The Challenge
The Problem with Smoking Cessation
Tobacco consumption is the single largest avoidable health risk in the EU. Around 700,000 deaths occur every year, with 26% of the overall population classified as smokers. Nicotine Replacement Therapy (NRT) is the primary approach for smoking cessation, yet it exhibits low success and adherence rates due to individual variations in nicotine metabolism.
The enzyme CYP2A6 plays a key role in nicotine metabolism. Inhibiting this enzyme can slow nicotine breakdown, enhancing the effectiveness of NRT protocols. NovaMechanics developed a strategy focusing on natural products as potential CYP2A6 inhibitors.
Our Methodology
A multi-stage computational pipeline validated through experimental testing
Database Curation
Curated three major natural product databases: Analyticon (6,500 compounds), EthnoHERBS (23,000 compounds), and COCONUT (682,114 compounds). Isolated unique compounds and filtered by molecular weight (<300 Da).
Virtual Screening with Enalos Asclepios
Used the Enalos Asclepios KNIME nodes to set up structure-based virtual screening workflows. Performed molecular docking simulations against the CYP2A6 crystal structure (PDB ID: 4EJJ) using AutoDock Vina. Zero-code, reproducible, and automated.
Machine Learning with Isalos
Built predictive ML models using the Isalos Analytics Platform. Used 77 molecular descriptors, 80/20 training/test split (3,600/790 compounds). Validated with Random Forest and XGBoost — XGBoost selected as the final model for CYP2A6 inhibition prediction.
In Vitro Evaluation
Top 30 compounds from Analyticon tested in vitro for CYP2A6 inhibitory activity using mouse microsomes. Multiple rounds of optimization identified 5 potent inhibitors with IC50 values close to the gold standard 8-methoxypsoralen (MOPS).
Toxicity Assessment
Best compounds (MO9, MO22) tested in HepG2 human hepatocyte cells using MTT assay. No toxicity observed at 24 or 48 hours at concentrations up to 100 µM.
In Vivo Validation
Compounds tested in mice for inhibition of nicotine metabolism. Both MO9 and MO22 exhibited significant, concentration-dependent inhibition of nicotine metabolism in vivo. Results confirmed against human recombinant CYP2A6 enzyme.
Results at a Glance
Powered by NovaMechanics Software
Enalos Asclepios KNIME Nodes
Used for structure-based virtual screening workflows including molecular docking with AutoDock Vina. Zero-code, reproducible, and fully automated pipeline for hit discovery.
Learn moreIsalos Analytics Platform
Used for building the XGBoost machine learning model for CYP2A6 inhibition prediction with 77 molecular descriptors, featuring comprehensive model validation and interpretability.
Learn moreNatural products present a cost-efficient and rich source of potential enzyme inhibitors. This project demonstrates that NovaMechanics' computational pipeline — from virtual screening through machine learning to experimental validation — can successfully identify novel drug candidates with real therapeutic potential.
Have a Similar Challenge?
Let's explore how our computational drug discovery pipeline can accelerate your research.
Get in Touch