CN103324933A - Membrane protein sub-cell positioning method based on complex space multi-view feature fusion - Google Patents

Membrane protein sub-cell positioning method based on complex space multi-view feature fusion Download PDF

Info

Publication number
CN103324933A
CN103324933A CN2013102270918A CN201310227091A CN103324933A CN 103324933 A CN103324933 A CN 103324933A CN 2013102270918 A CN2013102270918 A CN 2013102270918A CN 201310227091 A CN201310227091 A CN 201310227091A CN 103324933 A CN103324933 A CN 103324933A
Authority
CN
China
Prior art keywords
feature
pssm
protein
amino acid
features
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
Application number
CN2013102270918A
Other languages
Chinese (zh)
Inventor
於东军
胡俊
杨健
吴小伟
沈红斌
戚湧
唐振民
杨静宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Science and Technology Changshu Research Institute Co Ltd
Original Assignee
Nanjing University of Science and Technology Changshu Research Institute Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Science and Technology Changshu Research Institute Co Ltd filed Critical Nanjing University of Science and Technology Changshu Research Institute Co Ltd
Priority to CN2013102270918A priority Critical patent/CN103324933A/en
Publication of CN103324933A publication Critical patent/CN103324933A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Investigating Or Analysing Biological Materials (AREA)

Abstract

The invention discloses a membrane protein sub-cell positioning method based on complex space multi-view feature fusion. Firstly, features of pseudo amino acid composition of a protein sequence and features of a position-specific scoring matrix based on autocorrelation transform are extracted; secondly, the two kinds of features are combined into a feature vector in a complex space in a parallel mode; thirdly, dimension reduction is conducted on the complex features after parallel combination through general principal component analysis method so as to remove noise; finally, the fused features are classified through an optimization evidence theory based K nearest neighbor classifier, and the position of a sub-cell is determined. The membrane protein sub-cell positioning method has the advantages that the complex space multi-view feature fusion technology is adopted, so that the diagnostic features of the protein sequence are extracted effectively; the K nearest neighbor classifier based on the optimization evidence theory is used, so that the accuracy of the membrane protein sub-cell positioning is improved.

