##DMLDA-LocLIFT

Identification of multi-label protein subcellular localization using DMLDA dimensionality reduction and LIFT classifier

###Guiding principles:

**The dataset file contains Gram-negative bacteria dataset, Gram-positive bacteria dataset and plant dataset.

**Feature extraction

1. Evolutionary information: psepssm.m is the implementation of PsePSSM.
2. Physicochemical_information: PAAC.m,mainpseaac.m is the implementation of PseAAC. ebgw1.m,ebgw2.m,ebgw3.m,lizhuebgw.m is the implementation of EBGW.
3. Sequence_information: Dipeptide composition can be found from http://www.csbio.sjtu.edu.cn/bioinf/PseAAC/#.
4. Annotation information： Gene Ontology can be found from http://www.ebi.ac.uk/GOA/. ** Dimensional reduction: DMLDA_transform.m represents the DMLDA. MDDM_transform.m represents MDDM. PCA_transform.m represents PCA. MLSI_transform represents MLSI. MVMD_transform represents MVMD.

** Classifier: LIFT.m is the implementation of LIFT. MLKNN_test.m,MLKNN_train.m are the implementation of MLKNN. InsDif.m is the implementation of InsDif. MLLOC.m is the implementation of MLLOC.

** independent_test: The independent_test file contains the code of the test of independent dataset.

** Demo: Run the demo.m in MATLAB,and you can choose different dimensionality reduction methods and classifiers.