Biospectra analysis: model proteome characterizations for linking molecular structure and biological response.

Area of Science:

• Medicinal Chemistry
• Computational Biology
• Pharmacology

Background:

• Establishing quantitative structure-activity relationships (SAR) is crucial for drug discovery.
• Understanding how molecular structure relates to broad biological effects, including protein modulation, is essential.
• Measuring a molecule's functional activity across diverse proteins presents a significant challenge.

Purpose of the Study:

• To introduce biospectra analysis, a probabilistic SAR approach for identifying drug effect profiles.
• To demonstrate the utility of in vitro binding data for assessing functional similarity between medicinal agents.
• To investigate the relationship between biospectra similarity and agonist/antagonist properties of dopamine-aligned molecules.

Main Methods:

• Utilized in vitro binding data (percent inhibition) across a proteome-wide assay panel.
• Applied hierarchical clustering to a large dataset (1567 molecules) to identify relevant molecular clusters.
• Calculated biospectra similarity between molecules to determine functional effect profile similarity.

Main Results:

• Biospectra analysis effectively identified agonist and antagonist effect profiles based on pattern similarity.
• Functional effect profile similarity was maintained even after removing known drug targets from the analysis.
• Demonstrated a clear association between biospectra similarity and biological response profile similarity for 24 molecules.

Conclusions:

• Biospectra analysis offers a novel, unbiased method for predicting structure-response relationships.
• This approach facilitates the translation of broad biological effect information into actionable chemical structure design.
• Provides a valuable tool for drug discovery by forecasting molecular functional profiles.