PyFeat: A Python-based Effective Feature Generation Tool from DNA, RNA, and Protein Sequences

Authors: Rafsanjani Muhammod, Sajid Ahmed, Dewan Md Farid, Swakkhar Shatabda, Alok Sharma, and Abdollah Dehzangi

 

1. Download Package

1.1. Direct Download

We can directly download by clicking the link.

Note: The package will download in zip format (.zip) named PyFeat-master.zip.

or,

1.2. Clone a GitHub Repository (Optional)

Cloning a repository syncs it to our local machine (Example for Linux-based OS). After clone, we can add and edit files and then push and pull updates.

• Clone over HTTPS: user@machine:~$ git clone https://github.com/mrzResearchArena/PyFeat.git
• Clone over SSH: user@machine:~$ git clone [email protected]:mrzResearchArena/PyFeat.git

Note #1: If the clone was successful, a new sub-directory appears on our local drive. This directory has the same name (PyFeat) as the GitHub repository that we cloned.

Note #2: We can run any Linux-based command from any valid location or path, but by default, a command generally runs from /home/user/.

Note #2.1: user is the name of our computer but your computer name can be different (Example: /home/bioinformatics/).

2. Installation Process

2.1. Required Python Packages

Major (Generate Features):

• Install: python (version >= 3.5)
• Install: numpy (version >= 1.13.0)

Minor (Performance Measures):

• Install: sklearn (version >= 0.19.0)
• Install: pandas (version >= 0.21.0)
• Install: matplotlib (version >= 2.1.0)

2.2. How to download

Using PIP: pip install <package name>

user@machine:~$ pip install scikit-learn

or,

Using anaconda environment: conda install <package name>

user@machine:~$ conda install scikit-learn

3. Working Procedure

Run command on your console or terminal.

3.1. Generate Features

3.1.1. Training Purpose

user@machine:~$ python main.py --sequenceType=DNA --fullDataset=1 --optimumDataset=1 --fasta=/home/user/PyFeat/Datasets/DNA/FASTA.txt --label=/home/user/PyFeat/Datasets/DNA/Labels.txt --kTuple=3 --kGap=5 --pseudoKNC=1 --zCurve=1 --gcContent=1 --cumulativeSkew=1 --atgcRatio=1 --monoMono=1 --monoDi=1 --monoTri=1 --diMono=1 --diDi=1 --diTri=1 --triMono=1 --triDi=1

or,

user@machine:~$ python main.py -seq=DNA -full=1 -optimum=1 -fa=/home/user/PyFeat/Datasets/DNA/FASTA.txt -la=/home/user/PyFeat/Datasets/DNA/Label.txt -ktuple=3 -kgap=5 -pseudo=1 -zcurve=1 -gc=1 -skew=1 -atgc=1 -f11=1 -f12=1 -f13=1 -f21=1 -f22=1 -f23=1 -f31=1 -f32=1

3.1.2. Evaluation Purpose

user@machine:~$ python main.py --sequenceType=Protein --testDataset=1 --fasta=/home/user/PyFeat/Datasets/Protein/independentFASTA.txt --label=/home/user/PyFeat/Datasets/Protein/independentLabel.txt --kTuple=3 --kGap=5 --pseudoKNC=1 --zCurve=1 --gcContent=1 --cumulativeSkew=1 --atgcRatio=1 --monoMono=1 --monoDi=1 --monoTri=1 --diMono=1 --diDi=1 --diTri=1 --triMono=1 --triDi=1

or,

user@machine:~$ python main.py -seq=Protein -test=1 -fa=/home/user/PyFeat/Datasets/Protein/independentFASTA.txt -la=/home/user/PyFeat/Datasets/Protein/independentLabel.txt -ktuple=3 -kgap=5 -pseudo=1 -zcurve=1 -gc=1 -skew=1 -atgc=1 -f11=1 -f12=1 -f13=1 -f21=1 -f22=1 -f23=1 -f31=1 -f32=1

[ Comment: The = sign is optional. ]

Note #1: It will generate a full dataset named fullDataset.csv (if -full=1 or, --fullDataset==1)
Note #2: It will generate a selected features dataset named optimumDataset.csv (if -optimum=1 or, --optimumDataset==1), and It will also track the selected features index.
Note #3: It will generate a full dataset named testDataset.csv (if -test=1 or, --testDataset==1) [ Especially for the independent (testing) dataset purpose. ] [ 3.1.2. ]
Note #4: The process will run smoothly for valid FASTA sequences and row-wise class label.

