CN109727637A - Method based on shuffled frog leaping algorithm identification key protein matter - Google Patents
Method based on shuffled frog leaping algorithm identification key protein matter Download PDFInfo
- Publication number
- CN109727637A CN109727637A CN201811643461.5A CN201811643461A CN109727637A CN 109727637 A CN109727637 A CN 109727637A CN 201811643461 A CN201811643461 A CN 201811643461A CN 109727637 A CN109727637 A CN 109727637A
- Authority
- CN
- China
- Prior art keywords
- protein
- frog
- node
- formula
- worst
- 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
-
- 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
Landscapes
- Peptides Or Proteins (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of methods based on shuffled frog leaping algorithm identification key protein matter, by converting non-directed graph for protein-protein interaction network, obtain the corresponding subcellular localization information of protein, protein complex participation information and functional annotation information, node in protein-protein interaction network and side are handled, frog population is initialized according to the local average connectivity of protein node, group is divided according to the adaptive value of frog, frog carries out first evolution in group, execute local search, all frogs carry out global thoughts communication, execute global search, generate key protein matter.The method of the present invention can accurately identify key protein matter;The simulation experiment result show sensitivity, specificity, F estimate, the indexs such as positive predictive value, negative predictive value and accuracy it is more excellent;Compared with other key protein matter recognition methods, the biological nature of the topological characteristic and protein of the optimization characteristics of shuffled frog leaping algorithm and protein-protein interaction network itself is combined to identify key protein matter, improves the recognition accuracy of key protein matter.
Description
Technical field
The invention belongs to technical field of biological information, and in particular to one kind identifies key protein matter based on shuffled frog leaping algorithm
Method.
Background technique
Protein is the important component for forming all cells of organism, tissue, is the main undertaker of vital movement.It is different
Protein different vital movements is participated in biological cell, therefore, protein is divided into two major classes, key protein matter and
Non-key protein.Key protein matter is also lethal protein, the missing of key protein will lead to cell can not normally breed or
Person is dead, so that organism loses certain functions, or even can not survive.Identification to key protein is in life science
One important research content, it is correct to identify that key protein matter not only facilitates the Operational Mechanisms for understanding organism, and for disease
Disease diagnosis and drug design also have highly important application value.
Biologically, the identification of key protein matter is mainly struck using the method for some Bioexperiment, such as single-gene
It removes, RNA interference, conditional gene knockout etc..However these methods take time and effort, very expensive, and applicable species range
It is limited.With the development of high-throughput techniques, a large amount of biological data can obtain and the fast development of computer technology, so that
Identify that key protein matter becomes the new developing direction in the field using the method for calculation biology.Currently, utilizing calculation method
Identification key protein matter can be divided mainly into two classes: the method based on network topology and the method based on biological information fusion.
A large number of studies show that whether crucial the opening up in protein-protein interaction network with the node of a protein node
It is closely related to flutter characteristic.Based on this, it has been proposed that a series of centrality using node are estimated to identify key protein matter
Method.It such as spends centrality (Degree Centrality, DC), betweenness center (Betweenness Centrality, BC),
Degree of approach centrality (Closeness Centrality, CC), eigenvector centrality (Eigenvector Centrality,
EC), information centre's property (Information Centrality, IC), subgraph centrality (Subgraph Centrality, SC)
Deng.With the mining analysis that gos deep into network topology characteristic, the more key protein matter identification side of the topological property based on node
Method is suggested.Wang et al. proposes a kind of new centrality Measurement Method NC, and this method is by calculating side convergence factor simultaneously
The relationship between the characteristic and node and its neighbours of node is considered, thus to predict the key of protein;Li et al. people
A kind of local average connectivity method (LAC) is proposed, the neighbor node of each node is generated a new subgraph by this method,
Key protein matter is identified according to degree of each node in subgraph;Qi et al. proposes local interaction density method
(LID), this method identifies key protein matter based on the interaction relationship between the neighbor node of each node.These are based on
The centrality Measurement Method of network topology is largely dependent upon the reliability of protein-protein interaction network, and passes through height
The protein-protein interaction network data packet that the method for throughput biological experiment obtains contains a large amount of false positive, this is greatly influenced
The accuracy rate of key protein matter identification.
In order to overcome the method based on network topology to identify defect present in key protein matter, some researchers combine egg
The biological meaning of white matter proposes the identification that some methods based on biological information fusion are used for key protein matter.Such as crucial egg
White matter recognition methods PeC and WDC combine the network topology characteristic and gene expression data information of protein node;UDoNC is closed
The recognition methods of key albumen integrates protein-protein interaction network and protein structure domain information;TEO is mutual in protein
The functional annotation information and gene expression information of protein have been merged in effect network;SON combines subcellular localization information, straight
It is the topological property of homologous information and protein-protein interaction network.In addition, studies have shown that protein complex and crucial egg
There is close relationship between white matter, Hart et al. is experiments prove that key protein matter is usually enriched in and certain to have spy
Determine in the compound of function.Therefore, some key protein matter recognition methods based on protein complex are suggested, such as UC,
LIDC and LBCC etc..Experimental result shows that these incorporate the method for biological data in protein-protein interaction network
It is got well than being based only upon the recognition effect of method of network topology structure before, the identification for effectively improving key protein matter is accurate
Rate.
It is big although predicting that key protein matter makes some progress by the calculation method based on horizontal network at present
The recognition accuracy of part recognition methods is still lower, and robustness is poor, and this is mainly due to the incomplete of biological data
Property and the otherness between unreliability and the complexity and each species of vital movement, and most methods do not account for
It is special using a small number of topological characteristics or biology in isolation to the difference of the tightness degree and bonding strength that are contacted between network node
Property, analysis global and on the whole is lacked to key protein matter node.
Summary of the invention
In order to overcome the disadvantages of the above prior art, the purpose of the present invention is to provide one kind is known based on shuffled frog leaping algorithm
The method of other key protein matter identifies crucial egg using the optimization characteristics of shuffled frog leaping algorithm from protein-protein interaction network
White matter improves the recognition accuracy of key protein matter.
