Superzchen / iFeature

iFeature is a comprehensive Python-based toolkit for generating various numerical feature representation schemes from protein or peptide sequences. iFeature is capable of calculating and extracting a wide spectrum of 18 major sequence encoding schemes that encompass 53 different types of feature descriptors. Furthermore, iFeature also integrates five kinds of frequently used feature clustering algorithms, four feature selection algorithms and three dimensionality reduction algorithms.

update news!!!

• iFeatureOmega - an updated version of iFeature is now available! (2022-05-23)

iFeatureOmega is a comprehensive platform for generating, analyzing and visualizing more than 180 representations for biological sequences, 3D structures and ligands. To the best of our knowledge, iFeatureOmega supplies the largest number of feature extraction and analysis approaches for most molecule types compared to other pipelines. Three versions (i.e. iFeatureOmega-Web, iFeatureOmega-GUI and iFeatureOmega-CLI) of iFeatureOmega have been made available to cater to both experienced bioinformaticians and biologists with limited programming expertise. iFeatureOmega also expands its functionality by integrating 15 feature analysis algorithms (including ten cluster algorithms, three dimensionality reduction algorithms and two feature normalization algorithms) and providing nine types of interactive plots for statistical features visualization (including histogram, kernel density plot, heatmap, boxplot, line chart, scatter plot, circular plot, protein three dimensional structure plot and ligand structure plot). iFeatureOmega is an open-source platform for academic purposes. The web server can be accessed through https://ifeatureomega.erc.monash.edu and the GUI and CLI versions can be download at: https://github.com/Superzchen/iFeatureOmega-GUI and https://github.com/Superzchen/iFeatureOmega-CLI, respectively.

iFeatureOmega flowchart:

• iLearnPlus - an updated version of iFeature and iLearn is now available! (2021-02-28)

iLearnPlus is the first machine-learning platform with both graphical- and web-based user interface that enables the construction of automated machine-learning pipelines for computational analysis and predictions using nucleic acid and protein sequences. iLearnPlus integrates 21 machine-learning algorithms (including 12 conventional classification algorithms, two ensemble-learning frameworks and seven deep-learning approaches) and 19 major sequence encoding schemes (in total 147 feature descriptors), outnumbering all the current web servers and stand-alone tools for biological sequence analysis, to the best of our knowledge. In addition, the friendly GUI (Graphical User Interface) of iLearnPlus is available to biologists to conduct their analyses smoothly, significantly increasing the effectiveness and user experience compared to the existing pipelines. iLearnPlus is an open-source platform for academic purposes and is available at https://github.com/Superzchen/iLearnPlus/. The iLearnPlus-Basic module is online accessible at http://ilearnplus.erc.monash.edu/.

iLearnPlus-Basic module interface:

• iLearn - the updated version of iFeature is now available! (2019-03-13)

iLearn is a Python Toolkit and Web Server Integrating the Functionality of Feature Calculation, Extraction, Clustering, Feature Selection, Feature Normalization, Dimension Reduction and Model Construction for Classification, Best Model Selection, Ensemble Learning and Result Visualization for DNA, RNA and Protein Sequences. Please refer to https://github.com/Superzchen/iLearn for details.

iFeature: A python package and web server for features extraction and selection from protein and peptide sequences

iFeature is a comprehensive Python-based toolkit for generating various numerical feature representation schemes from protein or peptide sequences. iFeature is capable of calculating and extracting a wide spectrum of 18 major sequence encoding schemes that encompass 53 different types of feature descriptors. Furthermore, iFeature also integrates five kinds of frequently used feature clustering algorithms, four feature selection algorithms and three dimensionality reduction algorithms.

Installation

• Download iFeature by

git clone https://github.com/Superzchen/iFeature

iFeature is an open-source Python-based toolkit, which operates depending on the Python environment (Python Version 3.0 or above) and can be run on multi-OS systems (such as Windows, Mac and Linux operating systems). Before running iFeature, user should make sure all the following packages are installed in their Python environment: sys, os, shutil, scipy, argparse, collections, platform, math, re, numpy (1.13.1), sklearn (0.19.1), matplotlib (2.1.0), and pandas (0.20.1). For convenience, we strongly recommended users to install the Anaconda Python 3.0 version (or above) in your local computer. The software can be freely downloaded from https://www.anaconda.com/download/.

For users who want to generate descriptors by our provided iFeature package :

cd to the iFeature folder which contains iFeature.py, iFeaturePseKRAAC.py, cluster.py and feaSelector.py. All the functions regarding feature extraction, feature or sample clustering and feature selection analysis can be executed through these four main programs by specifying the parameter '--type'.

