LEMP (LSTM-based Ensemble Malonylation Predictor) Version 1.0
By Xuhan Liu & Zhen Chen, on July 1st 2017
Please see the LICENSE file for the license terms for the software. Basically it's free to academic users. If you do wish to sell the software or use it in a commercial product, then please contact the following E-mails,
Xuhan Liu: [email protected]
or
Zhen Chen: [email protected]
Firstly, ensure that the version of your Python >= 2.7.
Secondly, all the following packages are installed in your machine:
1. Numpy:
tcsh% pip install numpy
2. Scikit-learn
tcsh% pip install scikit-learn
3. H5py
tcsh% pip install h5py
4. Tensorflow
tcsh% pip install tensorflow (CPU version)
tcsh% pip install tensorflow_gpu (GPU version)
5. Keras
tcsh% pip install keras
Finally decompressing the software.
tcsh% tar -zxvf lemp-1.0.tar.gz
The executables will be placed in the this directory.
If you want to predict your own protein sequence, execute as follows:
python lemp.py -i <input_file> [-o <output_file>] [-h] [-v]
If you want to rebuild the new model with your own dataset,
1). Construct your dataset that has the same format as ours in 'dataset/chen_train.txt' and 'dataset/chen_test.txt',
which is training set and independent set, respectively.
2). excute following shell:
python model.py [-t <training_set>] [-i <independent_set>] [-d] [-n <int>] [-e <int>]
lemp.py
-i <input_file> : dataset file containing protein sequences as FASTA file format.
-o <output_file> : a directory containing the results of prediction of each sample.
-v : version information of this software.
-h : Help information, print USAGE and ARGUMENTS messages.
Note: Please designate each protein sequence in FASTA file with distinct name!
model.py
-t <training_set> : the file of training set, which must have the same format as "dataset/chen_train.txt".
-i <independent_set> : the file of independent set, which must have the same format as "dataset/chen_test.txt".
-d : rebuilt the LSTM-based deep learning model simultaneously (very time-consuming).
-n <int> : the number of CPU to train the model.
-e <int> : the number of trees in random forest classifier.
-v : Version information of this software.
-h : Help information, print USAGE and ARGUMENTS messages.
Note: Because maximum of file size in Github is 25M, our random forest model only contain 100 trees, if you want get more precise results, please run
python model.py
without any parameters to rebuild the random forest model.
lemp.py
Output columns : <Seq_ID> <Site> <Residue> <Score> <Y/N(sp=90%)> <Y/N(sp=95%)> <Y/N(sp=99%)>
1. Seq ID : The ID of protein sequence, which is as same as in the fasta file.
2. Site : The position of the residue at the protein sequence.
3. Residue : The type of residue which is possibly malonylated. In common, it must be lysine (i.e. "K").
4. Score : The score that computed by the LEMP.
5. Y/N(sp=90%) : Judging whether this residue was malonylation site or not under the specification = 90%.
5. Y/N(sp=95%) : Judging whether this residue was malonylation site or not under the specification = 95%.
5. Y/N(sp=99%) : Judging whether this residue was malonylation site or not under the specification = 99%.
Note: The larger value of specification under which it was judged as malonylation site, The higher confidence of the result will be.
Zhen Chen, Ningning He, Yu Huang, Wen Tao Qin, Xuhan Liu*, Lei Li*. Integration of A Deep Learning Classifier with A Random Forest Approach for Predicting Malonylation Sites. Genomics, Proteomics & Bioinformatics; ISSN(2019) pp. 1672-0229