User Guide

Contact:

Baoshan Ma(mabaoshan@dlmu.edu.cn)

Bingjie Chai(qfcbj0612@163.com)

Citation

Baoshan Ma, Bingjie Chai, Heng Dong, Jishuang Qi, Pengcheng Wang, Tong Xiong, Yi Gong, Di Li, Shuxin Liu, Fengju Song.Diagnostic classification of cancers using DNA methylation of paracancerous tissues (Under review)

1.Introduction

We provided a python program to build a classification model for separating early stage and late stage cancers. The classification model can then be applied to a new dataset. If you find the program useful, please cite the above reference.

2.Software requirement

python 3.7 version

3.The datasets of the program

DNA methylation data used in this research were collected from The Cancer Genome Atlas(TCGA) project and that are publicly available at https://portal.gdc.cancer.gov. The 'input data' file contains 'data.txt' and 'label.txt'. Rows and columns of the 'data.txt' correspond to samples and CpG sites, respectively.

4.How to use our program and obtain output metrics

(1) In the 'input_data' folder, you need to upload a txt file of DNA methylation data of paracancerous tissues and name it 'data.txt'. Besides, you need upload a txt file of sample label(early stage or late stage) and name it 'label.txt'. For 'data.txt', we recommend that you can set each row as a sample and each column as a CpG site. In the 'multiclass_input_data' folder, you need provide DNA methylation data of paracancerous tissues from some types of cancer and integrate them. You can download these data from TCGA and GEO database and name them 'tcga.txt' and 'geo.txt', respectively. Rows and columns of the data correspond to samples and CpG sites, respectively. Besides, you need provide 'tcga_label.txt' and 'geo_label.txt' that include the label of the corresponding cancers to realize multi classification of cancer types, respectively. 'tcgadata.txt' denotes tcga datasets after feature selection. 'tcgalabel.txt' is the corresponding file of 'tcgadata.txt'.

(2) Now, you can run the model in the 'machine_learning_models' folder. For each model, you need to input the data and label in turn.

(3) After the program is executed, you can obtain evaluation metrics of the model.

(4) In the 'tumor_specific_multiclass_classifier' folder, you can run 'model_tcga.txt' where tcga datasets are available. You can run 'model_geo' where TCGA and GEO datasets are available.