Flexible Fitting of PROTAC Concentration-Response Curves with Changepoint Gaussian Processes

Area of Science:

• Biochemistry
• Pharmacology
• Computational Biology

Background:

• Proteolysis-targeting chimeras (PROTACs) are novel therapeutics that induce targeted protein degradation.
• PROTAC compound screening relies on concentration-response (CR) assays to determine potency, often using the DC50 parameter.
• Traditional analysis of PROTAC CR data is hindered by complex, non-sigmoidal curve shapes, necessitating manual data masking.

Purpose of the Study:

• To develop an automated and reproducible method for analyzing PROTAC concentration-response data.
• To overcome the limitations of manual data masking in PROTAC screening.
• To accurately estimate key PROTAC potency parameters and rank compounds.

Main Methods:

• Implemented a Bayesian changepoint Gaussian processes model for flexible fitting of CR curves.
• The model accommodates both sigmoidal and non-sigmoidal PROTAC data without user intervention.
• Estimated parameters including DC50, Dmax, and Point of Departure (PoD) and propagated uncertainty.

Main Results:

• The Bayesian model successfully fitted complex, non-sigmoidal PROTAC CR curves.
• Automated parameter estimation and compound ranking were achieved, reducing subjectivity.
• Uncertainty propagation enabled confident comparison and ranking of PROTAC compounds.

Conclusions:

• The developed Bayesian approach offers a flexible, automated, and reproducible method for PROTAC screening.
• This minimizes manual decisions, reducing time and cost in compound evaluation.
• The method enhances the reliability of PROTAC compound potency assessment and ranking.