dA.cpp: implementation of the denoising autoencoder Compile: g++ -o spuda dA.cpp ./spuda -h

analysis: folder containing raw data, training results, code for analysis.

data: the data folder These data will be loaded automatically in the relevant R scripts

Figures: figures created by the codes will be stored in this folder genes_hidden_units: selected genes will be written into this folder for GO or network analysis, etc.

0_load_data.R: load and process data (processed data saved in data.RData) 1_gsea.R: run GSEA on the hidden units 2_predict_severity.R: prediction of severity (results saved in data.RData) 3_predict_clinical.R: prediction of FEV1/FVC ratio

plot_clinical.R: plot correlation between clinical traits plot_GO.R: plot p-value for gene ontology results (downloaded from https://david.ncifcrf.gov/) plot_network: plot gene interaction network (downloaded from https://string-db.org/)