 

Table 1: Arguments Details for the Features Generation

Argument	Corresponding Optional Argument	Type	Default	Help
--sequenceType	-seq	string	--sequenceType=DNA	We can use DNA, RNA, and protein or prot as option; Case is not sensitive.
--fasta	-fa	string		Enter a UNIX-like path; Example: /home/user/FASTA.txt
--label	-la	string		Enter a UNIX-like path; Example: /home/user/Label.txt
--kGap	-kgap	integer	--kGap=5	The number of gaps ranging from 1 to 5 inclusive; Example: -kGap=5
--kTuple	-ktuple	integer	--kTuple=3	The number of nucleotides ranging from 1 to 3 inclusive; Example: -kTuple=3
--fullDataset	-full	integer	--fullDataset=0	Set --fullDataset=1, if we don't want to save full dataset.
--testDataset	-test	integer	--testDataset=0	Set --testDataset=1, if we don't want to save test dataset.
--optimumDataset	-optimum	integer	--optimumDataset=0	Set --optimumDataset=1, if we don't want to save optimum dataset.
--pseudoKNC	-pseudo	integer	--pseudoKNC=0	Set --pseudoKNC=1, if we want to generate features.
--zCurve	-zcurve	integer	--zCurve=0	Set --zCurve=1, if we want to generate features.
--gcContent	-gc	integer	--gcContent=0	Set --gcContent=1, if we want to generate features.
--cumulativeSkew	-skew	integer	--cumulativeSkew=0	Set --cumulativeSkew=1, if we want to generate features.
--atgcRatio	-atgc	integer	--atgcRatio=0	Set --atgcRatio=1, if we want to generate features.
--monoMono	-f11	integer	--monoMono=0	Set --monoMono=1, if we want to generate features.
--monoDi	-f12	integer	--monoDi=0	Set --monoDi=1, if we want to generate features.
--monoTri	-f13	integer	--monoTri=0	Set --monoTri=1, if we want to generate features.
--diMono	-f21	integer	--diMono=0	Set --diMono=1, if we want to generate features.
--diDi	-f22	integer	--diDi=0	Set --diDi=1, if we want to generate features.
--diTri	-f23	integer	--diTri=0	Set --diTri=1, if we want to generate features.
--triMono	-f31	integer	--triMono=0	Set --triMono=1, if we want to generate features.
--triDi	-f32	integer	--triDi=0	Set --triDi=1, if we want to generate features.

     

Table 2: Feature Description

Feature Name	Feature Structure / Formula	Number of Features	Applicable
zCurve	x_axis = (A+G)-(C+T); y_axis = (A+C)-(G+T); z_axis = (A+T)-(G+C)	3 features for DNA/RNA	DNA, RNA
gcContent	( (G+C)/(A+C+G+T) ) x 100 %	1 features for DNA/RNA	DNA, RNA
atgcRatio	(A+T)/(G+C)	1 features for DNA/RNA	DNA, RNA
cumulativeSkew	gcSkew=(G-C)/(G+C); atSkew=(A-T)/(A+T)	2 features for DNA/RNA	DNA, RNA
pseudoKNC	X, XX, XXX	when --kTuple=3, 84 features for DNA/RNA and 8,420 features for protein	DNA, RNA, Protein
monoMonoKGap	X_X	when --kGap=1, 16 features for DNA/RNA and 400 features for protein	DNA, RNA, Protein
monoDiKGap	X_XX	when --kGap=1, 64 features for DNA/RNA and 8,000 features for protein	DNA, RNA, Protein
monoTriKGap	X_XXX	when --kGap=1, 256 features for DNA/RNA and 160,000 features for protein	DNA, RNA, Protein
diMonoKGap	XX_X	when --kGap=1, 64 features for DNA/RNA and 8,000 features for protein	DNA, RNA, Protein
diDiKGap	XX_XX	when --kGap=1, 256 features for DNA/RNA and 160,000 features for protein	DNA, RNA, Protein
diTriKGap	XX_XXX	when --kGap=1, 1024 features for DNA/RNA and 3,200,000 features for protein	DNA, RNA, Protein
triMonoKGap	XXX_X	when --kGap=1, 256 features for DNA/RNA and 160,000 features for protein	DNA, RNA, Protein
triDiKGap	XXX_XX	when --kGap=1, 1024 features for DNA/RNA and 3,200,000 features for protein	DNA, RNA, Protein

