michae-l / deepdn-iglu Goto Github PK

To facilitate the access to glutarylation site prediction data, DeepDN_iGlu has been implemented as an available web server ( https://bioinfo.wugenqiang.top/~smw/DeepDN_iGlu/ ).

The directory structure of this repository is as follows：
Images used in the paper: All images used in the paper
gitproject: gitproject is the project code used for this research
----subdir:dataset: benchmark data set
----subdir:feture_method: including feature encoding method and feature files that has been generated
----subdir:model: the final model generated in the experiment
----subdir:result: the result information
----main file:
    data_processing.py: process the raw data
    predicting: get start to predict

supplements: The 10-fold cross-validation results for the final model and comparison models are included in supplements
----subdir:10-fold cross-validation results of final model:
--------subdir:10-fold cross-validation: Contains the model generated by ten-folds cross-validation, related pictures(loss and MCC of training and validation sets) and log of history(tf log)
--------file:statistics_eval_param.csv: the performance of 10-fold cross-validation
----subdir:models of comparison:
--------subdir:Cross entropy loss:Ten-fold cross-validation results when using the cross-entropy loss function
--------subdir:dense blocks number selection:Ten-fold cross-validation results when choosing different numbers of dense blocks