Weight_EN

• Weight_elasticnet uses a statistic regularization method by integrating structure information and overlap of multiple pathways for analysis of high-dimensional gene expression data.

Installation

• It is required to install the following dependencies in order to be able to run the package: For the weighted elastic net method of the package named weight_elasticnet

python >= 3.5
numpy = 1.18.4
pandas = 0.23.4
joblib = 1.0.0
scipy >= 1.4.1
scikit-learn >= 0.23.2

You can also clone the repository and do a manual install.

git clone https://github.com/huanheaha/Weight_EN.git
python setup.py install

Running the Code

• Acquire all the data and code in Weight_EN to the local address.
• Open Python editor,then dictory to Weight_EN folder which contains example.py.
• Make sure the data and code are in one folder,or enter the exact data address when you run the code.

An example of applying the Stacked SGL is provided in example.py.

Running the example in example.py. Specific example including parameter description and parameter selection is provided in this file.

from weight_elasticnet.Weight_EN import LogisticRegElastic
logit = LogisticRegElastic()
coef_path = logit.fit(matrix, alpha, wj, precision = 0.0001,
                          lambda_grid =[lambd], pyspark=False)
"""matrix: inpute training matrix
alpha: α∈[0,1] which determines the relative weight of the l_1 and l_2 norm.
wj: feature weight
lambd: regularization parameter
"""                     
y_pre = logit.predict(X_test, pyspark=False)
"""y_pre: prediction probability of independent test set
"""  

Prediction results will show in the terminal.

print(y_pre[:10])
[0.77605729 0.49160284 0.77872351 0.66705835 0.58397401 0.73701146
 0.76280227 0.62244287 0.48524927 0.62714534]

The parameters alpha and lambd can be adjusted according to user requirements.

Announcements

• The input data set should be standardized.
• The csv files of LIHC_kegg_path.csv, THCA_kegg_path.csv and Lung_kegg_path.csv represent the final mRNA gene expression data of three cancer genes trained in the model. The csv file of gene_weight.csv is gene weight data, i.e. [degree,rank,1/rank].
• For example, the optimal parameters are trained in THCA_best.model.You can run predict.py to obtain the prediction results of the real data.