In order to achieve the above object, the present invention is achieved by the following scheme:
A kind of method based on shuffled frog leaping algorithm identification key protein matter disclosed by the invention, comprising the following steps:
1) non-directed graph is converted by protein-protein interaction network
Protein-protein interaction network is converted to a non-directed graph G=(V, E), wherein V={ vi, i=1,2 ..., n }
For node viSet, E be side e set, node viIndicate that protein, side e indicate the interaction between protein;
2) in protein-protein interaction network side and node handle
Calculate the local average connectivity LAC of protein node, the subcellular localization score value SC and albumen of protein node
Matter compound score value PC calculates the structural similarity SS and functional similarity FS on the side of connection two proteins node;
3) initial frog population is randomly generated
Enabling F is frog population scale, C be the number of candidate key protein for needing to identify, i.e. a frog is individual
All proteins node is taken preceding 2 × C according to LAC value descending sort for the search range for reducing key protein matter by length
Biggish node generates initial population in LAC value, and TopV is these protein node sets;
4) frog group is divided group by global search process
Descending sort is carried out to frog population by the adaptive value Essentiality (f) of frog individual, wherein f=1,2 ...
F records the highest frog Px of adaptive value, F frog is assigned to m group Y1, Y2..., YmIn, meet Yk=[X (j) | X
(j)=X (k+m × (j-1)), j=1,2 ..., n, k=1,2 ..., m], wherein X (j) indicates the jth after sorting in frog group only
Frog;
5) first evolution, i.e. progress local search are carried out in each group: k, iter respectively indicate group's counter drawn game
Evolution counter in portion's is respectively intended to be compared with group sum m and local maxima evolution number maxiter, k=1, iter=
1, maxiter ∈ [50,100];
6) frog of all groups is mixed, all frog individuals is ranked up again by new adaptive value and race
Group divides, and records new global optimum frog individual Px (new), if the difference of the adaptive value of Px (new) and Px is not less than 10-4,
Turn to step 5;Otherwise, step 7 is turned to;
7) key protein matter is generated
Protein in optimal frog individual is exported as key protein matter.
Preferably, in step 2), the local average connectivity LAC of protein node is obtained by formula (1):
In formula,Indicate node viNeighbor node collection,Be byIn node constitute subgraph,Indicate setIn any node vjIn subgraphIn neighbor node number.
Preferably, the subcellular localization score value SC of protein node is obtained by formula (2):
In formula, ClIndicate a kind of subcellular components, l=1,2 ... 11, SI (Cl) indicate subcellular components ClImportance obtain
Point, it is obtained by formula (3):
In formula, num (l) represents ClIncluded in key protein matter number, Tnum represent is key protein sum
Mesh;
The protein complex score value of protein node is calculated by formula (4):
In formula, F (vi) indicate node viThe frequency in known protein complex is appeared in, is obtained by formula (5), FM is institute
There is the maximum frequency appeared in known protein complex in protein node;
In formula, N represents known protein complex total number, if protein node appears in protein complex PtIn,
Then Pt(vi)=1, otherwise Pt(vi)=0;
The initial weight of each protein node is obtained by formula (6):
InW(vi)=SC (vi)×PC(vi) formula (6).
Preferably, in step 2), the structural similarity SS for connecting the side of two proteins node is calculated by formula (7):
In formula, Γ (i), Γ (j) respectively indicate node vi, vjNeighbor node collection add vi, vj;
The functional similarity on the side of connection two proteins node is calculated by formula (8):
In formula, g (i), g (j) respectively indicate annotation node viAnd vjGO term set;
The weight for connecting the side of two proteins node is obtained by formula (9):
Weij=SSij×FSijFormula (9)
The final weight of each protein node is obtained by formula (10):
Preferably, the adaptive value Essentiality (f) of frog individual is obtained by formula (11) in step 4):
Preferably, step 5) concrete operations are as follows:
Local thoughts communication, i.e. progress local updating, k=k+1 5-1) are carried out to the frog in k-th of frog group;
5-2) in frog group YkIn, it selects s frog and enters subfamily group sub_Yk, (s < n), the choosing of frog in subfamily group
It takes based on wheel disc bet method, i.e., a possibility that adaptive value of frog individual is bigger in group, which is selected is bigger, enables Pb
Optimal and worst frog, iter=iter+1 in subfamily group are respectively indicated with Pw;
The position that worst frog Pw 5-3) is updated according to local optimum frog Pb in subfamily group, for worst frog individual
Whether appearing in local optimum frog individual Pb per one-dimensional component protein for it judged, if occurring, then makes the component by Pw
Protein remains unchanged;Otherwise the one-component albumen chosen in Pb is replaced with certain probability, i.e., worst frog Pw
Position according to formula Pnl1=update1 (Pw, Pb, r1) be updated, r in formula1For with the component protein in Pb in Pw
The probability that is replaced of protein, Pnl1New position after being updated for worst frog Pw according to local optimum frog Pb;
If 5-4) improving the position of worst frog, i.e., the adaptive value of worst frog in a new location by step 5-2)
It is higher than the adaptive value on original position, just with newly generated position Pnl1Replace original position Pw, otherwise just uses global optimum
Frog Px updates the position of worst frog individual again, judges whether worst frog individual Pw appears in per one-dimensional component protein
In global optimum frog individual Px, if occurring, then remain unchanged the component protein;Otherwise choosing one in Px
A component albumen is replaced with certain probability, i.e., the position of worst frog Pw is according to formula Pnl2=update2 (Pw, Px, r2)
It is updated, r in formula2For the probability being replaced with the component protein in Px to the protein in Pw, Pnl2For worst frog
Pw is according to the new position after global optimum frog Px update;
If 5-5) improving the position of worst frog, i.e., the adaptive value of worst frog in a new location by step 5-3)
It is higher than the adaptive value on original position, just with newly generated position Pnl2Replace original position Pw, is otherwise randomly generated in wet
The frog of any position in ground substitutes the worst frog, i.e., the position of worst frog Pw is according to formula Pnl3=update3 (Pw,
TopV, r3) be updated, r in formula3For the probability being replaced per one-dimensional component protein in Pw, Pnl3For worst frog Pw with
New position after machine update;
As long as execute above step 5-3), step 5-4) and 5-5 in any primary update, this subgroup will be recalculated
Optimal frog individual Pb and worst frog individual Pw;
If 5-6) iter≤maxiter turns to step 5-2);
If 5-7) k≤m turns to step 5-1), otherwise turn to step 6.
