A lazy learning-based QSAR classification study for screening potential histone deacetylase 8 (HDAC8) inhibitors

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Histone deacetylases 8 (HDAC8) regulates gene transcription, including tumor suppressor genes, making it a key target in cancer therapy.
• Developing effective HDAC8 inhibitors is crucial for advancing human cancer treatment strategies.

Purpose of the Study:

• To develop and validate quantitative structure-activity relationship (QSAR) classification models for identifying potential HDAC8 inhibitors.
• To employ virtual screening and molecular docking to discover novel HDAC8 inhibitor candidates.
• To assess the stability of potential inhibitor-HDAC8 complexes using molecular dynamics simulations.

Main Methods:

• Development of two QSAR classification models: K-nearest neighbours (KNN) and Neighbourhood Classifier (NEC).
• Calculation of molecular descriptors using ADRIANA.Code, followed by Principal Component Analysis (PCA) for descriptor selection.
• Validation of models using Leave-One-Out Cross-Validation (LOO CV) and an external test set.
• Virtual screening of a compound database, followed by molecular docking and 5 ns molecular dynamics (MD) simulations for hit validation.

Main Results:

• Two predictive QSAR models (KNN and NEC) were successfully developed and validated.
• Virtual screening identified five potential HDAC8 inhibitor compounds based on scoring functions and binding affinity.
• Molecular dynamics simulations confirmed the stability of the identified hit compounds in complex with HDAC8.

Conclusions:

• The NEC classification model represents a novel application in virtual screening for drug discovery.
• The identified compounds show potential as HDAC8 inhibitors for cancer treatment.
• This study demonstrates the utility of QSAR modeling and computational approaches in accelerating the drug discovery process for HDAC8 targets.