Protein subcellular localization prediction using graph convolutional network and protein embedding

• /preprocess contains scripts for preprocessing data.
• /data contains preprocessed data and some data statistics.
• /model contains model definitions, scripts to run models and a performance evaluation pipeline.
• /result contains summary plots for model performance.
• /report contains latex source and pdf rendering of project report

Scripts have been tested under the following environment:

• python 3.6.3
  • numpy 1.12.1
  • scipy 1.1.0
  • sklearn 0.19.1
  • tensorflow 1.8.0
  • h5py 2.7.1
  • tqdm 4.23.4
• R 3.4.3
  • rhdf5 2.22.0
  • Biostrings 2.46.0
  • pROC 1.12.1
  • mccr 0.4.4
  • dplyr 0.7.4
  • reshape2 1.4.3
  • ggplot2 2.2.1