Integrated Approach of Machine Learning and High-Throughput Screening to Identify Chemical Probe Candidates Targeting...

Area of Science:

• Biochemistry and Medicinal Chemistry
• Computational Drug Discovery
• Enzyme Inhibitor Development

Background:

• Developing selective chemical probes for the aldehyde dehydrogenase (ALDH) enzyme family is crucial but challenging for biological pathway dissection and drug discovery.
• Existing methods often lack efficiency and chemical diversity in identifying potent inhibitors for specific ALDH isoforms.

Purpose of the Study:

• To present a novel integrated approach combining quantitative high-throughput screening (qHTS), machine learning (ML), and pharmacophore (PH4) modeling for rapid identification of selective ALDH inhibitors.
• To enhance chemical diversity and identify potent, isoform-selective ALDH inhibitors using a combined in vitro and in silico strategy.

Main Methods:

• Screened approximately 13,000 compounds using qHTS against biochemical and cellular ALDH assays.
• Developed ML and PH4 models using screening data to virtually screen 174,000 additional compounds.
• Validated identified inhibitors through cellular target engagement assays and confirmed potency in biochemical and cell-based assays.

Main Results:

• Identified chemically diverse, potent, and isoform-selective inhibitors for ALDH1A2, ALDH1A3, ALDH2, and ALDH3A1.
• Successfully employed a single iteration of quantitative structure-activity relationship (QSAR) and PH4 modeling for efficient virtual screening.
• Generated a publicly available dataset of characterized compounds for future research and probe development.

Conclusions:

• The integrated qHTS, ML, and PH4 modeling approach significantly accelerates the discovery of biologically relevant chemical probes for the ALDH enzyme family.
• This platform expands accessible chemical diversity for probe development, offering a rapid and resource-efficient solution.
• The generated dataset serves as a valuable resource for training future models and advancing ALDH-targeted research.