Descriptor-free molecular discovery in large libraries by adaptive substituent reordering

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Optimizing molecular structures for desired properties is crucial in drug discovery.
• The vast number of possible substituent combinations on a molecular scaffold presents a significant challenge.
• Efficient exploration of chemical space is needed to identify novel active compounds.

Purpose of the Study:

• To introduce a novel strategy for efficiently optimizing substituent combinations on molecular scaffolds.
• To reduce the synthetic and assaying effort required in molecular discovery.
• To demonstrate the effectiveness of the strategy in identifying active compounds from large libraries.

Main Methods:

• Iterative rounds of compound sampling.
• Substituent reordering to regularize the property landscape.
• Property estimation over the generated landscape.

Main Results:

• The strategy successfully identified active compounds within a large pharmaceutical library.
• Achieved a threefold reduction in synthetic and assaying effort compared to conventional methods.
• Demonstrated efficacy even without prior knowledge of any specific compound's molecular identity.

Conclusions:

• The proposed strategy offers an efficient approach to navigating complex chemical spaces in molecular discovery.
• This method significantly enhances the efficiency of identifying promising drug candidates.
• The approach is applicable to large compound libraries and can accelerate the drug discovery pipeline.