enpkg_graph_builder

Build the Experimental Natural Products Knowledge Graph

⚙️ Module part of enpkg_workflow.

The aim of this repository is to format as RDF the data produced previously.

0. Clone repository and install environment

1. Clone this repository.
2. Create environment:

   conda env create -f environment.yml

3. Activate environment:

   conda activate graph_builder

   1. Format data as RDF

For each sample, a new folder containing all the produced .ttl files will be created. The workflow is splitted into different scripts so that you can adapt it easily.

Desired architecture

data/
└─── sample_a/
      └─── rdf/
            metadata_enpkg.ttl                        # Sample's metadata
            features_pos.tll                          # Features data in PI mode
            features_neg.ttl                          # Features data in NI mode
            features_spec2vec_pos.tll                 # Features spec2vec data in PI mode
            features_spec2vec_neg.ttl                 # Features spec2vec data in NI mode
            sirius_pos.tll                            # SIRIUS data in PI mode
            sirius_neg.ttl                            # SIRIUS data in NI mode
            canopus_pos.tll                           # CANOPUS data in PI mode
            canopus_neg.ttl                           # CANOPUS data in NI mode
            isdb_pos.tll                              # ISDB data in PI mode
            isdb_neg.ttl                              # ISDB data in NI mode
            individual_mn_pos.tll                     # MN data in PI mode
            individual_mn_neg.ttl                     # MN data in NI mode
            structures_metadata.ttl                   # Structures' metadata

            {sample_a}_merged_graph.ttl               # Data from all above subgraphs

1.1 Format sample's metadata

Generate metadata_enpkg.ttl

python .\src\individual_processing\01_a_rdf_enpkg_metadata_indi.py -p path/to/your/data/directory/

1.2 Format features' data

Generate features_{ion}.ttl

python .\src\individual_processing\02_a_rdf_features_indi.py -p path/to/your/data/directory/ -ion {pos} or {neg}

1.3 Format features' spec2vec data

Generate features_spec2vec_{ion}.ttl

python .\src\individual_processing\02_b_rdf_features_spec2vec_indi.py -p path/to/your/data/directory/ -ion {pos} or {neg}

1.4 Format SIRIUS annotations

Generate sirius_{ion}.ttl

python .\src\individual_processing\03_rdf_csi_annotations_indi.py -p path/to/your/data/directory/ -ion {pos} or {neg}

1.5 Format CANOPUS annotations

Generate canopus_{ion}.ttl

python .\src\individual_processing\04_rdf_canopus_indi.py -p path/to/your/data/directory/ -ion {pos} or {neg}

1.6 Format ISDB annotations

Generate isdb_{ion}.ttl

python .\src\individual_processing\05_rdf_isdb_annotations_indi.py -p path/to/your/data/directory/ -ion {pos} or {neg}

1.7 Format MN data

Generate individual_mn_{ion}.ttl

python .\src\individual_processing\06_rdf_individual_mn_indi.py -p path/to/your/data/directory/ -ion {pos} or {neg}

1.8 Format structures' metadata

Generate structures_metadata.ttl

python .\src\individual_processing\07_rdf_structures_metadata_indi.py -p path/to/your/data/directory/ -db {The path to the structures metadata SQL DB}

1.9 Merge previously generated .tll files

Generate {sample_a}_merged_graph.ttl

python .\src\individual_processing\08_rdf_merger.py -p path/to/your/data/directory/

1.10 Export all merged .ttl files into the same folder

python .\src\individual_processing\09_rdf_exporter.py -s path/to/your/data/directory/ -t path/to/your/target/directory/ - c compress to gz (optional)

2. Steps that are dataset specific/optional

2.1. Format ChEMBL structures' metadata

python .\src\rdf_biodereplication.py -chemdb {The path to the structures metadata SQL DB} -biodb {The path to the samples ChEMBL metadata FOLDER (will integrate all ChEMBL files)} -o path/to/your/target/directory/result.ttl

3. Explore your data! :rocket:

Different graph database softwares can be used to explore your .ttl files. You can for example use GraphDB. Load the generated .ttl files (samples' merged files, the ENPKG schema, and the optional ChEMBL .ttl file(s)) and explore the chemistry of your dataset! 🌴 🦠