GPDBN: Deep bilinear network integrating both genomic data and pathological images for breast cancer prognosis prediction

This is an implementation of GPDBN in Python 2.7.12 under Linux with CPU Intel Xeon 4110 @ 2.10GHz, GPU NVIDIA GeForce RTX 2080 Ti, and 192GB of RAM. It needs Keras libraries to be installed.

It uses Keras library with the Tensorflow backend, and does not work on the Theano backend, as the loss function of the network is written with Tensorflow.

Installation

git clone https://github.com/isfj/GPDBN.git
cd GPDBN

Running Experiments

Our proposed GPDBN framework is in the GPDBN/model/inter_intra.py directorty. Before running the experiments, make sure the required environment is configured. After configuring the required environment, you can run GPDBN by the following codes

cd GPDBN/model
python inter_intra.py

Dataset

Breast cancer patient samples adopted in this study include matched digital whole-slide images and gene expression profiles, which are acquired from The Cancer Genome Atlas (TCGA) data portal.

Each row of 'mRNA_data_slct32.xlsx' and 'patho_data_slct32.xlsx' represents the feature vector of one patient,and there are totally 345 patients.

Contact

Please feel free to contact us if you need any help: ustc023w@mail.ustc.edu.cn

All rights reserved