Design of focused and restrained subsets from extremely large virtual libraries

Area of Science:

• Computational Chemistry
• Cheminformatics
• Drug Discovery

Background:

• Combinatorial chemistry enables the creation of virtual libraries with billions of compounds.
• Selecting practical subsets for experimentation is crucial but challenging for large virtual libraries.
• Traditional computational library design often involves impractical full virtual library enumeration.

Purpose of the Study:

• To introduce and evaluate a novel computational approach for efficient combinatorial library design.
• To demonstrate the advantages of calculating molecular descriptors dynamically within an optimization cycle.

Main Methods:

• Developed and implemented an 'on the fly optimization' (OTFO) method.
• Integrated an ultra-fast descriptor calculation engine with subset optimization capabilities.
• Compared OTFO performance against simulated annealing (SA) and genetic algorithms (GA).

Main Results:

• OTFO successfully designs focused and restrained chemical library subsets by sampling only a fraction of the virtual library.
• The method avoids the impracticality and wastefulness of enumerating properties for entire virtual libraries.
• Coupling rapid descriptor calculations with optimization significantly enhances efficiency.

Conclusions:

• On-the-fly optimization (OTFO) provides an efficient and practical strategy for combinatorial library design.
• This approach is stable and effective for generating targeted compound subsets from massive virtual libraries.
• OTFO offers a significant improvement over traditional methods requiring full library enumeration.