A computational method for detection of ligand-binding proteins from dose range thermal proteome profiles

Area of Science:

• Molecular Biology
• Chemical Biology
• Drug Discovery

Background:

• Detecting ligand-protein interactions in living cells is crucial for molecular biology and drug discovery.
• Current proteome-wide thermal stability profiling methods have limitations in sensitivity, specificity, and cross-dataset comparability due to preset thresholds.
• Improved analytical strategies are needed to accurately interpret thermal profiling data.

Purpose of the Study:

• To develop a novel data analysis method for thermal stability profiling that improves the detection of ligand-protein interactions.
• To provide better control over the false discovery rate in thermal profiling experiments.
• To enable more reliable identification of drug targets and off-target interactions.

Main Methods:

• Development of a statistical hypothesis testing framework applied to curves derived from thermal stability profiling data.
• Implementation of the method to analyze datasets from epigenetic drugs and a metabolite.
• Utilizing an R package available from Bioconductor for data analysis.

Main Results:

• The new method demonstrates enhanced sensitivity and specificity in detecting ligand-protein interactions compared to traditional threshold-based approaches.
• Application of the method identified off-target engagement of epigenetic drugs.
• Specifically, the HDAC8 inhibitor PCI-34051 and its analog BRD-3811 were found to bind and inhibit leucine aminopeptidase 3.

Conclusions:

• The proposed statistical method offers a robust approach for analyzing thermal stability profiling data.
• This method facilitates accurate identification and prioritization of drug targets by controlling the false discovery rate.
• The findings highlight the potential for discovering unexpected drug-target interactions, such as the inhibition of leucine aminopeptidase 3 by HDAC8 inhibitors.