It is further preferred that step 5-3) in, the new position Pnl that is obtained after the location updating of worst frog Pw1Calculating
Method uses algorithm update1 (Pw, Pb, r1), the specific method is as follows:
Step1: finding out and occur in Pb, the protein set Pset1 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Pb;
Step3: ifAnd random number rand > r1, then a protein v is randomly selected from set Pset1j
Replace vi, and Pset1=Pset1- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
It is further preferred that in step 5-4), the new position Pnl that is obtained after the location updating of worst frog Pw2Meter
Calculation method uses algorithm update2 (Pw, Px, r2), the specific method is as follows:
Step1: finding out and occur in Px, the protein set Pset2 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Px;
Step3: ifAnd random number rand > r2, then a protein v is randomly selected from set Pset2j
Replace vi, and Pset2=Pset2- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
It is further preferred that in step 5-5), the new position Pnl that is obtained after the location updating of worst frog Pw3Meter
Calculation method uses algorithm update3 (Pw, TopV, r3), the specific method is as follows:
Step1: finding out and occur in TopV, the protein set Pset3 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in TopV;
Step3: ifAnd random number rand > r3, then a protein is randomly selected from set Pset3
vjReplace vi, and Pset3=Pset3- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
Compared with prior art, the invention has the following advantages:
1, the present invention is utilized subcellular localization information and protein is multiple when assigning initial weight to protein node
Object information is closed, participation situation by the subcellular localization characteristic of protein and in the composite measures the important of protein
Property improves the recognition accuracy of key protein matter in a certain degree.
2, the present invention not only allows for the characteristic of protein itself when assigning final weight to protein node, and
The connectivity between its neighbours' characteristic and protein is also contemplated, the bonding strength between protein is by calculating two eggs
What topology connection structure similitude and functional similarity between white matter obtained, while network topology and biological information are considered,
Manifold fusion use allows the invention to more accurately and effectively identify key protein matter.
3, jump is carried out by information interchange between present invention simulation frog individual to find the more place of food
Process identifies key protein matter, and one group of candidate key protein is regarded as a frog, is searched by executing part in group
Rope strategy, and after each group evolves to certain phase, entire frog group carries out global information exchange, last algorithm
When termination, a histone matter corresponding to the highest frog of adaptive value is exactly the key protein matter identified, with other keys
Albumen recognition methods is compared, by the topological characteristic and egg of the optimization characteristics of shuffled frog leaping algorithm and protein-protein interaction network
The biological nature of white matter itself is combined the identification process for realizing key protein matter, and the identification for improving key protein matter is accurate
Rate.
4, key protein matter can effectively be identified from protein-protein interaction network using the present invention, not only facilitated
Understand the adjusting and controlling growth process of cell and the Operational Mechanisms of vital movement, helps people to understand organism and sustain life activity institute
The primary demand needed, and important theoretical base can be provided for related researcher from genome and protein assembly level
Plinth, research and development preparation of diagnoses and treatment and drug for disease etc. have extremely important application value.
Detailed description of the invention
Fig. 1 is the flow chart of the method for the invention that key protein matter is identified based on shuffled frog leaping algorithm;
Fig. 2 is the frog group Y of step 5) of the inventionkCarry out the method flow diagram of local search.
Specific embodiment
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work
It encloses.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so as to the embodiment of the present invention described herein can in addition to illustrating herein or
Sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that cover
Cover it is non-exclusive include, for example, the process, method, system, product or equipment for containing a series of steps or units are not necessarily limited to
Step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, product
Or other step or units that equipment is intrinsic.
The invention will be described in further detail with reference to the accompanying drawing:
As shown in Figure 1, the present invention is based on the methods of shuffled frog leaping algorithm identification key protein matter, comprising the following steps:
1) non-directed graph is converted by protein-protein interaction network
Protein-protein interaction network is converted to a non-directed graph G=(V, E), wherein V={ vi, i=1,2 ..., n }
For node viSet, E be side e set, node viIndicate that protein, side e indicate the interaction between protein;
2) in protein-protein interaction network side and node handle
The local average connectivity of protein node is calculated by formula (1):
In formula,Indicate node viNeighbor node collection,Be byIn node constitute subgraph,
Indicate setIn any node vjIn subgraphIn neighbor node number;
The subcellular localization score value of protein node is calculated by formula (2):
In formula, ClIndicate a kind of subcellular components, l=1,2 ... 11, SI (Cl) indicate subcellular components ClImportance obtain
Point, it is obtained by formula (3):
In formula, num (l) represents ClIncluded in key protein matter number, Tnum represent is key protein sum
Mesh;
The protein complex score value of protein node is calculated by formula (4):
In formula, F (vi) indicate node viThe frequency in protein complex is appeared in, is obtained by formula (5), FM is institute
There is protein node to appear in the maximum frequency in known protein complex;
In formula, N represents known protein complex total number, if protein node appears in protein complex PtIn,
Then Pt(vi)=1, otherwise Pt(vi)=0;
The initial weight of each protein node is obtained by formula (6):
InW(vi)=SC (vi)×PC(vi) formula (6)
The structural similarity on the side of connection two proteins node is calculated by formula (7):
In formula, Γ (i), Γ (j) respectively indicate node vi, vjNeighbor node collection add vi, vj;
The functional similarity on the side of connection two proteins node is calculated by formula (8):
In formula, g (i), g (j) respectively indicate annotation node viAnd vjGO term set;
The weight for connecting the side of two proteins node is obtained by formula (9):
Weij=SSij×FSijFormula (9)
The final weight of each protein node is obtained by formula (10):
3) initial frog population is randomly generated
Enabling F is frog population scale, C be the number of candidate key protein for needing to identify, i.e. a frog is individual
All proteins node is taken preceding 2 × C according to LAC value descending sort for the search range for reducing key protein matter by length
Biggish node generates initial population in LAC value, and TopV is these protein node sets;
4) frog group is divided group by global search process
Descending sort is carried out to frog population by the adaptive value Essentiality (f) of frog individual, wherein f=1,2 ...