Description

Memebrane protein subcellular location localization method based on complex space various visual angles Fusion Features
Technical field
The invention belongs to memebrane protein subcellular location positioning field, relate more specifically to a kind of memebrane protein subcellular location localization method based on complex space various visual angles Fusion Features, particularly a kind of memebrane protein subcellular location localization method that merges and adopt the broad sense principal component analysis to carry out feature extraction based on the various visual angles feature parallel.
Background technology
Memebrane protein (Transmembrane Protein) is the very important protein of a class in biosome, and it all plays very important effect for nutriment transportation, intercellular signal transmission and the energy exchange of cell.Simultaneously, memebrane protein also is a lot of pharmaceutically-active target spots, and most typical is the G protein family.There are some researches show, 60% ~ 70% target protein is G protein family member in the medicament research and development.In genomic data, there is 20% ~ 30% gene outcome to be predicted to be memebrane protein.
Existing research shows, there is close contacting the function of memebrane protein and its residing position in subcellular fraction.Subcellular location is exactly that memebrane protein appears at that organelle or compartment.The subcellular location information of obtaining protein is very helpful for the 26S Proteasome Structure and Function of understanding protein.The subcellular location of orientation film albumen has important practical significance exactly.Yet memebrane protein is the physiology course of a complexity from the initial position that is generated to final appearance.Although it is effective method that biology is measured, and takes time and effort.At rear era gene, memebrane protein sequence magnanimity presents, however corresponding subcellular location measure and make slow progress, need the automatic mode of the efficient memebrane protein subcellular location of research and development location badly.
In recent years, many Forecasting Methodologies for the Protein Subcellular position have appearred.For example, targetP (Emanuelsson, O., et al. Locating proteins in the cell using TargetP, SignalP and related tools. Nat Protoc, 2007,2 (4): 953-971) predict Eukaryotic Protein Subcellular position by the particular sequence that detects the N-end.Cell-Ploc (Kuo-Chen Chou, H.-B. Shen. Cell-PLoc 2.0:an improved package of web-servers for predicting subcellular localization of proteins in various organisms. Natural Science, 2010,2 (10): 1090-1103) Integrated using elementary sequence signature and senior feature come the Protein Subcellular position of variety classes biosome is predicted.CE-Ploc (A. Khan, A. Majid, and M. Hayat, CE-PLoc:An ensemble classifier for predicting protein subcellular locations by fusing different modes of pseudo amino acid composition, Comput. Biol. Chem., vol. 35, no. 4, pp. 218 – 229,2011.) then the diversity of fusion feature space and decision space is come the Protein Subcellular position is predicted.
Yet in the existing Protein Subcellular position predicting method, most methods are not distinguished the type of protein when predicting, only having only a few is to design for memebrane protein subcellular location location specially.The people such as Chou (Chou, K.C. and D.W. Elrod. Prediction of membrane protein types and subcellular locations. Proteins, 1999,34 (1): 137-153) ground-breaking work has been carried out in memebrane protein subcellular location location, pointed out close relation between the subcellular location of memebrane protein and their amino acid composition.After this, the people such as Sharpe (Sharpe, H.J., T.J. Stevens, and S. Munro. A comprehensive comparison of transmembrane domains reveals organelle-specific properties. Cell, 2010,142 (1): 158-169) successfully in forecast model, introduce the topology information of memebrane protein, and develop on this basis one based on the memebrane protein subcellular location localization method of neural network.Yet, these methods when using the various visual angles feature, employing be traditional serial combination strategy, the intrinsic dimensionality after causing making up is too high, thereby affects computing velocity and performance.
