ProteomeScan

ProteomeScan is a Python toolkit for large-scale human proteome scan screening against a given set of drugs/ligands.

Requirements

• Python 3.10 and above
• Create a conda environment and install dependencies from requirements.txt
• Internet connection (for UniProt/PDBe/PDBe-KB access)
• Your ligand files (see /data folder below)

Follow the documentation links to setup deepchem-server and PyDS.

Data Folder

• data/gene_selection_raw: Contains files used to filter Human Proteome and shortlist gene names for ProteomeScan.
• data/ligands/processed/: Place your ligand SDF files here before running the pipeline. Each ligand should be provided as its own file named as Processed_<LigandName>.sdf.
• If you have raw Ligand SMILES, use data/ligands/prepare_ligands.py.

Quick Start

To run the full pipeline, simply execute:

python example.py

example.py is provided as an executable, end-to-end pipeline with example gene and ligand lists. You can customize these lists or provide your own, and place the corresponding ligand .sdf files in data/ligands/processed/.

What happens:

• For each gene, the tool fetches optimal (curated, best) PDBs by scraping UniProt, PDBe, and PDBe-KB.
• Each ligand is docked against optimal PDBs for each gene target.
• All results and logs are saved in the provided output directory.
• Post-scan run includes promiscuity filtering and pose analysis.

Pseudo-Code

For workflow logic at a high level, see the pseudo-code below:

Require: ligands list, gene names list

N_ligands ← length(ligands list)
N_gene_names ← length(gene names list)
Optimal PDBs ← ∅
for each gene name in gene names list do
    Optimal PDBs ← get_optimal_cleaned pdbs(gene name)
end for
Scores ← ∅
for each ligand in ligands list do
    for each gene name in gene names list do
        Scores[ligand][gene name] ← run_gene_based_docking(ligand, gene name)
    end for
end for
Ranked_Scores ← ∅
for each ligand in ligands list do
    Ranked_Scores[ligand] ← Sort(Scores[ligand])
end for
Promiscuous targets ← Analyse_Target_Promiscuity(N_ligands, N_gene_names, Ranked_Scores)
Updated_Promiscuous_targets ← [ ]
for each target in Promiscuous targets do
    if Check_binding(target, ligands list) then
        Append target to Updated_Promiscuous_targets
    end if
end for
for each ligand in ligands list do
    Remove Updated_Promiscuous_targets from Ranked_Scores[ligand]
end for
return Ranked_Scores per ligand