F records the highest frog Px of adaptive value.F frog is assigned to m group Y1, Y2..., YmIn, meet Yk=[X (j) | X
(j)=X (k+m × (j-1)), j=1,2 ..., n, k=1,2 ..., m], wherein X (j) indicates the jth after sorting in frog group only
Frog;
5) first evolution, i.e. progress local search are carried out in each group: k, iter respectively indicate group's counter drawn game
Evolution counter in portion's is respectively intended to be compared with group sum m and local maxima evolution number maxiter, k=1, iter=
1, maxiter ∈ [50,100];Referring to fig. 2, specifically comprise the following steps:
Local thoughts communication, i.e. progress local updating, k=k+1 5-1) are carried out to the frog in k-th of frog group;
5-2) in frog group YkIn, it selects s frog and enters subfamily group sub_Yk, (s < n), the choosing of frog in subfamily group
A possibility that taking is to be based on wheel disc bet method, and the adaptive value of frog individual is bigger in group, and the frog is selected is bigger, enables Pb
Optimal and worst frog, iter=iter+1 in subfamily group are respectively indicated with Pw;
The position that worst frog Pw 5-3) is updated according to local optimum frog Pb in subfamily group, for worst frog individual
Whether appearing in local optimum frog individual Pb per one-dimensional component protein for it judged, if occurring, then makes the component by Pw
Protein remains unchanged;Otherwise choosing the one-component albumen (protein does not occur in Pw) in Pb with certain general
Rate is replaced, i.e., the position of worst frog Pw is according to formula Pnl1=update1 (Pw, Pb, r1) be updated, r in formula1For
The probability that the protein in Pw is replaced with the component protein in Pb, Pnl1It is worst frog Pw according to local optimum blueness
New position after frog Pb update;
If 5-4) improving the position of worst frog by step 5-2, i.e., the adaptive value of worst frog in a new location
It is higher than the adaptive value on original position, just with newly generated position Pnl1Replace original position Pw, otherwise just uses global optimum
Frog Px updates the position of worst frog individual again, judges whether worst frog individual Pw appears in per one-dimensional component protein
In global optimum frog individual Px, if occurring, then remain unchanged the component protein;Otherwise choosing one in Px
A component albumen is replaced with certain probability, i.e., the position of worst frog Pw is according to formula Pnl2=update2 (Pw, Px, r2)
It is updated, r in formula2For the probability being replaced with the component protein in Px to the protein in Pw, Pnl2For worst frog
Pw is according to the new position after global optimum frog Px update;
If 5-5) improving the position of worst frog by step 5-3, i.e., the adaptive value of worst frog in a new location
It is higher than the adaptive value on original position, just with newly generated position Pnl2Replace original position Pw, is otherwise randomly generated in wet
The frog of any position in ground substitutes the worst frog, i.e., the position of worst frog Pw is according to formula Pnl3=update3 (Pw,
TopV, r3) be updated, r in formula3For the probability being replaced per one-dimensional component protein in Pw, Pnl3For worst frog Pw with
New position after machine update;
No matter performing any primary update in above 5-3,5-4 and 5-5, require to recalculate the optimal of this subgroup
Frog individual Pb and worst frog individual Pw;
If 5-6) iter≤maxiter, step 5-2 is turned to;
If 5-7) k≤m, step 5-1 is turned to, otherwise turns to step 6;
6) frog of all groups is mixed, all frog individuals is ranked up again by new adaptive value and race
Group divides, and records new global optimum frog individual Px (new), if the difference of the adaptive value of Px (new) and Px is not less than 10-4,
Turn to step 5;Otherwise, step 7 is turned to;
7) key protein matter is generated
Protein in optimal frog individual is exported as key protein matter.
The adaptive value Essentiality (f) of frog individual is obtained by formula (11) in step 4) of the invention:
Step 5-3 of the invention) in, the new position Pnl that is obtained after the location updating of worst frog Pw1Calculation method adopt
With algorithm 1update1 (Pw, Pb, r1), the specific method is as follows:
Step1: finding out and occur in Pb, the protein set Pset1 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Pb;
Step3: ifAnd random number rand > r1, then a protein v is randomly selected from set Pset1j
Replace vi, and Pset1=Pset1- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
Step 5-4 of the invention) in, the new position Pnl that is obtained after the location updating of worst frog Pw2Calculation method adopt
With algorithm 2update2 (Pw, Px, r2), the specific method is as follows:
Step1: finding out and occur in Px, the protein set Pset2 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Px;
Step3: ifAnd random number rand > r2, then a protein v is randomly selected from set Pset2j
Replace vi, and Pset2=Pset2- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
Step 5-5 of the invention) in, the new position Pnl that is obtained after the location updating of worst frog Pw3Calculation method adopt
With algorithm 3update3 (Pw, TopV, r3), the specific method is as follows:
Step1: finding out and occur in TopV, the protein set Pset3 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in TopV;
Step3: ifAnd random number rand > r3, then a protein is randomly selected from set Pset3
vjReplace vi, and Pset3=Pset3- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
Below by way of specific embodiment, the present invention is described in more detail:
Here is a kind of method based on shuffled frog leaping algorithm identification key protein matter by taking protein network as an example, specifically
It operates as follows:
The present embodiment is to pick up from the saccharomyces cerevisiae data set (DIP 2010.10.10 editions) of DIP database as emulation data
Collection removes self-interaction and duplicate interaction, in total includes 5093 protein, 24743 sides.Subcellular localization number
It downloads to obtain from COMPARTMENTS (20140830 editions) database according to being, including 6002 yeast proteins and 238657
Subcellular location record.Known protein complex data are by integrating CM270, CM425, CYC408 and CYC428 tetra-
What the data that a data are concentrated obtained, include altogether 745 protein complexes, covers 2167 protein.GO data are
The compact version of GO ontologies.Key protein prime number is according to by integrating in tetra- databases of MIPS, SGD, DEG and SGDP
Data obtain, contain 1285 key protein matter altogether.Experiment porch is 10 operating system of Windows, Intel Intel Core i5-
6600 double-core 3.31GHz processors, 8GB physical memory, with Matlab R2014a software realization method of the invention.