The present invention proposes a kind of new subcellular location localization method for memebrane protein, and the method at first extracts the feature of memebrane protein from different visual angles, then the various visual angles feature is carried out the parallel combined and obtains the complex space feature; On this basis, further use broad sense principal component analysis (PCA) (GPCA, J. Yang, J.Y.Y., D. Zhang. Feature fusion:parallel strategy vs. serial strategy. Pattern Recognition, 2003,36 (6): 1369-1381) carry out Fusion Features; Use at last based on the feature of the k nearest neighbor sorter OET-KNN that optimizes evidence theory after to dimensionality reduction and classify, thereby determine the subcellular location at memebrane protein place.
Summary of the invention
The object of the invention is to overcome the shortcoming that serial various visual angles Fusion Features calculated amount of the prior art is large, predetermined speed is slow, precision is not high.The memebrane protein subcellular location localization method based on complex space various visual angles Fusion Features that a kind of predetermined speed is fast, precision of prediction is high has been proposed.
Technical scheme of the present invention is: at first, memebrane protein to be positioned is carried out the various visual angles feature extraction: the one, extract feature from the amino acid composition visual angle; Another is to extract feature from the protein evolution visual angle.Secondly, with the various visual angles feature parallel combination of extracting, obtain the multiple feature of memebrane protein; Then, use the broad sense principal component analysis (PCA) that multiple feature is carried out the dimensionality reduction denoising; At last, use forecast model that the position of this memebrane protein is positioned.
The first step: various visual angles feature extraction
Pseudo amino acid composition composition characteristics (PseAAC): extract Pseudo amino acid composition composition characteristics (PseAAC) from memebrane protein sequence to be positioned, this feature is one
Figure 210834DEST_PATH_IMAGE001
The proper vector of dimension.
Position-specific scoring matrices feature (PSSM-ACT) based on the auto-correlation conversion: for one by
Figure 198381DEST_PATH_IMAGE002
The memebrane protein that individual amino acid forms can obtain its ad-hoc location score matrix (Position Specific Scoring Matrix, PSSM) by the PSI-BLAST program, and this matrix is
Figure 422689DEST_PATH_IMAGE002
Row 20 row.Then on this basis, extraction is based on position-specific scoring matrices (PSSM-ACT) feature of auto-correlation conversion.
Second step: Concurrent Feature merges
For the PseAAC that extracts in the first step and PSSM-ACT feature, by the mode of the parallel combined, obtain parallel multiple feature, this feature is arranged in complex space.Then, the proper vector after using the broad sense principal component analysis (PCA) to the parallel combined is carried out the dimensionality reduction denoising, the feature after obtaining merging.
The 3rd step: subcellular location location
Use is classified to the feature after merging based on the k nearest neighbor sorter (OET-KNN) of optimizing evidence theory, thereby determines the subcellular location at memebrane protein place.
Beneficial effect of the present invention:
The present invention compares its remarkable advantage with existing memebrane protein subcellular location location technology: (1) has used parallel various visual angles Fusion Features, rather than traditional serial merges from various visual angles, has significantly reduced the dimension of primitive character; (2) in complex space, use the broad sense principal component analysis to carry out Feature Dimension Reduction and denoising, effectively improved bearing accuracy.
Description of drawings
Fig. 1 is based on the memebrane protein subcellular location localization method system construction drawing of complex space various visual angles Fusion Features.
Embodiment
The present invention is further illustrated below in conjunction with accompanying drawing.
Fig. 1 has provided the system construction drawing of the inventive method:
1. use Chou (K. C. Chou, Prediction of protein cellular attributes using pseudo-amino acid composition, Proteins, vol. 43, no. 3, pp. 246-255,2001) method that proposes, extract Pseudo amino acid composition composition characteristics PseAAC from memebrane protein sequence to be positioned, this feature is one
Figure 562815DEST_PATH_IMAGE001
The proper vector of dimension.In this proper vector, front 20 components are classical amino acid compositions; The back
Figure 909483DEST_PATH_IMAGE003
Individual component is this memebrane protein Rank amino acid order correlative character,
Figure 458330DEST_PATH_IMAGE005