Note: When sequence becomes DNA, RNA, and Protein then X = {A,C,G,T}, X = {A,C,G,U}, and X = {A,C,D,E,F,G,H,I,K,L,M,N,P,Q,R,S,T,V,W,Y} respectively.

Arguments Details for the Features Generation and Feature Description are provided in Tables 1 and Table 2, respectively.

     

3.2. Run Machine Learning Classifiers (Optional)

user@machine:~$ python runClassifiers.py --nFCV=10 --dataset=optimumDataset.csv --auROC=1 --boxPlot=1

Note #1: It will provide classification results (evaluationResults.txt) from the user provides binary class dataset (.csv format).
Note #2: Generate a ROC Curve (auROC.png).
Note #3: Generate an accuracy comparison via boxPlot (AccuracyBoxPlot.png).

 

Table 3: Arguments Details for the Machine Learning Classifiers

Argument	Corresponding Optional Argument	Type	Default	Help
--nFCV	-cv	integer	--nFCV=10	How many numbers of cross-validation?
--dataset	-data	string	--dataset=optimumDataset.csv	Enter a UNIX-like path for a .csv file; Example: /home/User/dataset.csv
--auROC	-roc	integer	--auROC=1	Set --auROC=0, if we didn't want to generate the ROC Curve.
--boxPlot	-box	integer	--boxPlot=1	Set --boxPlot=0, if we didn't want to generate the accuracy box-plot.

 

3.3. Training Model (Optional)

user@machine:~$ python trainModel.py --dataset=optimumDataset.csv --model=LR

Note #1: It will provide a dumpModel.pkl from the user provides binary class dataset (.csv format).

 

Table 4: Arguments Details for Training Model

Argument	Corresponding Optional Argument	Type	Default	Help
--dataset	-data	string	--dataset=optimumDataset.csv	Enter a UNIX-like path for a .csv file; Example: /home/User/dataset.csv
--model	-m	string	--model=LR	We can use LR, SVM, KNN, DT, SVM, NB, Bagging, RF, AB, GB, and LDA as an option; All options are case sensitive.
--K	-k	integer	--K=5	Only for the KNN classifier; Number of neighbor

Note: LR, SVM, KNN, DT, NB, Bagging, RF, AB, GB, and LDA represents Logistics Regression, Support Vector Machine, k-Nearest Neighbor, Decision Tree, Naive Bayes, Bagging, Random Forest, AdaBoost, Gradient Boosting, Linear Discriminant Analysis classifier respectively.

     

3.4. Evaluation Model (Optional)

user@machine:~$ python evaluateModel.py --optimumDatasetPath=optimumDataset.csv --testDatasetPath=testDataset.csv

Note #1: Here, optimumDataset.csv, and testDataset.csv using as a traing dataset and test dataset respectively.