Specific step is as follows:
1, non-directed graph is converted by protein-protein interaction network
Protein-protein interaction network comprising 5093 protein and 24743 interaction relationships is converted to one
Non-directed graph G=(V, E), wherein V={ vi, i=1,2 ..., 5093 } it is node viSet, E be 24743 side e set,
Node viIndicate that protein, side e indicate the interaction between protein.
2, in protein-protein interaction network side and node handle
To node viPretreatment: i=1,2 ..., 5093 often give a determining i, can calculate node viPart it is flat
Equal connectivity is calculated the local average connectivity of protein node by formula (1):
In formula,Indicate node viNeighbor node collection,Be byIn node constitute subgraph,
Indicate setIn any node vjIn subgraphIn neighbor node number;Protein node is calculated by formula (2)
Subcellular localization score value:
In formula, ClIndicate a kind of subcellular components, l=1,2 ... 11, SI (Cl) indicate subcellular components ClImportance obtain
Point, it is obtained by formula (3):
In formula, num (l) represents ClIncluded in key protein matter number, Tnum represent be the key that yeast egg
White sum, Tnum=1285;The protein complex score value of protein node is calculated by formula (4):
In formula, F (vi) indicate node viThe frequency in protein complex is appeared in, is obtained by formula (5), FM is institute
There is protein node to appear in the maximum frequency in known protein complex;
In formula, N represents known protein complex total number, N=745, if to appear in protein compound for protein node
Object PtIn, then Pt(vi)=1, otherwise Pt(vi)=0;The initial weight of each protein node is obtained by formula (6):
InW(vi)=SC (vi)×PC(vi) formula (6)
The structural similarity on the side of connection two proteins node is calculated by formula (7):
In formula, Γ (i), Γ (j) respectively indicate node vi, vjNeighbor node collection add vi, vj;It calculates and connects by formula (8)
Connect the functional similarity on the side of two proteins node:
In formula, g (i), g (j) respectively indicate annotation node viAnd vjGO term set;
The weight for connecting the side of two proteins node is obtained by formula (9):
Weij=SSij×FSijFormula (9)
The final weight of each protein node is obtained by formula (10):
3, initial frog population is randomly generated
Enabling F is frog population scale, and F=100, C are the number for needing the candidate key protein identified, i.e. a frog
All proteins node is taken preceding 2 according to LAC value descending sort for the search range for reducing key protein matter by the length of individual
× C biggish the node of LAC value generates initial population, and TopV is these protein node sets;
4, frog group is divided group by global search process
Descending sort is carried out to frog population by the adaptive value Essentiality (f) of frog individual, wherein f=1,2 ...
F records the highest frog Px of adaptive value.F frog is assigned to m group Y1, Y2..., YmIn, meet Yk=[X (j) | X
(j)=X (k+m × (j-1)), j=1,2 ..., n, k=1,2 ..., m], wherein m=10, n=10, X (j) indicate the frog after sequence
Jth frog in group, adaptive value Essentiality (f) are obtained by formula (11):
5, first evolution, i.e. progress local search are carried out in each group: k, iter respectively indicate group's counter drawn game
Evolution counter in portion's is respectively intended to be compared with group sum m and local maxima evolution number maxiter, k=1, iter=
1, maxiter ∈ [50,100];
5-1, local thoughts communication, i.e. progress local updating, k=k+1 are carried out to the frog in k-th of frog group;
5-2, in frog group YkIn, it selects s frog and enters subfamily group sub_Yk, (s < n), the choosing of frog in subfamily group
A possibility that taking is to be based on wheel disc bet method, and the adaptive value of frog individual is bigger in group, and the frog is selected is bigger, enables Pb
Optimal and worst frog, iter=iter+1 in subfamily group are respectively indicated with Pw;
5-3, the position that worst frog Pw is updated according to local optimum frog Pb in subfamily group, for worst frog individual
Whether appearing in local optimum frog individual Pb per one-dimensional component protein for it judged, if occurring, then makes the component by Pw
Protein remains unchanged;Otherwise choosing the one-component albumen (protein does not occur in Pw) in Pb with certain general
Rate is replaced, i.e., the position of worst frog Pw is according to formula Pnl1=update1 (Pw, Pb, r1) be updated, r in formula1For
The probability that the protein in Pw is replaced with the component protein in Pb, Pnl1It is worst frog Pw according to local optimum blueness
New position after frog Pb update, Pnl1It can be obtained by algorithm 1:
Algorithm 1update1 (Pw, Pb, r1)
Step1: finding out and occur in Pb, the protein set Pset1 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Pb;
Step3: ifAnd random number rand > r1, then a protein v is randomly selected from set Pset1j
Replace vi, and Pset1=Pset1- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
If 5-4, the position for improving worst frog by step 5-2, i.e., the adaptive value of worst frog in a new location
It is higher than the adaptive value on original position, just with newly generated position Pnl1Replace original position Pw, otherwise just uses global optimum
Frog Px updates the position of worst frog individual again, judges whether worst frog individual Pw appears in per one-dimensional component protein
In global optimum frog individual Px, if occurring, then remain unchanged the component protein;Otherwise choosing one in Px
A component albumen is replaced with certain probability, i.e., the position of worst frog Pw is according to formula Pnl2=update2 (Pw, Px, r2)
It is updated, r in formula2For the probability being replaced with the component protein in Px to the protein in Pw, Pnl2For worst frog
Pw is according to the new position after global optimum frog Px update, Pnl2It can be obtained by algorithm 2:
Algorithm 2update2 (Pw, Px, r2)
Step1: finding out and occur in Px, the protein set Pset2 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Px;
Step3: ifAnd random number rand > r2, then a protein v is randomly selected from set Pset2j
Replace vi, and Pset2=Pset2- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
If 5-5) improving the position of worst frog by step 5-3, i.e., the adaptive value of worst frog in a new location
It is higher than the adaptive value on original position, just with newly generated position Pnl2Replace original position Pw, is otherwise randomly generated in wet
The frog of any position in ground substitutes the worst frog, i.e., the position of worst frog Pw is according to formula Pnl3=update3 (Pw,
TopV, r3) be updated, r in formula3For the probability being replaced per one-dimensional component protein in Pw, Pnl3For worst frog Pw with
New position after machine update, Pnl3It can be obtained by algorithm 3:
Algorithm 3update3 (Pw, TopV, r3)
Step1: finding out and occur in TopV, the protein set Pset3 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in TopV;
Step3: ifAnd random number rand > r3, then a protein is randomly selected from set Pset3
vjReplace vi, and Pset3=Pset3- { vj};
Step4: Step2-3 is repeated, until protein all in Pw all judges to finish.