It is the kind of the amino acid attribute used.
2. from memebrane protein sequence to be positioned, extract the position-specific scoring matrices feature PSSM-ACT based on the auto-correlation conversion.The below is specifically addressed:
The first step uses PSI-BLAST software to obtain the original position specificity score matrix (PSSM) of memebrane protein to be positioned, is expressed as , illustrate suc as formula (1):
Figure 2013102270918100002DEST_PATH_IMAGE001
(1)
Second step, right
Figure 57305DEST_PATH_IMAGE006
Carry out normalized.With
Figure 685732DEST_PATH_IMAGE008
With
Figure 33406DEST_PATH_IMAGE009
Respectively expression
Figure 659559DEST_PATH_IMAGE006
The
Figure 108995DEST_PATH_IMAGE010
Mean value and the standard deviation of 20 scores in the row, that is:
Figure 26267DEST_PATH_IMAGE011
(2)
(3)
PSSM is after the normalization
Figure 41813DEST_PATH_IMAGE013
, wherein
Figure 962234DEST_PATH_IMAGE014
Obtain by following formula:
Figure 135726DEST_PATH_IMAGE015
(4)
Length is
Figure 208724DEST_PATH_IMAGE002
Memebrane protein standardization after PSSM can be expressed as:
(5)
The 3rd goes on foot, and calculates first the PSSM constituent features of 20 dimensions of this memebrane protein by formula (6) :
Figure 531886DEST_PATH_IMAGE018
(6)
Wherein,
Figure 724970DEST_PATH_IMAGE019
(7)
Then, exist
Figure 863828DEST_PATH_IMAGE020
Upper calculating
Figure 758840DEST_PATH_IMAGE021
The rank correlative character:
Figure 8556DEST_PATH_IMAGE022
(8)
Wherein,
，
Figure 365905DEST_PATH_IMAGE024
(9)
So, the position-specific scoring matrices feature (PSSM-ACT) based on the auto-correlation conversion is expressed as follows:
Figure 2013102270918100002DEST_PATH_IMAGE003
(10)
The 4th step, parallel various visual angles Fusion Features
Obtain to use first the mode of the parallel combined that two feature vectors are carried out Feature Combination after PseAAC and two kinds of features of PSSM-ACT, obtain a multiple feature.Specifically be described below:
If AWith BTwo feature spaces, such as PseAAC and PSSM-ACT feature space,
Figure 303085DEST_PATH_IMAGE026
With
Figure 205182DEST_PATH_IMAGE027
It is respectively feature space AWith BDimension.
Figure 685842DEST_PATH_IMAGE028
Be the training sample space.For given memebrane protein sample
Figure 289867DEST_PATH_IMAGE029
, its PseAAC and PSSM-ACT feature are respectively
Figure 146965DEST_PATH_IMAGE030
With
Figure 903568DEST_PATH_IMAGE031
And be about to these two features and make up:
Figure 555130DEST_PATH_IMAGE032
(11)
Wherein,
Figure 413495DEST_PATH_IMAGE033
It is an imaginary unit.Note, when With When dimension was unequal, before combination, that vector that dimension is low was used 0 polishing, until two vectorial dimensions equate.
For the proper vector after the parallel combined, re-use the broad sense principal component method and carry out dimensionality reduction and denoising, reach the purpose of Fusion Features.
The 5th step, the subcellular location location
Use is classified to the feature after merging based on the k nearest neighbor sorter OET-KNN that optimizes evidence theory, thereby determines the subcellular location at memebrane protein place.
Example:
Use the in the world MemLoci (Pierleoni of up-to-date issue, A., P.L. Martelli, and R. Casadio. MemLoci:predicting subcellular localization of membrane proteins in eukaryotes. Bioinformatics, 2011,27 (9): 1224-1230) memebrane protein subcellular location data set.
This data set comprises 10634 sequences altogether, wherein, has 4016 to come from plasmalemma, and 2308 from the organelle film, also has 4310 and comes from inner membrance.The composition of detailed benchmark dataset is as shown in table 1.
Table 1 memebrane protein Subcellular Localization data set forms
Table 2 has been listed this patent method, MemLoci method and the BLAST method positioning performance on above-mentioned data set relatively.In table 2, we are very easy to find, and this patent method will obviously be better than other two kinds of methods, not only have higher overall precision (
Figure 86474DEST_PATH_IMAGE037
) and generalized correlation correlativity (Generalized Correlation, GC), and the REC in every class (Recall), FPR (False Positive Rate) and MCC (Matthews correlation coefficients) preferably result is arranged.
Table 2 this method and MemLoci and BLAST method performance comparison
Figure DEST_PATH_IMAGE005
Above-described embodiment does not limit the present invention in any way, and every employing is equal to replaces or technical scheme that the mode of equivalent transformation obtains all drops in protection scope of the present invention.

