Temporally Resolved Gene Co-expressions with Semantic Modelling

This repo is dedicated to the share the final project code for CSC 2611 🌟

• The model training and validation code can be found in the main folder:

  • co_exp_lineage_wandb_DiffTime.py for the Co-expression Prediction model training and testing
  • co_exp_lineage_wandb_ExpEst.py for the Expression Estimation model training and testing
  • coexp_change_inference.ipynb for time-course construction, change point detection, Pearson correlation calculation and result visualizations

• The datasets can be found in the ./datasets folder:

  • ./dyn-LI-2000-1 for the Linear dataset
  • ./dyn-BF-2000-1 for the Birfurcation dataset
  • ./dyn-TF-2000-1 for the Trifurcation dataset
  • The datasets were produced by Pratapa et al. and retrieved from https://zenodo.org/record/3701939#.Y5a5huzMJwo.

• The trained models can be found in the ./models folder:

  • There are two models associated with each dataset, with a total of six sub-folders
  • The Co-expression Prediction model sub-folders are named with "DiffTime", and the Expression Estimation model sub-folders named with "ExpEstimator"
  • In coexp_change_inference.ipynb, the model from the latest epoch is used for inference

• requirements.txt includes all packages needed for this repo

Please feel free to let me know if any issues. Thank you for reviewing! 🙌 😃