Beyond profiling: using ADMET models to guide decisions.

Area of Science:

• Drug Discovery
• Computational Chemistry
• Medicinal Chemistry

Background:

• ADMET models (in silico and in vitro) are standard for molecule profiling and identifying liabilities.
• Current applications are often limited to basic filtering, missing opportunities for strategic decision-making.

Purpose of the Study:

• To demonstrate how ADMET models can be utilized beyond basic profiling to guide critical drug discovery decisions.
• To illustrate methods for using model data and uncertainty analysis to confidently select chemical series and design improved molecules.

Main Methods:

• Defining project-specific success criteria and property importance for prioritization.
• Utilizing experimental or predicted ADMET model data to assess molecule property balance.
• Incorporating uncertainty analysis into decision-making processes.
• Interpreting in silico models to understand structure-property relationships and sensitivity to modifications.

Main Results:

• Models can guide the selection of promising chemical series for focused resource allocation.
• Interpretation of model data aids in designing improved molecules by targeting specific structural modifications.
• Visualizing structure-property relationships facilitates the redesign process.

Conclusions:

• ADMET models offer significant value beyond simple profiling, enabling confident decision-making in drug discovery.
• Integrating uncertainty assessment and detailed model interpretation enhances the strategic application of these models.
• The described methods empower researchers to drive drug discovery decisions effectively.