No matter performing any primary update in above 5-3,5-4 and 5-5, require to recalculate the optimal of this subgroup
Frog individual Pb and worst frog individual Pw;
If 5-6, iter≤maxiter turn to step 5-2;
If 5-7, k≤m turn to step 5-1, step 6 is otherwise turned to;
6, the frog of all groups is mixed, all frog individuals is ranked up again by new adaptive value and race
Group divides, and records new global optimum frog individual Px (new), if the difference of the adaptive value of Px (new) and Px is not less than 10-4,
Turn to step 5;Otherwise, step 7 is turned to;
7, key protein matter is generated
Protein in optimal frog individual is exported as key protein matter.
In order to verify effectiveness of the invention, inventor identifies crucial egg using 1 shuffled frog leaping algorithm of the embodiment of the present invention
The method of white matter carries out the identification of key protein matter to the protein network in DIP database, to the candidate identified
Key protein prime number mesh (C) is respectively 1% of all proteins number of network nodes in protein-protein interaction network, 5%, 10%,
15%, 20%, 25% when, is analyzed, and the results are shown in Table 1, table 2, and table 1 shows the side with other current identification key protein matter
The result that method identifies carries out the comparison of recognition accuracy, and table 2 is shown with the method for other identification key protein matter each
Comparison in a evaluation index.
1 present invention of table is compared with the key protein matter of other methods identification is in accuracy rate
The comparison in each evaluation index of table 2 present invention and other methods
Table 1, which is shown, identifies 1%, 5%, 10%, 15%, 20% using the method for the present invention, 25% protein conduct
The recognition accuracy that candidate key protein is compared with the key protein matter in java standard library, and closed with other 9 kinds identifications
The comparison of key method of protein recognition result.As can be seen from Table 1, it is compared with other methods, the method for the present invention can be more effectively
Identify key protein matter, for the number of candidate key protein from 1% to 25%, the method for the present invention has highest identification accurate
Rate.
Table 2 shows when the candidate key protein number identified is 25%, the method for the present invention and other 9 kinds of methods
Estimate in sensitivity, specificity, F, knot is compared in the assessment on the evaluation indexes such as positive predictive value, negative predictive value and accuracy
Fruit.As can be seen from Table 2, it is compared with other methods, the present invention can predict more key protein matter, and prediction accuracy
It is higher.
In conclusion the present invention is based on the methods of shuffled frog leaping algorithm identification key protein matter, by the way that protein is mutual
Effect network be converted into non-directed graph, obtain the corresponding subcellular localization information of protein, protein complex participation information and
Functional annotation information is handled node in protein-protein interaction network and side, according to the local average of protein node
Connectivity initialization frog population carries out first evolution according to the adaptive value of frog division group, frog in group, executes part
Search, all frogs carry out global thoughts communication, execute global search, generate key protein matter.The method of the present invention can be accurately
Identify key protein matter;The simulation experiment result show sensitivity, specificity, F estimate, positive predictive value, negative predictive value and
The indexs such as accuracy are more excellent;Compared with other key protein matter recognition methods, by the optimization characteristics and albumen of shuffled frog leaping algorithm
The biological nature of the topological characteristic and protein of matter interactive network itself is combined to identify key protein matter, improves
The recognition accuracy of key protein matter.
The above content is merely illustrative of the invention's technical idea, and this does not limit the scope of protection of the present invention, all to press
According to technical idea proposed by the present invention, any changes made on the basis of the technical scheme each falls within claims of the present invention
Protection scope within.
Claims (9)
1. a kind of method based on shuffled frog leaping algorithm identification key protein matter, which comprises the following steps:
1) non-directed graph is converted by protein-protein interaction network
Protein-protein interaction network is converted to a non-directed graph G=(V, E), wherein V={ vi, i=1,2 ..., n } it is knot
Point viSet, E be side e set, node viIndicate that protein, side e indicate the interaction between protein;
2) in protein-protein interaction network side and node handle
It is multiple to calculate the local average connectivity LAC of protein node, the subcellular localization score value SC of protein node and protein
Object score value PC is closed, the structural similarity SS and functional similarity FS on the side of connection two proteins node are calculated;
3) initial frog population is randomly generated
Enabling F is frog population scale, and C be the number of candidate key protein that needs identify, i.e., the length of one frog individual,
All proteins node is taken into preceding 2 × C LAC value according to LAC value descending sort for the search range for reducing key protein matter
In biggish node generate initial population, TopV is these protein node sets;
4) frog group is divided group by global search process
Descending sort is carried out to frog population by the adaptive value Essentiality (f) of frog individual, wherein f=1,2 ... F, note
The highest frog Px of adaptive value is recorded, F frog is assigned to m group Y1, Y2..., YmIn, meet Yk=[X (j) | X (j)
=X (k+m × (j-1)), j=1,2 ..., n, k=1,2 ..., m], wherein X (j) indicates that the jth after sorting in frog group is only green
The frog;
5) first evolution is carried out in each group, i.e. progress local search: k, iter respectively indicate group's counter and part into
Change counter, be respectively intended to be compared with group sum m and local maxima evolution number maxiter, k=1, iter=1,
Maxiter ∈ [50,100];
6) frog of all groups is mixed, all frog individuals is ranked up again by new adaptive value and group draws
Point, and new global optimum frog individual Px (new) is recorded, if the difference of the adaptive value of Px (new) and Px is not less than 10-4, turn to
Step 5);Otherwise, step 7) is turned to;
7) key protein matter is generated
Protein in optimal frog individual is exported as key protein matter.
2. the method according to claim 1 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that step
2) in, the local average connectivity LAC of protein node is obtained by formula (1):
In formula,Indicate node viNeighbor node collection,Be byIn node constitute subgraph,Table
Show setIn any node vjIn subgraphIn neighbor node number.