Claims (4)

1. based on the memebrane protein subcellular location localization method of complex space various visual angles Fusion Features, it is characterized in that comprising following three steps:
The first step: various visual angles feature extraction: the amino acid based on protein forms information and amino acid sequence order information, extracts Pseudo amino acid composition composition characteristics PseAAC; Use PSI-BLAST obtains the evolution information PSSM of protein sequence to be positioned, then from the position-specific scoring matrices feature PSSM-ACT of PSSM extraction based on the auto-correlation conversion;
Second step: first two kinds of features extracting in the first step are carried out the Concurrent Feature combination, the feature that obtains is positioned at complex space; Then, use the broad sense principal component analysis that the feature after making up is carried out dimensionality reduction, the feature after obtaining merging;
The 3rd step: set up the location prediction model, use as the location prediction model, is predicted the subcellular location of memebrane protein based on the k nearest neighbor sorter OET-KNN that optimizes evidence theory.
2. the memebrane protein subcellular location localization method based on complex space various visual angles Fusion Features according to claim 1 is characterized in that: in the described first step for one by
Figure 2013102270918100001DEST_PATH_IMAGE001
The protein that individual amino acid forms can obtain its ad-hoc location score matrix PSSM by the PSI-BLAST program, and this matrix is Row 20 row; Then PSSM is extracted the position-specific scoring matrices feature PSSM-ACT of auto-correlation conversion; Secondly, the amino acid based on protein forms information and amino acid sequence order information, extraction Pseudo amino acid composition composition characteristics PseAAC.
3. albuminous membranae protein subcellular location prediction method according to claim 1 and 2, it is characterized in that: the proper vector of the principal component analytical method GPCA that uses broad sense after to Parallel Fusion carried out dimension-reduction treatment in complex space, when reducing calculated amount, eliminate the noise in the data.
4. albuminous membranae protein subcellular location prediction method according to claim 2 is characterized in that: the various visual angles feature extraction extracts Pseudo amino acid composition composition characteristics PseAAC from memebrane protein sequence to be positioned, this feature is one
Figure DEST_PATH_IMAGE003
The proper vector of dimension, front 20 components are classical amino acid compositions, the back
Figure 165870DEST_PATH_IMAGE004
Individual component is this memebrane protein
Figure 902882DEST_PATH_IMAGE006
Rank amino acid order correlative character,
Figure DEST_PATH_IMAGE007
It is the kind of the amino acid attribute used; Extract the position-specific scoring matrices feature PSSM-ACT based on the auto-correlation conversion from memebrane protein sequence to be positioned, concrete steps are as follows:
The first step uses PSI-BLAST software to obtain the original position specificity score matrix (PSSM) of memebrane protein to be positioned, is expressed as
Figure 488584DEST_PATH_IMAGE008
, illustrate suc as formula (1):
Figure 811840DEST_PATH_IMAGE001
(1)
Second step, right
Figure 731478DEST_PATH_IMAGE008
Carry out normalized, use
Figure DEST_PATH_IMAGE011
With
Figure DEST_PATH_IMAGE013
Respectively expression
Figure 81426DEST_PATH_IMAGE008
The
Figure DEST_PATH_IMAGE015
Mean value and the standard deviation of 20 scores in the row, that is:
Figure 497495DEST_PATH_IMAGE016
(2)
(3)
PSSM is after the normalization
Figure 203283DEST_PATH_IMAGE018
, wherein
Figure DEST_PATH_IMAGE019
Obtain by following formula:
Figure 850033DEST_PATH_IMAGE020
(4)
Length is Memebrane protein standardization after PSSM can be expressed as:
Figure 2013102270918100001DEST_PATH_IMAGE002
(5)
The 3rd goes on foot, and calculates first the PSSM constituent features of 20 dimensions of this memebrane protein by formula (6)
Figure DEST_PATH_IMAGE023
:
Figure 595453DEST_PATH_IMAGE024
(6)
Wherein,
Figure DEST_PATH_IMAGE025
(7)
Then, exist
Figure 405014DEST_PATH_IMAGE026
Upper calculating
Figure 911082DEST_PATH_IMAGE028
The rank correlative character:
Figure DEST_PATH_IMAGE029
(8)
Wherein,
Figure 2666DEST_PATH_IMAGE030
，
Figure DEST_PATH_IMAGE031
(9)
So, the position-specific scoring matrices feature (PSSM-ACT) based on the auto-correlation conversion is expressed as follows:
Figure 853614DEST_PATH_IMAGE003
(10)
According to claim 1 or 4 described albuminous membranae protein subcellular location prediction methods, it is characterized in that: described parallel various visual angles Fusion Features, obtain after PseAAC and two kinds of features of PSSM-ACT, use first the mode of the parallel combined that two feature vectors are carried out Feature Combination, obtain a multiple feature:
Two feature spaces of PseAAC and PSSM-ACT,
Figure 299972DEST_PATH_IMAGE034
With
Figure DEST_PATH_IMAGE035
It is respectively feature space AWith BDimension,
Figure 554105DEST_PATH_IMAGE036
Be the training sample space, for given memebrane protein sample , its PseAAC and PSSM-ACT feature difference,
Figure 54356DEST_PATH_IMAGE038
With,
Figure DEST_PATH_IMAGE039
, and be about to these two features and make up:
(11)
Wherein,
Figure DEST_PATH_IMAGE041
An imaginary unit, when
Figure DEST_PATH_IMAGE043
With
Figure 134494DEST_PATH_IMAGE044
When dimension was unequal, before combination, that vector that dimension is low was used 0 polishing, until two vectorial dimensions equate.
CN2013102270918A 2013-06-08 2013-06-08 Membrane protein sub-cell positioning method based on complex space multi-view feature fusion Pending CN103324933A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2013102270918A CN103324933A (en) 2013-06-08 2013-06-08 Membrane protein sub-cell positioning method based on complex space multi-view feature fusion