 

Table 5: Arguments Details for Evaluation Model

Argument	Corresponding Optional Argument	Type	Default	Help
--optimumDatasetPath	-optimumPath	string	--optimumDatasetPath=optimumDataset.csv	Enter a UNIX-like path for a .csv file; Example: /home/User/dataset.csv
--testDatasetPath	-testPath	string	--testDatasetPath=testDataset.csv	Enter a UNIX-like path for a .csv file; Example: /home/User/dataset.csv

     

References

[1] Bin Liu, Fule Liu, Longyun Fang, Xiaolong Wang, and Kuo-Chen Chou. repdna: a python package to generate various modes of feature vectors for dna sequences by in- corporating user-defined physicochemical properties and sequence-order effects. Bioin- formatics, 31(8):1307–1309, 2014.

[2] Dong-Sheng Cao, Qing-Song Xu, and Yi-Zeng Liang. propy: a tool to generate various modes of chous pseaac. Bioinformatics, 29(7):960–962, 2013.

[3] 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, et al. ifeature: a python package and web server for features extraction and selection from protein and peptide sequences. Bioinformatics, 1:4, 2018.

[4] Md Rafsan Jani, Md Toha Khan Mozlish, Sajid Ahmed, Dewan Md Farid, and Swakkhar Shatabda. irecspot-ef: Effective sequence based features for recombination hotspot prediction. Computers in biology and medicine, 2018.

[5] Bin Liu, Fule Liu, Longyun Fang, Xiaolong Wang, and Kuo-Chen Chou. reprna: a web server for generating various feature vectors of rna sequences. Molecular Genetics and Genomics, 291(1):473–481, 2016.

[6] Nan Xiao, Dong-Sheng Cao, Min-Feng Zhu, and Qing-Song Xu. protr/protrweb: R package and web server for generating various numerical representation schemes of pro- tein sequences. Bioinformatics, 31(11):1857–1859, 2015.

[7] Bin Liu. Bioseq-analysis: a platform for dna, rna and protein sequence analysis based on machine learning approaches. Briefings in bioinformatics, 2017.

[8] Bin Liu, Hao Wu, Deyuan Zhang, Xiaolong Wang, and Kuo-Chen Chou. Pse-analysis: a python package for dna/rna and protein/peptide sequence analysis based on pseudo components and kernel methods. Oncotarget, 8(8):13338, 2017.

[9] Bin Liu, Fule Liu, Xiaolong Wang, Junjie Chen, Longyun Fang, and Kuo-Chen Chou. Pse-in-one: a web server for generating various modes of pseudo components of dna, rna, and protein sequences. Nucleic acids research, 43(W1):W65–W71, 2015.

[10] Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python. Journal of machine learning research, 12(Oct):2825–2830, 2011.

[11] Eamonn Keogh and Abdullah Mueen. Curse of dimensionality. In Encyclopedia of Machine Learning and Data Mining, pages 314–315. Springer, 2017.

[12] Ruihu Wang. Adaboost for feature selection, classification and its relation with svm, a review. Physics Procedia, 25:800–807, 2012.

[13] Hao Lin, En-Ze Deng, Hui Ding, Wei Chen, and Kuo-Chen Chou. ipro54-pseknc: a sequence-based predictor for identifying sigma-54 promoters in prokaryote with pseudo k-tuple nucleotide composition. Nucleic acids research, 42(21):12961–12972, 2014.

[14] Wei Chen, Pengmian Feng, Hui Ding, and Hao Lin. Pai: Predicting adenosine to inosine editing sites by using pseudo nucleotide compositions. Scientific reports, 6:35123, 2016.

[15] Hao Lin, Zhi-Yong Liang, Hua Tang, and Wei Chen. Identifying sigma70 promoters with novel pseudo nucleotide composition. IEEE/ACM transactions on computational biology and bioinformatics, 2017.

[16] Mahmoud Ghandi, Dongwon Lee, Morteza Mohammad-Noori, and Michael A Beer. Enhanced regulatory sequence prediction using gapped k-mer features. PLoS computa- tional biology, 10(7):e1003711, 2014.

[17] Shahana Yasmin Chowdhury, Swakkhar Shatabda, and Abdollah Dehzangi. Idnaprot- es: Identification of dna-binding proteins using evolutionary and structural features. Scientific Reports, 7(1):14938, 2017.