Next-Gen Bioinformatics

Predicting Toxicity with Ensemble Machine Learning

Stack-Tox bridges the gap between high-performance computing and drug discovery. We provide instant toxicity profiling using advanced stacking algorithms and standard druglikeness filters.

The Pipeline

1. Input & Parsing

Users submit SMILES via draw, paste, or upload. The system standardizes molecules using RDKit and generates 2D/3D descriptors.

2. Stacked Inference

Data is sent securely to our HPC cluster. A meta-model aggregates predictions from Random Forest, SVM, and XGBoost classifiers.

3. Profiling Report

Results are returned instantly with detailed violation counts (Lipinski, Ghose, etc.) and a BOILED-Egg plot for BBB permeation.

Built With Precision

Python & RDKit

Cheminformatics Core

PHP 8.2

Server Orchestration

Scikit-Learn

ML Model Training

Linux HPC

High Performance Computing

Research Team

Team Member

Dheeraj Kumar Chaurasia

PhD Scholar

Specializing in Computational Drug Discovery and ML Pipelines.

Lab Logo

SCFBio Lab

Research Group

Supercomputing Facility for Bioinformatics & Computational Biology.