3. the method according to claim 1 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that albumen
The subcellular localization score value SC of matter node is obtained by formula (2):
In formula, ClIndicate a kind of subcellular components, l=1,2 ... 11, SI (Cl) indicate subcellular components ClImportance score, by
Formula (3) obtains:
In formula, num (l) represents ClIncluded in key protein matter number, Tnum represent be key protein total number;
The protein complex score value of protein node is calculated by formula (4):
In formula, F (vi) indicate node viThe frequency in known protein complex is appeared in, is obtained by formula (5), FM is all eggs
The maximum frequency in known protein complex is appeared in white matter node;
In formula, N represents known protein complex total number, if protein node appears in protein complex PtIn, then Pt
(vi)=1, otherwise Pt(vi)=0;
The initial weight of each protein node is obtained by formula (6):
InW(vi)=SC (vi)×PC(vi) formula (6).
4. the method according to claim 1 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that step
2) in, the structural similarity SS for connecting the side of two proteins node is calculated by formula (7):
In formula, Γ (i), Γ (j) respectively indicate node vi, vjNeighbor node collection add vi, vj;
The functional similarity on the side of connection two proteins node is calculated by formula (8):
In formula, g (i), g (j) respectively indicate annotation node viAnd vjGO term set;
The weight for connecting the side of two proteins node is obtained by formula (9):
Weij=SSij×FSijFormula (9)
The final weight of each protein node is obtained by formula (10):
5. the method according to claim 1 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that step
4) the adaptive value Essentiality (f) of frog individual is obtained by formula (11) in:
6. the method according to claim 1 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that step
5) concrete operations are as follows:
Local thoughts communication, i.e. progress local updating, k=k+1 5-1) are carried out to the frog in k-th of frog group;
5-2) in frog group YkIn, it selects s frog and enters subfamily group sub_Yk, (s < n), the selection base of frog in subfamily group
A possibility that in wheel disc bet method, i.e., the adaptive value of frog individual is bigger in group, and the frog is selected is bigger, enables Pb and Pw
Respectively indicate optimal and worst frog, iter=iter+1 in subfamily group;
The position that worst frog Pw 5-3) is updated according to local optimum frog Pb in subfamily group, for worst frog individual Pw,
Whether appearing in local optimum frog individual Pb per one-dimensional component protein for it judged, if occurring, then makes the component egg
White matter remains unchanged;Otherwise the one-component albumen chosen in Pb is replaced with certain probability, i.e., worst frog Pw's
Position is according to formula Pnl1=update1 (Pw, Pb, r1) be updated, r in formula1For with the component protein in Pb in Pw
The probability that protein is replaced, Pnl1New position after being updated for worst frog Pw according to local optimum frog Pb;
If 5-4) improving the position of worst frog by step 5-2), i.e., the adaptive value of worst frog in a new location is than former
Adaptive value on position is high, just with newly generated position Pnl1Replace original position Pw, otherwise just uses global optimum frog
Px updates the position of worst frog individual again, judges whether worst frog individual Pw appears in the overall situation per one-dimensional component protein
In optimal frog individual Px, if occurring, then remain unchanged the component protein;Otherwise choosing one point in Px
Amount albumen is replaced with certain probability, i.e., the position of worst frog Pw is according to formula Pnl2=update2 (Pw, Px, r2) carry out
It updates, r in formula2For the probability being replaced with the component protein in Px to the protein in Pw, Pnl2For worst frog Pw root
New position after being updated according to global optimum frog Px;
If 5-5) improving the position of worst frog by step 5-3), i.e., the adaptive value of worst frog in a new location is than former
Adaptive value on position is high, just with newly generated position Pnl2Replace original position Pw, is otherwise randomly generated in wetland
The frog of any position substitute the worst frog, i.e., the position of worst frog Pw is according to formula Pnl3=update3 (Pw, TopV,
r3) be updated, r in formula3For the probability being replaced per one-dimensional component protein in Pw, Pnl3At random more for worst frog Pw
New position after new;
As long as execute above step 5-3), step 5-4) and 5-5 in any primary update, this subgroup will be recalculated most
Excellent frog individual Pb and worst frog individual Pw;
If 5-6) iter≤maxiter turns to step 5-2);
If 5-7) k≤m turns to step 5-1), otherwise turn to step 6.
7. the method according to claim 6 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that step
In 5-3), the new position Pnl that is obtained after the location updating of worst frog Pw1Calculation method using algorithm update1 (Pw, Pb,
r1), the specific method is as follows:
Step1: finding out and occur in Pb, the protein set Pset1 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Pb;
Step3: ifAnd random number rand > r1, then a protein v is randomly selected from set Pset1jReplacement
vi, and Pset1=Pset1- { vj};
Step4: Step 2-3 is repeated, until protein all in Pw all judges to finish.
8. the method according to claim 6 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that in step
Rapid 5-4) in, the new position Pnl that is obtained after the location updating of worst frog Pw2Calculation method using algorithm update2 (Pw,
Px, r2), the specific method is as follows:
Step1: finding out and occur in Px, the protein set Pset2 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in Px;
Step3: ifAnd random number rand > r2, then a protein v is randomly selected from set Pset2jReplacement
vi, and Pset2=Pset2- { vj};
Step4: Step 2-3 is repeated, until protein all in Pw all judges to finish.
