Drug Properties Tools

Configuration File: drug_properties_tools.json Tool Type: Local Tools Count: 3

This page contains all tools defined in the drug_properties_tools.json configuration file.

Available Tools

DrugProps_calculate_qed (Type: DrugPropertiesTool)

Calculate Quantitative Estimate of Drug-likeness (QED) score for a compound using the Bickerton e…

DrugProps_calculate_qed tool specification

Tool Information:

  • Name: DrugProps_calculate_qed

  • Type: DrugPropertiesTool

  • Description: Calculate Quantitative Estimate of Drug-likeness (QED) score for a compound using the Bickerton et al. (2012) method. QED integrates 8 molecular descriptors (MW, AlogP, HBA, HBD, PSA, rotatable bonds, aromatic rings, structural alerts) into a single score from 0 (non-drug-like) to 1 (most drug-like). Median QED of approved oral drugs is ~0.49. High QED (≥0.67) = drug-like; Medium (0.34–0.67) = borderline; Low (<0.34) = non-drug-like. Use after DrugProps_lipinski_filter for a more nuanced drug-likeness assessment.

Parameters:

  • smiles (string) (required) SMILES string of the molecule. Examples: ‘CC(=O)Oc1ccccc1C(=O)O’ (aspirin, QED~0.55), ‘c1ccc2c(c1)cc1ccc3cccc4ccc2c1c34’ (pyrene, low QED).

Example Usage:

query = {
    "name": "DrugProps_calculate_qed",
    "arguments": {
        "smiles": "example_value"
    }
}
result = tu.run(query)

DrugProps_lipinski_filter (Type: DrugPropertiesTool)

Check drug-likeness of a compound using Lipinski Rule of Five (Ro5), Veber, Pfizer 3/75, Egan, an…

DrugProps_lipinski_filter tool specification

Tool Information:

  • Name: DrugProps_lipinski_filter

  • Type: DrugPropertiesTool

  • Description: Check drug-likeness of a compound using Lipinski Rule of Five (Ro5), Veber, Pfizer 3/75, Egan, and Ghose filters. Returns MW, cLogP, HBD, HBA, TPSA, RotBonds, Csp3 fraction, and pass/fail status for each filter. Essential first step in hit-to-lead optimization. Lipinski Ro5 (pass = drug-like): MW≤500, cLogP≤5, HBD≤5, HBA≤10 (max 1 violation). Veber (oral bioavailability): RotBonds≤10, TPSA≤140. Pfizer 3/75 (reduces promiscuity): cLogP≤3, TPSA≥75. Examples: aspirin (‘CC(=O)Oc1ccccc1C(=O)O’) passes all filters; venetoclax fails Ro5 due to MW>900 (beyond-Ro5 territory).

Parameters:

  • smiles (string) (required) SMILES string of the molecule. Examples: ‘CC(=O)Oc1ccccc1C(=O)O’ (aspirin), ‘CC(C)Cc1ccc(cc1)C(C)C(=O)O’ (ibuprofen), ‘c1ccc2c(c1)cc1ccc3cccc4ccc2c1c34’ (pyrene, non-drug-like).

Example Usage:

query = {
    "name": "DrugProps_lipinski_filter",
    "arguments": {
        "smiles": "example_value"
    }
}
result = tu.run(query)

DrugProps_pains_filter (Type: DrugPropertiesTool)

Screen a compound for PAINS (Pan-Assay Interference Compounds), Brenk undesirable substructures, …

DrugProps_pains_filter tool specification

Tool Information:

  • Name: DrugProps_pains_filter

  • Type: DrugPropertiesTool

  • Description: Screen a compound for PAINS (Pan-Assay Interference Compounds), Brenk undesirable substructures, and NIH alerts using RDKit FilterCatalog. PAINS compounds produce false positives in biochemical assays due to non-specific activity mechanisms (rhodanines, quinones, catechols, etc.). Brenk filter flags reactive/toxic groups (Michael acceptors, epoxides, nitroaromatics, etc.). Returns match details and remediation recommendations. Essential quality filter before committing to high-throughput screening hits. Use before investing in hit validation.

Parameters:

  • smiles (string) (required) SMILES string of the molecule to screen. Examples: ‘O=C1CSC(=S)N1’ (rhodanine PAINS), ‘O=C1C=CC(=O)c2ccccc21’ (quinone PAINS), ‘CC(=O)Oc1ccccc1C(=O)O’ (aspirin, clean).

Example Usage:

query = {
    "name": "DrugProps_pains_filter",
    "arguments": {
        "smiles": "example_value"
    }
}
result = tu.run(query)