Hierarchical QSAR technology based on the Simplex representation of molecular structure

Area of Science:

• Computational Chemistry
• Cheminformatics
• Drug Discovery

Background:

• Quantitative Structure-Activity Relationship (QSAR) and Quantitative Structure-Property Relationship (QSPR) are crucial for predicting molecular activity and properties.
• Existing QSAR methods often face challenges like molecular alignment and limited interpretability.
• There is a need for advanced QSAR methodologies that offer improved accuracy, interpretability, and applicability in drug design.

Purpose of the Study:

• To introduce and elaborate on the Hierarchical Quantitative Structure-Activity Relationship (HiT QSAR) technology based on the Simplex Representation of Molecular Structure (SiRMS).
• To demonstrate the application and advantages of HiT QSAR for various QSAR/QSPR tasks across different levels of molecular description (1D-4D).
• To validate the reliability of HiT QSAR models as predictive virtual screening tools and their utility in directed drug design.

Main Methods:

• Utilized the Simplex Representation of Molecular Structure (SiRMS) to represent molecules as systems of simplexes (tetratomic fragments).
• Employed a sequential, hierarchical approach to QSAR modeling, progressively increasing descriptor detailing from 1D to 4D.
• Developed and applied the HIT QSAR software, incorporating a statistical block and utilities for QSAR/QSPR analysis.

Main Results:

• HiT QSAR eliminates the need for molecular alignment and effectively incorporates various physical-chemical properties of atoms.
• The approach yielded models with high adequacy and good interpretability, outperforming popular QSAR methods on test sets.
• Validated QSAR models demonstrated reliability for virtual screening and served as a basis for successful directed drug design through synthetic and biological experiments.

Conclusions:

• HiT QSAR, based on SiRMS, provides a robust and versatile framework for QSAR/QSPR modeling.
• The technology offers significant advantages in terms of model accuracy, interpretability, and applicability to molecular design.
• HiT QSAR represents a valuable advancement in computational chemistry, facilitating efficient drug discovery and development.