Abstract:Alzheimer's disease (AD) is a gradually advancing neurodegenerative disorder characterized by a concealed onset. Acetylcholinesterase (AChE) is an efficient hydrolase that catalyzes the hydrolysis of acetylcholine (ACh), which regulates the concentration of ACh at synapses and then terminates ACh-mediated neurotransmission. There are inhibitors to inhibit the activity of AChE currently, but its side effects are inevitable. In various application fields where Al have gained prominence, neural network-based models for molecular design have recently emerged and demonstrate encouraging outcomes. However, in the conditional molecular generation task, most of the current generation models need additional optimization algorithms to generate molecules with intended properties which make molecular generation inefficient. Consequently, we introduce a cognitive-conditional molecular design model, termed PED, which leverages the variational auto-encoder. Its primary function is to adeptly produce a molecular library tailored for specific properties. From this library, we can then identify molecules that inhibit AChE activity without adverse effects. These molecules serve as lead compounds, hastening AD treatment and concurrently enhancing the AI's cognitive abilities. In this study, we aim to fine-tune a VAE model pre-trained on the ZINC database using active compounds of AChE collected from Binding DB. Different from other molecular generation models, the PED can simultaneously perform both property prediction and molecule generation, consequently, it can generate molecules with intended properties without additional optimization process. Experiments of evaluation show that proposed model performs better than other methods benchmarked on the same data sets. The results indicated that the model learns a good representation of potential chemical space, it can well generate molecules with intended properties. Extensive experiments on benchmark datasets confirmed PED's efficiency and efficacy. Furthermore, we also verified the binding ability of molecules to AChE through molecular docking. The results showed that our molecular generation system for AD shows excellent cognitive capacities, the molecules within the molecular library could bind well to AChE and inhibit its activity, thus preventing the hydrolysis of ACh.

PED: a novel predictor-encoder-decoder model for Alzheimer drug molecular generation

Multi-site Dynamic Recording for Aβ Oligomers-Induced Alzheimer's Disease in Vitro Based on Neuronal Network Chip.

[In Vitro Pathological Model of Alzheimer's Disease Based on Neuronal Network Chip and Its Real-Time Dynamic Analysis].

MTDL-GAN: De novo Design of Multi-Target Directed Ligands for Alzheimer’s Disease from Unpaired Sets of Target-Focused Chemical Library

Design of vilazodone-donepezil chimeric derivatives as acetylcholinesterase inhibitors by QSAR, molecular docking and molecular dynamics simulations

Molecular generation strategy and optimization based on A2C reinforcement learning in de novo drug design

AlzhCPI: A Knowledge Base for Predicting Chemical-Protein Interactions Towards Alzheimer’s Disease

In Silico Discovery and Predictive Modeling of Novel Acetylcholinesterase (AChE) Inhibitors for Alzheimer's Treatment

An AI-Driven Framework for Discovery of BACE1 Inhibitors for Alzheimer's Disease

Screening Targets and Therapeutic Drugs for Alzheimer's Disease Based on Deep Learning Model and Molecular Docking

Hybrid neural network approaches to predict drug-target binding affinity for drug repurposing: screening for potential leads for Alzheimer's disease

Discovery of Multitarget-Directed Ligands Against Alzheimer'S Disease Through Systematic Prediction of Chemical Protein Interactions

Discovery of novel acetylcholinesterase inhibitors through integration of machine learning with genetic algorithm based in silico screening approaches

Structural and molecular features of acetylcholinesterase targeting small molecules: Leveraging in-silico tools to combat Alzheimer's disease

Discovery of Novel Acetylcholinesterase Inhibitors as Potential Candidates for the Treatment of Alzheimer's Disease

Computational analysis of novel drugs designed for use as acetylcholinesterase inhibitors and histamine H3 receptor antagonists for Alzheimer's disease by docking, scoring and de novo evolution

Virtual screening of lead compounds for the treatment of Alzheimer's disease based on multi-target strategy

CKG-IMC: An inductive matrix completion method enhanced by CKG and GNN for Alzheimer's disease compound-protein interactions prediction

Computational and ADMET Predictions of Novel Compounds as Dual Inhibitors of BuChE and GSK-3β to Combat Alzheimer's Disease

Robust and predictive 3D-QSAR models for predicting the activities of novel oxadiazole derivatives as multifunctional anti-Alzheimer agents

Investigation of Anti-Alzheimer’s Mechanisms of Sarsasapogenin Derivatives by Network-Based Combining Structure-Based Methods