Mycobacterium abscessus drug discovery using machine learning.

Area of Science:

• Microbiology
• Drug Discovery
• Computational Chemistry

Background:

• Nontuberculous mycobacteria (NTM) infections are increasing globally, exceeding tuberculosis (TB) in some regions.
• NTM pose significant challenges for patients with underlying lung conditions like COPD, bronchiectasis, and cystic fibrosis.
• Current treatments for NTM infections are often ineffective, lengthy, toxic, and associated with high recurrence rates, particularly for Mycobacterium abscessus.

Purpose of the Study:

• To identify novel therapeutic agents for Mycobacterium abscessus infections.
• To explore the application of machine learning in drug discovery for NTM.
• To address limitations in current machine learning approaches for M. abscessus drug discovery.

Main Methods:

• Integrated machine learning, medicinal chemistry, and in vitro testing.
• Applied a previously established drug discovery approach to M. abscessus.
• Synthesized and tested novel 1-(phenylsulfonyl)-1H-benzimidazol-2-amine compounds.

Main Results:

• Identified several novel 1-(phenylsulfonyl)-1H-benzimidazol-2-amine compounds.
• Observed weak in vitro activity of these compounds against M. abscessus.
• Highlighted limitations and areas for improvement in the machine learning approach for M. abscessus.

Conclusions:

• The identified compounds represent a potential starting point for medicinal chemistry optimization.
• Further research is needed to develop effective treatments for M. abscessus infections.
• Machine learning holds promise for NTM drug discovery but requires further refinement for M. abscessus.