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2013102270918A CN103324933A (en) 2013-06-08 2013-06-08 Membrane protein sub-cell positioning method based on complex space multi-view feature fusion

Publications (1)

Publication Number Publication Date
CN103324933A true CN103324933A (en) 2013-09-25

Family

ID=49193661

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2013102270918A Pending CN103324933A (en) 2013-06-08 2013-06-08 Membrane protein sub-cell positioning method based on complex space multi-view feature fusion

Country Status (1)

Country Link
CN (1) CN103324933A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104615911A (en) * 2015-01-12 2015-05-13 上海交通大学 Method for predicting membrane protein beta-barrel transmembrane area based on sparse coding and chain training
CN104636635A (en) * 2015-01-29 2015-05-20 南京理工大学 Protein crystallization predicting method based on two-layer SVM learning mechanism
CN105046103A (en) * 2015-07-03 2015-11-11 景德镇陶瓷学院 Novel representation method for protein sequence fusing genetic information
CN105046106A (en) * 2015-07-14 2015-11-11 南京农业大学 Protein subcellular localization and prediction method realized by using nearest-neighbor retrieval
CN106022366A (en) * 2016-07-04 2016-10-12 杭州电子科技大学 Rotary mechanical equipment fault diagnosis method based on neighbor evidence fusion
CN108197427A (en) * 2018-01-02 2018-06-22 山东师范大学 Proteins subcellular location method and apparatus based on depth convolutional neural networks
CN108595909A (en) * 2018-03-29 2018-09-28 山东师范大学 TA targeting proteins prediction techniques based on integrated classifier
CN112242179A (en) * 2020-09-09 2021-01-19 天津大学 Method for identifying type of membrane protein

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102760210A (en) * 2012-06-19 2012-10-31 南京理工大学常熟研究院有限公司 Adenosine triphosphate binding site predicting method for protein
CN102779240A (en) * 2012-06-21 2012-11-14 哈尔滨工程大学 Inherent irregular protein structure forecasting method based on kernel canonical correlation analysis

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102760210A (en) * 2012-06-19 2012-10-31 南京理工大学常熟研究院有限公司 Adenosine triphosphate binding site predicting method for protein
CN102779240A (en) * 2012-06-21 2012-11-14 哈尔滨工程大学 Inherent irregular protein structure forecasting method based on kernel canonical correlation analysis

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
DONGJUN YU等: "《Enhancing membrane protein subcellular localization prediction by parallel fusion of multi-view features》", 《IEEE TRANACTIONS ON NANOBIOSCIENCE》 *
DONGJUN YU等: "《Enhancing membrane protein subcellular localization prediction by parallel fusion of multi-view features》", 《IEEE TRANACTIONS ON NANOBIOSCIENCE》, vol. 11, no. 4, 31 December 2012 (2012-12-31), pages 375 - 385 *
DONG-JUN YU等: "《Learning protein multi-view features in complex space》", 《AMINO ACIDS》 *

