Project for course "Machine Learning for Genomics" at ETH.
Dependencies for conda and pip are listed in environment.yml and requirements.txt.
The main notable requirement is python=3.9.0 due to 3.10 not working with some other requirements.
environment.yml contains my whole setup, however, most is macOS Metal specific libraries.
Every python dependency is listed in requirement.txt without nested dependencies.
tensorflow-macos and tensorflow-metal are also system specific dependencies, and varies for other systems
(installation guide).
Most reliable way to replicate environment:
conda create -n gene_exp_rnn python=3.9
conda activate gene_exp_rnn
pip install -r requirements.txt
(..and install tensorflow according to guide above and remove other tensorflow dependencies if not using Metal.)
The project data should be unzipped into ./data. For example, the X1 dataset train info should be available at ./data/CAGE-train/X1_train_info.tsv.
- Python 3.9.0
- Histone modification data processing with pyBigWig.