This is a repository containing all ME-models reconstructed using coralME.

Automatically reconstructed ME-models (dME-models) are available in clean/, and updated ME-models (uME-models) are available in published/. Code to reproduce Figures of our manuscript and to reconstruct ME-models are provided here as python scripts and Jupyter notebooks.

The image below explains the structure of the dME- and uME-model folders. We provide JSON files for minimal (step1), fully reconstructed (step2) and troubleshot (step3) ME-models. For uME-models, step2 and step3 are the same model, just that step3 will have a solution property.

git clone --depth 1 <repository>

1. Go to the desired directory, clean/ or published/.
2. Open Supplementary Notebook and follow instructions.

1. Go to the desired directory, e.g. cd clean/ or cd published/
2. run bash build.sh 0 1 1 1 ORGANISM_NAME

ORGANISM_NAME must exist in directory. The four numbers are binary flags to run (if 1) / not run (if 0):

1. BLASTp
2. Data synchronization, complementation and OSM generation
3. ME-model reconstruction
4. ME-model troubleshooting.

Since we provide here BLASTp results, the first number can be set to 0. Setting it to 1 will run a BLASTp with 4 CPU cores and overwrite files in blast_files_and_results under each organisms directory.

Important: ME-models are saved as Pickle files.

1. Go to the desired directory, e.g. cd clean/ or cd published/
2. run bash parallel.sh 0 1 1 1. The four numbers represent the same flags as described before. Modify N in parallel.sh to set number of cores, default 4.

In a python environment, run

model = coralme.io.pickle.load_pickle_me_model("PATH/TO/ORGANISM_MEMODEL_FILE.pkl")

1. run gzip -d PATH/TO/ORGANISM_MEMODEL_FILE.json.gz to extract the JSON file.
2. In a python environment, run

model = coralme.io.json.load_json_me_model("PATH/TO/ORGANISM_MEMODEL_FILE.json")