9. the method according to claim 6 based on shuffled frog leaping algorithm identification key protein matter, which is characterized in that in step
Rapid 5-5) in, the new position Pnl that is obtained after the location updating of worst frog Pw3Calculation method using algorithm update3 (Pw,
TopV, r3), the specific method is as follows:
Step1: finding out and occur in TopV, the protein set Pset3 not occurred in Pw;
Step2: for component protein vi∈ Pw judges whether occur in TopV;
Step3: ifAnd random number rand > r3, then a protein v is randomly selected from set Pset3jIt replaces
Change vi, and Pset3=Pset3- { vj};
Step4: Step 2-3 is repeated, until protein all in Pw all judges to finish.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811643461.5A CN109727637B (en) | 2018-12-29 | 2018-12-29 | Method for identifying key proteins based on mixed frog-leaping algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811643461.5A CN109727637B (en) | 2018-12-29 | 2018-12-29 | Method for identifying key proteins based on mixed frog-leaping algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109727637A true CN109727637A (en) | 2019-05-07 |
CN109727637B CN109727637B (en) | 2023-09-05 |
Family
ID=66299447
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811643461.5A Active CN109727637B (en) | 2018-12-29 | 2018-12-29 | Method for identifying key proteins based on mixed frog-leaping algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109727637B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110400599A (en) * | 2019-07-22 | 2019-11-01 | 陕西师范大学 | Method based on dove colony optimization algorithm identification key protein matter |
CN111584010A (en) * | 2020-04-01 | 2020-08-25 | 昆明理工大学 | Key protein identification method based on capsule neural network and ensemble learning |
CN111951951A (en) * | 2020-07-14 | 2020-11-17 | 西安电子科技大学 | Disease module detection method and system based on connectivity significance |
CN113255889A (en) * | 2021-05-26 | 2021-08-13 | 安徽理工大学 | Occupational pneumoconiosis multi-modal analysis method based on deep learning |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030092752A1 (en) * | 2000-06-21 | 2003-05-15 | Chen-Yong Lin | Structure based discovery of inhibitors of matriptase for the treatment of cancer and other conditions |
CN107330235A (en) * | 2017-05-08 | 2017-11-07 | 华南理工大学 | A kind of molecule power science study method of polypeptide chain self assembling process |
CN107633105A (en) * | 2017-07-12 | 2018-01-26 | 西北工业大学 | A kind of four rotor wing unmanned aerial vehicle parameter identification methods based on improvement shuffled frog leaping algorithm |
-
2018
- 2018-12-29 CN CN201811643461.5A patent/CN109727637B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030092752A1 (en) * | 2000-06-21 | 2003-05-15 | Chen-Yong Lin | Structure based discovery of inhibitors of matriptase for the treatment of cancer and other conditions |
CN107330235A (en) * | 2017-05-08 | 2017-11-07 | 华南理工大学 | A kind of molecule power science study method of polypeptide chain self assembling process |
CN107633105A (en) * | 2017-07-12 | 2018-01-26 | 西北工业大学 | A kind of four rotor wing unmanned aerial vehicle parameter identification methods based on improvement shuffled frog leaping algorithm |
Non-Patent Citations (3)
Title |
---|
MIN LI ET AL.: "United Complex Centrality for Identification of Essential Proteins from PPI Networks", 《IEEE/ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS 》 * |
刘洲洲等: "基于混合蛙跳算法的非均匀分簇WSNs路由协议", 《计算机应用研究》 * |
吴超: "蛋白质功能预测的半监督学习算法研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110400599A (en) * | 2019-07-22 | 2019-11-01 | 陕西师范大学 | Method based on dove colony optimization algorithm identification key protein matter |
CN111584010A (en) * | 2020-04-01 | 2020-08-25 | 昆明理工大学 | Key protein identification method based on capsule neural network and ensemble learning |
CN111584010B (en) * | 2020-04-01 | 2022-05-27 | 昆明理工大学 | Key protein identification method based on capsule neural network and ensemble learning |
CN111951951A (en) * | 2020-07-14 | 2020-11-17 | 西安电子科技大学 | Disease module detection method and system based on connectivity significance |
CN111951951B (en) * | 2020-07-14 | 2023-06-23 | 西安电子科技大学 | Disease module detection method and system based on connected significance |
CN113255889A (en) * | 2021-05-26 | 2021-08-13 | 安徽理工大学 | Occupational pneumoconiosis multi-modal analysis method based on deep learning |
Also Published As
Publication number | Publication date |
---|---|
CN109727637B (en) | 2023-09-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109727637A (en) | Method based on shuffled frog leaping algorithm identification key protein matter | |
CN108319812B (en) | Method for identifying key protein based on cuckoo search algorithm | |
Ciriello et al. | AlignNemo: a local network alignment method to integrate homology and topology | |
CN105279397B (en) | A kind of method of key protein matter in identification of protein interactive network | |
Marinakis et al. | A hybrid discrete artificial bee colony-GRASP algorithm for clustering | |
CN109637579B (en) | Tensor random walk-based key protein identification method | |
CN106874708B (en) | Using the method for the artificial bee colony optimization algorithm identification key protein matter for the mechanism of looking for food | |
CN109801674B (en) | Key protein identification method based on heterogeneous biological network fusion | |
CN107784196B (en) | Method for identifying key protein based on artificial fish school optimization algorithm | |
CN105117618B (en) | It is a kind of based on the drug targets of guilt by association principle and network topology structure feature interact recognition methods | |
CN107885971B (en) | Method for identifying key protein by adopting improved flower pollination algorithm | |
Lei et al. | Detecting protein complexes from DPINs by density based clustering with Pigeon-Inspired Optimization Algorithm | |
CN109686403A (en) | Based on key protein matter recognition methods in uncertain protein-protein interaction network | |
CN109509509A (en) | Protein complex method for digging based on dynamic weighting protein-protein interaction network | |
CN107092812A (en) | A kind of method based on genetic algorithm in identification key protein matter in PPI networks | |
CN111145830A (en) | Protein function prediction method based on network propagation | |
CN113688383A (en) | Attack defense testing method based on artificial intelligence and artificial intelligence analysis system | |
Yin et al. | Coco: A coupled contrastive framework for unsupervised domain adaptive graph classification | |
CN108804871B (en) | Key protein identification method based on maximum neighbor subnet | |
CN108229643A (en) | A kind of method using drosophila optimization algorithm identification key protein matter | |
Ji et al. | Ant colony optimization with multi-agent evolution for detecting functional modules in protein-protein interaction networks | |
CN108388769B (en) | Protein functional module identification method based on edge-driven label propagation algorithm | |
Babu et al. | A simplex method-based bacterial colony optimization algorithm for data clustering analysis | |
CN109616153A (en) | A method of key protein matter is identified using improved HITS algorithm | |
Panagiotopoulos et al. | MEvA-X: a hybrid multiobjective evolutionary tool using an XGBoost classifier for biomarkers discovery on biomedical datasets |
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 |