breast

breast

First download data from TCGA using

mkdir -p data && mv manifest.txt data/. && cd data
gdc-client download -m manifest.txt

Run breast.ipynb to preprocess data

Go trough predictor_BRCA.ipynb to execute the predictor discussed in the paper, this can be run on Colab on in any machine with tensorflow (possibly with GPU enabled)

model_Subtype_Selected_Lum.h5 contains the trained weights, load it using

from tensorflow.keras.models import load_model
model = load_model(f"model_{label}_{L}.h5", custom_objects={'f1': f1})
model.summary()

as described in the notebook, to reproduce our evaluation analyses.

Output for hierarchical stochastic block model is in topsbm folder