GOEGCN_BRCA_Subtypes

Paper: RNA-Seq-based breast cancer subtype classification using machine learning approaches

###################### 1. Data ##############################

Please extract the Data compressed file to this "Data" folder.

ds1_brca_clinic.Rdata: The clinic data of BRCA

ds1_brca_tumor_counts.Rdata: The BRCA tumor Counts data with the same genes and samples as the BRCA tumor FPKM data of 3)

ds1_brca_tumor_fpkm.Rdata: The tumor data of BRCA after filtering out the genes whose mean values are less than 0.2 and variations are less than 2 across the tumor samples.

ref_brca_normal_counts.Rdata: The BRCA normal Counts data with same genes as 2), the data is for just for reference, and itis not used in this paper.

ref_brca_normal_fpkm.Rdata: The BRCA normal FPKM data with same genes as 3), the data is for just for reference, and itis not used in this paper.

###################### 2. Main functions ####################

There are 6 main functions and the corresponding descriptions are as follows,

1. Main1_weighted_DEGs.R

This main function is used to obtain the final weighted DEGs for classification and pathway enrichment analysis.

2. Main2_classifier.R

This function is adopted to conduct the BRCA subtypes classification using weighted DEGs (for classification) Learn the corresponding classifiers in training set and evaluate the performance in the testing set for each subtype

3. Main2_classifier_ROC.R

This function is similar with the "Main2_classifier.R", the difference is the "ROC" metric is adopted when training the model. This function can plot the corresponding ROC curves and the ROC curves for each subtype are saved at directory "ROC_graphs".

4. Main2_classifier_ROC_Multiple.R

This function is similar with the "Main2_classifier_ROC.R". This function can plot the multiple ROC curves in one figure and the ROC curves including different machine learning approaches for each subtype are saved at directory "ROC_graphs".

5. main3_GeneCouples_pathways.R

This is the subfunction which is called by Main3_GOEGCN.R. This function is used to conduct the GO enrichment analysis using GOEGCN with weighted DEGs (for GO enrichment analysis)

6. Main3_GOEGCN.R

This function is used to conduct the GO enrichment analysis using GOEGCN with weighted DEG, and the save the corresponding enriched GO terms. The results for each subtype is saved into Excel file and the list results are saved into Rdata file.

###################### 3. Supplementary file ####################

Please extract the supplementary compressed file to this "supplementary" folder. These the are supplementary files which are needed in this papper.

If you think this research is useful to you, please cite the following article,

Zhezhou Yu, Zhuo Wang, Xiangchun Yu*, and Zhe Zhang, RNA-Seq-based breast cancer subtypes classification using machine learning approaches, Computational Intelligence and Neuroscience.