Discovery of senolytics using machine learning

Area of Science:

• Biochemistry
• Computational Biology
• Pharmacology

Background:

• Cellular senescence is a key factor in aging and various diseases, including cancer and diabetes.
• Targeting senescent cells with senolytics is a promising therapeutic strategy.
• A limited number of senolytics are known due to the lack of identified molecular targets.

Purpose of the Study:

• To discover novel senolytic compounds using machine learning algorithms.
• To validate the senolytic activity of computationally identified compounds in human cell lines.
• To assess the cost-effectiveness and potential of AI in early-stage drug discovery.

Main Methods:

• Machine learning algorithms were trained on published data to screen chemical libraries.
• Computational screening identified potential senolytic compounds.
• In vitro validation of ginkgetin, periplocin, and oleandrin in human cell lines undergoing senescence.

Main Results:

• Ginkgetin, periplocin, and oleandrin were identified and validated as senolytics.
• These compounds demonstrated potency comparable to existing senolytics.
• Oleandrin showed improved potency compared to current best-in-class alternatives.
• The AI-driven approach significantly reduced drug screening costs.

Conclusions:

• Artificial intelligence can effectively leverage diverse drug screening data for novel drug discovery.
• The identified senolytics offer new therapeutic avenues for age-related diseases.
• This study highlights the potential of AI and open science in accelerating early-stage drug discovery.