liveBook · Manning

Structural alert

As explained in Build AI Drug Discovery Pipelines

A structural alert is a chemical substructure associated with undesirable reactivity, toxicity, instability, or other drug-development risk.

A structural alert is a molecular substructure associated with undesirable reactivity, toxicity, instability, or other drug-development risk. Matching alerts provides a fast, interpretable way to triage screening libraries before experimental assays. section 2.3.2

Common examples include thiophene, epoxides, and acid anhydrides, which may indicate risks such as hepatotoxicity, mutagenicity, or skin sensitization. PAINS and BRENK catalogs extend filtering to assay interference and unwanted functionality. section 2.3.2

Why an alert is not a verdict

The same thiophene alert appears in methapyrilene and eprosartan, but only methapyrilene is associated with a hepatotoxic metabolic pathway. Structural alerts can therefore overestimate hazard, miss toxic molecules without known alerts, or require biochemical context for interpretation. section 2.3.2

Listing 2.5 Applying PAINS & BRENK Filters
from rdkit.Chem import FilterCatalog
def apply_pains_brenk_filters(df, mol_col='mol'):
    df_copy = df.copy()
    filter_params = FilterCatalog.FilterCatalogParams()    #A
    filter_params.AddCatalog(filter_params.FilterCatalogs.PAINS)
    filter_params.AddCatalog(filter_params.FilterCatalogs.BRENK)
    catalog = FilterCatalog.FilterCatalog(filter_params)
    
    df_copy.loc[:, 'PAINS_BRENK_match'] = \
        df_copy[mol_col].apply(catalog.HasMatch)    #B
    df_copy.loc[:, 'PAINS_BRENK_compliant'] = ~df_copy['PAINS_BRENK_match']
    
    filtered_df = df_copy[df_copy['PAINS_BRENK_compliant']]
    
    print(f"Compounds before PAINS/BRENK filter: {len(df_copy)}")    #C
    return df_copy, filtered_df
 
specs_ro5_compliant, specs_ro5_pains_brenk_compliant = \
    apply_pains_brenk_filters(specs_ro5_compliant)

An RDKit workflow combines PAINS and BRENK catalogs, records whether each molecule matches, and retains only compliant compounds for subsequent screening. section 2.3.2

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Defines alerts, gives examples, demonstrates RDKit filtering, and stops short of treating matches as predictive toxicity judgments.
Examples
Uses thiophene, epoxides, acid anhydrides, Glaxo alerts, and PAINS/BRENK screening results.
What you can do after
Apply via RDKit to flag and remove matching molecules while retaining match records for review.
How it connects
Connects structural alerts to PAINS, BRENK filters, SMARTS, substructure matching, toxicity models, and biochemical context.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.