mgcnn_alkaloid: Molecular Graph Convolutional Neural Networks for Prediction of Alkaloid Biosynthesis Bathways
Source code for our paper "Classification of alkaloids into starting substances on biosynthetic pathway using graph convolutional neural networks".
Python >= 3.6.6 NumPy >= 1.15.4 Pandas >= 0.22.0 TensorFlow 1.6.0 DeepChem 2.1.0
- mgcnn_alkaloid.py: Training and evaluation of Molecular Graph Convolutional Neural Networks for alkaloid biosynthesis pathway prediction.
- data/alkaloid_data.csv: Molecular data of alkaloids.
- Five fold cross validation
python alkaloid_data.csv
- Verify one segment of five fold cross validation
python alkaloid_data.csv [0-4]
Naoaki ONO
[Currently under review]