponding medium and equipment |
| CN108712663B (en) * | 2018-05-03 | 2021-02-02 | 武汉斗鱼网络科技有限公司 | Live broadcast room recommendation method based on bipartite graph, related storage medium and device |
| CN112270958A (en) * | 2020-10-23 | 2021-01-26 | 大连民族大学 | Prediction method based on hierarchical deep learning miRNA-lncRNA interaction relation |
| CN112270958B (en) * | 2020-10-23 | 2023-06-20 | 大连民族大学 | Prediction method based on layered deep learning miRNA-lncRNA interaction relationship |
| CN112289373A (en) * | 2020-10-27 | 2021-01-29 | 齐齐哈尔大学 | lncRNA-miRNA-disease association method fusing similarity |
| CN112289373B (en) * | 2020-10-27 | 2021-07-06 | 齐齐哈尔大学 | lncRNA-miRNA-disease association method fusing similarity |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106934252A (en) | A kind of triple net Resources Spread method | |
| CN111428147B (en) | Social recommendation method of heterogeneous graph volume network combining social and interest information | |
| Cui et al. | A novel artificial bee colony algorithm with an adaptive population size for numerical function optimization | |
| Costa et al. | Analyzing and modeling real-world phenomena with complex networks: a survey of applications | |
| Dede et al. | Weight minimization of trusses with genetic algorithm | |
| Hasançebi et al. | Performance evaluation of metaheuristic search techniques in the optimum design of real size pin jointed structures | |
| Kaveh et al. | Optimum design of skeletal structures using imperialist competitive algorithm | |
| Shi et al. | A link clustering based overlapping community detection algorithm | |
| CN102413029B (en) | Method for partitioning communities in complex dynamic network by virtue of multi-objective local search based on decomposition | |
| Bravo et al. | An integrated behavioral model of the land-use and transport systems with network congestion and location externalities | |
| Masuda et al. | Evolutionary dynamics and fixation probabilities in directed networks | |
| Wang et al. | Artificial immune optimization methods and applications-a survey | |
| CN110232971A (en) | A kind of doctor's recommended method and device | |
| Liu | Automatic calibration of a rainfall–runoff model using a fast and elitist multi-objective particle swarm algorithm | |
| Zhao et al. | Minimum dominating set of multiplex networks: Definition, application, and identification | |
| Maheri et al. | Size and topology optimization of trusses using hybrid genetic-particle swarm algorithms | |
| Le et al. | Drug response prediction by globally capturing drug and cell line information in a heterogeneous network | |
| Sun | A study of solving traveling salesman problem with genetic algorithm | |
| de Araujo Lima et al. | A hybrid approach based on genetic algorithm and nearest neighbor heuristic for solving the capacitated vehicle routing problem | |
| Peterson et al. | Simulated evolution of protein-protein interaction networks with realistic topology | |
| Wei et al. | A bi-objective evolutionary algorithm for multimodal multi-objective optimization | |
| CN104348695B (en) | A kind of mapping method of virtual network and its system based on artificial immune system | |
| Liu et al. | MEACCP: A membrane evolutionary algorithm for capacitated clustering problem | |
| Pluhacek et al. | PSO with partial population restart based on complex network analysis | |
| CN109410078A (en) | A kind of information propagation prediction method for the mobile social networking shared suitable for object oriented file |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170707 |