Classification scheme for the design of serine protease targeted compound libraries

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Serine proteases (SPs) are crucial drug targets.
• Accurate classification of SP-active molecules is essential for drug development.
• Existing methods for classifying SP activity can be time-consuming.

Purpose of the Study:

• To develop and validate a novel scoring scheme for classifying molecules as serine protease (SP) active or inactive.
• To create an efficient computational tool for identifying potential SP ligands.
• To accelerate the drug discovery process for SP-targeted therapies.

Main Methods:

• Utilized pre-selected molecular descriptors for structural encoding.
• Employed a trained neural network for molecular classification.
• Validated the method using large, diverse databases of SP-active and non-SP-active molecules and Sensitivity Analysis.

Main Results:

• Successfully developed a scoring scheme for classifying molecules into SP-active and inactive categories.
• Demonstrated efficient qualification and disqualification of potential serine protease ligands.
• The method proved effective in profiling molecular requirements for SP activity.

Conclusions:

• The developed scoring scheme provides an efficient computational tool for drug discovery.
• This method aids in constraining virtual libraries, accelerating the identification of novel SP-active drug candidates.
• The approach facilitates the development of new drugs targeting serine proteases.