Insight into potent leads for alzheimer's disease by using several artificial intelligence algorithms

Area of Science:

• Pharmacology
• Computational Biology
• Traditional Chinese Medicine

Background:

• Alzheimer's disease (AD) pathogenesis involves proteins like S-nitrosoglutathione reductase (GSNOR), complement Factor D, complement 3b (C3b), and Protein Kinase R-like Endoplasmic Reticulum Kinase (PERK).
• These proteins represent potential therapeutic targets for AD treatment.

Purpose of the Study:

• To identify Traditional Chinese Medicine (TCM) candidates with multi-target potential against GSNOR, C3b, Factor D, and PERK using network pharmacology.
• To validate the efficacy and binding stability of identified TCM candidates using Artificial Intelligence (AI) and molecular dynamics (MD) simulations.

Main Methods:

• Network pharmacology approach to screen TCM candidates against GSNOR, C3b, Factor D, and PERK.
• AI algorithms (machine learning and deep learning) and Comparative Molecular Field Analysis (CoMFA/CoMSIA) for bioactivity prediction and model construction.
• 100 ns molecular dynamics (MD) simulations to evaluate binding stability of potential inhibitors.

Main Results:

• AI models demonstrated high predictive power with R² values generally above 0.75.
• Identified potential TCM candidates including Ardisia japonica, Ligusticum chuanxiong, Lippia nodiflora, and Mirabilis jalapa.
• Specific compounds like Glyasperin B and Nodifloridin A showed promising binding stability.

Conclusions:

• The study proposes a novel combined AI method for predicting drug-target interactions.
• Identified herbs and compounds may form a potential multi-target therapeutic formula for Alzheimer's disease.
• This research provides a computational basis for developing novel AD treatments from TCM.