Graph Convolutional Neural Network-Enabled Frontier Molecular Orbital Prediction: A Case Study with Neurotransmitters...

Area of Science:

• Computational Chemistry
• Neuroscience
• Artificial Intelligence

Background:

• Predicting molecular properties is crucial for drug discovery.
• Understanding neurochemical interactions with receptors is key to treating neurological disorders.
• Artificial intelligence (AI) offers novel approaches to complex chemical and biological problems.

Purpose of the Study:

• To investigate the relationship between neurochemical hardness and receptor binding affinities using AI.
• To apply a graph convolutional neural network artificial neural network (GCN-ANN) for predicting molecular properties.
• To explore the applicability of the Hard-Soft Acid-Base (HSAB) principle in neurochemical interactions.

Main Methods:

• Developed and trained a GCN-ANN model using B3LYP-calculated HOMO and LUMO energies (>110,000 molecules).
• Performed benchmark studies on 45 neurochemicals using B3LYP, ωB97XD, and M06-2X density functionals.
• Analyzed binding affinities, hardness, and GCN-ANN-derived substructures.

Main Results:

• The GCN-ANN model successfully probed the link between chemical hardness and receptor affinity.
• Consistent hardness and electronegativity values were observed across multiple density functionals.
• The study confirmed that neuroreceptor interactions align with the Hard-Soft Acid-Base (HSAB) principle.

Conclusions:

• AI-driven methods, specifically GCN-ANN, provide valuable physical insights into neurochemical-neuroreceptor interactions.
• Findings support the HSAB principle in governing these biological interactions.
• This research paves the way for developing more precise and effective antidepressants for anxiety and depression.