"iFeature.py" is the main program used to extract 37 different types of feature descriptors. For details of other parameters, run:

python iFeature.py --help 

"iFeaturePseKRAAC.py" is the program used to extract the 16 types of pseudo K-tuple reduced amino acid composition (PseKRAAC) feature descriptors. For details of other parameters, run:

python iFeaturePseKRAAC.py --help

"cluster.py" is the program used for running the feature or sample clustering algorithms. For details of other parameters, run:

python cluster.py --help

"feaSelector.py" is the fourth main program used to implement the feature selection algorithms. For details of other parameters, run:

python feaSelector.py --help

Furthermore, the iFeature package contains other Python scripts to generate the position-specific scoring matrix (PSSM) profiles, predicted protein secondary structure and predicted protein disorder, which have also been often used to improve the prediction performance of machine learning-based classifiers in conjunction with sequence-derived information. The three dimensionality reduction algorithms are also included in the scripts directory.

Examples for users to extract descriptors from iFeature.py. All files in the example commands can be found in the examples directory.

The input protein or peptide sequences for iFeature.py and iFeaturePseKRAAC.py should be in fasta format, Please find the example in example folder. The following parameters are required by iFeature.py:

• --help show help of 'iFeature.py'
• --file protein/peptide sequence file in fasta format
• --type feature types for protein sequence analysis
• --path data file path used for 'PSSM', 'SSEB(C)', 'Disorder(BC)', 'ASA' and 'TA' encodings
• --train training file in fasta format only used for 'KNNprotein' or 'KNNpeptide' encodings
• --label sample label file only used for 'KNNprotein' or 'KNNpeptide' encodings
• --order output order for of Amino Acid Composition (i.e. AAC, EAAC, CKSAAP, DPC, DDE, TPC) descriptors
• --userDefinedOrder user defined output order for of Amino Acid Composition (i.e. AAC, EAAC, CKSAAP, DPC, DDE, TPC) descriptors
• --out the generated descriptor file Running the following command to obtain the Composition of k-spaced Amino Acid Pairs (CKSAAP) descriptor:

python python iFeature.py --file examples/test-protein.txt --type CKSAAP

Generally, users can generate different descriptors by changing the descriptor type by specifying '--type', For example, run the following command to generate the Dipeptide Deviation from Expected Mean (DDE) descriptor:

python python iFeature.py --file examples/test-protein.txt --type DDE

For some descriptors (e.g. PSSM, Disorder, ASA, TA and SSEB), the predicted protein property file should be supplied by the --path parameter. For example, run the following command to generate PSSM descriptor:

python iFeature.py --file examples/test-peptide.txt --type PSSM --path examples/predictedProteinProperty

KNNprotein and KNNpeptide descriptors requires an extra training file and a label file, which is spedified by --train and --label. Run the following command to generate the KNNprotein descriptor:

python iFeature.py --file examples/test-peptide.txt --type KNNpeptide --train examples/train-peptide.txt --label examples/label.txt

For the six descriptors in Amino Acid Composition group, user can specify the output order by --order and --userDefinedOrder, three amino acids order (i.e. alphabetically, polarity and side chaim volume) were supplied by iFeature. Run the following command to generate the AAC descriptor with the 'polarity' order:

python iFeature.py --file examples/test-protein.txt --type AAC --order polarity

Run the following command to generate the AAC descriptor with a user-defined order:

python iFeature.py --file examples/test-protein.txt --type AAC --order userDefined --userDefinedOrder YWVTSRQPNMLKIHGFEDCA

For some of the descriptors, user can adjust the default parameters by advanced usage. The detatiled advanced usage for each type of descriptor can be found in iFeatureManual.pdf in iFeature directory. For example, for CKSAAP descriptor, advanced users can adjust the size of the sliding window to 3 (the default is 5) by running the following Python command:

python codes/EAAC.py examples/test-peptide.txt 3 EAAC.tsv

The default output file is encoding.tsv, which can be specified by --out. For example:

python iFeature.py --file examples/test-protein.txt --type AAC --out AAC.txt

Examples for users to extract descriptors from iFeaturePseKRAAC.py.

The 16 types of reduced amino acid alphabets with different clustering approaches can be used to generate different versions of pseudo reduced amino acid compositions (PseRAACs).

