Getting Started

Installation

  1. Clone the repository:

    git clone <repo-url>
cd proteINET

  2. Create and activate the conda environment:

    conda env create -f environment.yml
conda activate proteINET

If not using conda, install the pip packages listed in environment.yml.

Basic Workflow

The typical workflow involves three steps:

  1. Fetch data from STRING

  2. Build network and compute metrics

  3. Generate reports and visualizations

Fetching Data from STRING

Use graph.StringLoader.StringLoader to query the STRING API:

from graph.StringLoader import StringLoader

loader = StringLoader(
    protein_query=["TP53", "BRCA1", "EGFR"],
    species=9606,  # Human
    add_nodes=20,
    required_score=400
)

# Retrieve and standardize data
data = (
    loader
    .retrieve_data()
    .standardize_data_format()
    .get_data()
)

print(data["nodes"].head())
print(data["edges"].head())

Building a Network

Use graph.PPINetwork.PPINetwork to construct and analyze the network:

from graph.PPINetwork import PPINetwork

ppin = PPINetwork(data=data)

# Compute network metrics
ppin.compute_all_global_metrics(weighted=False)
ppin.compute_all_node_metrics(weighted=False)
ppin.compute_all_edge_metrics(weighted=False)

# Access computed metrics
print(ppin.graph_metrics)
print(ppin.node_metrics_df.head())

Exporting for Cytoscape

Export the network for use in Cytoscape:

exported_files = ppin.export_for_cytoscape(
    output_dir="cytoscape_export",
    basename="ppi_network_TP53-BRCA1"
)

for file_type, path in exported_files.items():
    print(f"{file_type}: {path}")

Generating Interactive HTML Reports

Use visual.ReportGenerator.ReportGenerator to create an interactive report:

from visual.ReportGenerator import ReportGenerator

ReportGenerator().add_cytoscape_html_report(
    ppin,
    title="STRING Protein Network",
    node_color_attr="pagerank",
    node_size_attr="degree"
).generate_report_file("protein_network_report.html")

Running Enrichment Analysis (GSEA)

Compute gene set enrichment on ranked proteins:

ranked_genes, gsea_results = ppin.run_gsea(
    node_attribute="degree_centrality",
    gene_sets="GO_Biological_Process_2023",
    outdir="gsea_results"
)

print(gsea_results.head(10))

Example: Complete Pipeline

See main.py for a complete example that ties all components together:

python main.py

This runs a full pipeline that:

  1. Queries STRING for a set of proteins

  2. Builds the PPI network

  3. Computes metrics

  4. Runs GSEA enrichment

  5. Exports Cytoscape-ready files

  6. Generates an interactive HTML report

Troubleshooting

Network timeout errors:

Increase the timeout parameter in StringLoader:

loader = StringLoader(..., timeout=60)
No identifiers found:

Ensure input protein names are correct. STRING accepts: - Gene names (e.g., “TP53”) - UniProt IDs - Ensembl IDs

Large networks are slow:

Reduce the number of additional nodes or filter by required_score:

loader = StringLoader(..., add_nodes=5, required_score=800)