A unified platform for ADMET profiling, scoring, and visualization. GAI integrates data from multiple predictive tools into a single, interpretable index (0-100) to accelerate drug discovery workflows.
Install R (>= 4.0) and the required dependencies:
install.packages("remotes")
remotes::install_deps()Launch the interactive Shiny App:
https://sulkysubject37.shinyapps.io/GAI-Analyzer/
Process a file from your terminal:
./gai data/ADMETlab3_result64.csv -o results/mol_64- Multi-Tool Support: Natively parses results from ADMETlab 3.0, SwissADME, pkCSM, and DeepPK.
- Robust Normalization: Automatically handles unit conversions and "desirability" directions (e.g., Higher is Better for HIA, Lower is Better for Toxicity).
- Batch Processing: Score hundreds of molecules simultaneously and generate comparative boxplots.
- Rich Visuals: Detailed molecule reports including Gauge plots, Radar charts, and Waterfall contribution charts.
Detailed documentation is available in the docs/ directory:
| Document | Description |
|---|---|
| Mathematical Model | GAI aggregation formulas, weights, and normalization logic. |
| Architecture | File structure, data flow pipeline, and "Unflawed" logic. |
| Supported Endpoints | Comprehensive list of canonical parameters and their logic. |
| Usage Guide | In-depth instructions for the Shiny App and CLI options. |
app.R: Shiny web application entry point.scripts/run_gai.R: Core CLI processing script.R/: Scoring engine, data loaders, and master configuration.data/: Example ADMET results for testing.www/: Application styling and assets.
If you use this toolkit in your research, please cite:
Arshad, M. (2025). Global ADMET Index (GAI): v2.0.1 (v2.0.1). Zenodo. https://doi.org/10.5281/zenodo.18030940
This project is licensed under the MIT License - see the LICENSE file for details.
Author: MD. Arshad
Institution: Jamia Millia Islamia