The following parameters are required by iFeaturePseKRAAC.py:

• --file protein/peptide sequence file in fasta format
• --type descriptor type in the PseKRAAC feature group
• feature types for protein sequence analysis, two alternative modes (g-gap and lambda-correlation) are available, with the ‘g-gap’ model as the default.
• K-tuple value, three K-tuple values (i.e. 1, 2 and 3) are available, default is 2
• --gap_lambda gap value for the g-gap model or lambda value for the `lambda-correlation’ model, 10 values are available (i.e. 0, 1, 2, …, 9)
• the reduced amino acids cluster type. Users can run the following command to view the available values for each descriptor type:

python iFeaturePseKRAAC.py --show

Use the following command to extract the PseKRAAC feature descriptors:

python iFeaturePseKRAAC.py --file examples/test-protein.txt --type type1 --subtype lambda-correlation --ktuple 2 --gap_lambda 2 --raactype 5

Examples for Feature analysis.

iFeature integrates several commonly used and useful clustering, feature selection and dimensionality reduction algorithms. In order to facilitate the understanding of the results for non-experts, a scatter diagram will be ploted to show the distribution of the cluster result. When ploting the scatter diagram, the t-SNE algorithm was used to reduce the high-dimensional descriptors to 2 dimensions, and then the scatter diagram was plotted by using the reduced descriptors and clustering result. The clustering result will be also stored in a file with text format. The file contained the information of clusters number and ownership for each sample. The clustering algorithms can be run by the following command:

python cluster.py --file descriptor.tsv --type <clustering_algorithm> --sof <sample/feature>

• --file is the descriptor output file, which is generated by iFeature.py or iFeaturePseKRAAC.py
• --type specifies the clustering algorithm (kmeans, hcluster, apc, meanshift and dbscan)
• --sof specifies the option for performing clustering for samples or features (default: samples) Use the following command to perform the K-Means clustering:

python cluster.py --file examples/example.tsv --type kmeans --sof feature --nclusters 2

The feature selection algorithms can be run by the following command:

python feaSelector.py --file descriptor.tsv --type <feature_selection_algorithm> --label sample_label_file

• --label specify the sample class of the samples in descriptor.tsv. Use the following command to perform the Chi-Square feature selection:

python feaSelector.py --file examples/example.tsv --type CHI2 --label examples/label.txt --out CHI2_feature.txt

After running the feature selection algorithm the descriptors will be ranked according to their importance, the higher the ranking is, the more important the descriptor is.

In addition, three dimensionality reduction algorithms (PCA,LDA and t-SNE) have been implemented in the iFeature package and can be run using the following commands. To facilitate understanding, the dimensionality reduction result can be visualized by the scatter diagram.

Use the following command to perform the PCA analysis:

python scripts/pcaAnalysis.py --file examples/example.tsv --ncomponents 3 --out pcaResult.txt

• --ncomponents specifies the number of principal components, --out specifies the name of the output file of PCA. Use the following command to perform the LDA analysis:

python scripts/ldaAnalysis.py --file examples/example.tsv --ncomponents 3 --label examples/label.txt --out ldaResult.txt

Use the following command to perform the t-SNE analysis:

python scripts/tsneAnalysis.py --file examples/example.tsv

Citation：

If you find iFeature useful, please kindly cite the following paper:

Zhen Chen, Xuhan Liu, Pei Zhao, Chen Li, Yanan Wang, Fuyi Li, Tatsuya Akutsu, Chris Bain, Robin B. Gasser, Zuoren Yang*, Lukasz Kurgan*, Jiangning Song*, iFeatureOmega – an integrative platform for the feature engineering, visualization and analysis of features from molecular sequence, structural and ligand data sets. Nucleic Acids Research , 2022. https://doi.org/10.1093/nar/gkac351

Zhen Chen, Pei Zhao, Fuyi Li, André Leier, Tatiana T Marquez-Lago, Yanan Wang, Geoffrey I Webb, A Ian Smith, Roger J Daly*, Kuo-Chen Chou*, Jiangning Song*, iFeature: a Python package and web server for features extraction and selection from protein and peptide sequences. Bioinformatics, 2018, 34(14): 2499–2502. https://doi.org/10.1093/bioinformatics/bty140

Zhen Chen, Pei Zhao, Fuyi Li, Tatiana T Marquez-Lago, André Leier, Jerico Revote, Yan Zhu, David R Powell, Tatsuya Akutsu, Geoffrey I Webb, Kuo-Chen Chou, A Ian Smith, Roger J Daly, Jian Li, Jiangning Song*, iLearn: an integrated platform and meta-learner for feature engineering, machine-learning analysis and modeling of DNA, RNA and protein sequence data. Briefings in Bioinformatics, 2020, 21(3): 1047–1057. https://doi.org/10.1093/bib/bbz041

Zhen Chen, Pei Zhao, Chen Li, Fuyi Li, Dongxu Xiang, Yong-Zi Chen, Tatsuya Akutsu, Roger J Daly, Geoffrey I Webb, Quanzhi Zhao*, Lukasz Kurgan*, Jiangning Song*, iLearnPlus: a comprehensive and automated machine-learning platform for nucleic acid and protein sequence analysis, prediction and visualization. Nucleic Acids Research , 2021;, gkab122, https://doi.org/10.1093/nar/gkab122