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104615911A (en) * 2015-01-12 2015-05-13 上海交通大学 Method for predicting membrane protein beta-barrel transmembrane area based on sparse coding and chain training
CN104615911B (en) * 2015-01-12 2017-07-18 上海交通大学 Method based on sparse coding and chain study prediction memebrane protein beta barrel trans-membrane regions
CN104636635A (en) * 2015-01-29 2015-05-20 南京理工大学 Protein crystallization predicting method based on two-layer SVM learning mechanism
CN104636635B (en) * 2015-01-29 2018-06-12 南京理工大学 Crystallization of protein Forecasting Methodology based on two layers of SVM study mechanism
CN105046103A (en) * 2015-07-03 2015-11-11 景德镇陶瓷学院 Novel representation method for protein sequence fusing genetic information
CN105046106A (en) * 2015-07-14 2015-11-11 南京农业大学 Protein subcellular localization and prediction method realized by using nearest-neighbor retrieval
CN105046106B (en) * 2015-07-14 2018-02-23 南京农业大学 A kind of Prediction of Protein Subcellular Location method realized with nearest _neighbor retrieval
CN106022366A (en) * 2016-07-04 2016-10-12 杭州电子科技大学 Rotary mechanical equipment fault diagnosis method based on neighbor evidence fusion
CN108197427A (en) * 2018-01-02 2018-06-22 山东师范大学 Proteins subcellular location method and apparatus based on depth convolutional neural networks
CN108595909A (en) * 2018-03-29 2018-09-28 山东师范大学 TA targeting proteins prediction techniques based on integrated classifier
CN112242179A (en) * 2020-09-09 2021-01-19 天津大学 Method for identifying type of membrane protein

Similar Documents

Publication Publication Date Title
CN103324933A (en) Membrane protein sub-cell positioning method based on complex space multi-view feature fusion
Delsuc et al. Phylogenomics and the reconstruction of the tree of life
CN107506617B (en) Half-local social information miRNA-disease association prediction method
CN102184349A (en) System and method for clustering gene expression data based on manifold learning
Xu et al. Prediction of posttranslational modification sites from amino acid sequences with kernel methods
CN103942568A (en) Sorting method based on non-supervision feature selection
CN102842043B (en) Particle swarm classifying method based on automatic clustering
CN109448787B (en) Protein subnuclear localization method for feature extraction and fusion based on improved PSSM
Du et al. SubMito‐PSPCP: Predicting Protein Submitochondrial Locations by Hybridizing Positional Specific Physicochemical Properties with Pseudoamino Acid Compositions
CN103617203B (en) Protein-ligand bindings bit point prediction method based on query driven
CN104462365A (en) Multi-view target searching method based on probability model
CN105389326A (en) Image annotation method based on weak matching probability canonical correlation model
CN112489723B (en) DNA binding protein prediction method based on local evolution information
WO2018045529A1 (en) Network station site planning method and device
CN104392459A (en) Infrared image segmentation method based on improved FCM (fuzzy C-means) and mean drift
CN102663681B (en) Gray scale image segmentation method based on sequencing K-mean algorithm
CN102750532B (en) Method for detecting targets based on components
CN103559363A (en) Method for calculating optimum response strategy in imperfect information extensive game
CN110660448B (en) Key protein identification method based on topological and functional characteristics of protein
CN103177121A (en) Locality preserving projection method for adding pearson relevant coefficient
Jiao et al. Identification of plant vacuole proteins by exploiting deep representation learning features
CN109492770A (en) A kind of net with attributes embedding grammar based on the sequence of personalized relationship
CN107526937B (en) MiRNA-disease association prediction method based on collaborative filtering
CN106250818A (en) A kind of total order keeps the face age estimation method of projection
Song et al. A New Approach for Parameter Estimation in the Sequence-structure Alignment of Non-coding RNAs.

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20130925