- A Unix-style platform (ideally, Linux)
- Docker
-
Install Docker on your system.
-
Build our Docker image (it is large, at ~15 GB).
git clone https://github.com/GreenCUBIC/MethylSight2.git
cd MethylSight2
docker build -t methylsight2 .- Create an empty (blank) output file where the results should be stored.
touch /path/to/my/results/file.csv- Run MethylSight 2.0
If a GPU is available:
input=<ABSOLUTE_PATH_TO_INPUT_FILE>;output=<ABSOLUTE_PATH_TO_OUTPUT_FILE>;docker run -v "$input:/input.fasta:ro" --mount type=bind,source="$output",target="/output.csv" --gpus 1 methylsight2 "/env/bin/python3 model.py -i /input.fasta -w weights.ckpt -o /output.csv"If no GPUs are available:
input=<ABSOLUTE_PATH_TO_INPUT_FILE>;output=<ABSOLUTE_PATH_TO_OUTPUT_FILE>;docker run -v "$input:/input.fasta:ro" --mount type=bind,source="$output",target="/output.csv" methylsight2 "/env/bin/python3 model.py -i /input.fasta -w weights.ckpt -o /output.csv"The model weights can be downloaded here.
curl -X POST -H "Content-Type: application/json" -d '{"sequence": "SEQUENCE"}' https://fcharih-methylsight2.hf.space/predict