BAD Molecule Filter

A free machine learning tool by Abdallah Abou Hajal that predicts whether a small molecule will clump into a colloidal aggregate and show up as a false positive in a drug screen.

What it checks

  1. Colloidal aggregation

    Give it a molecule as a SMILES string. It predicts whether the molecule is likely to form colloidal aggregates, which inhibit proteins nonspecifically and produce false hits in high-throughput screening. It was Abdallah's MSc work at Al Ain University.

  2. What it does not check

    Aggregation only. It does not flag PAINS substructures or other kinds of assay interference, such as reactive or redox cycling compounds.

Run it

  1. Desktop app for Windows and Linux

    Version 1.1, free, from the releases page on GitHub.

  2. Source code

    Open under the GPL 3.0 licence.

Cite it

  1. Boosting the Accuracy and Chemical Space Coverage of the Detection of Small Colloidal Aggregating Molecules Using the BAD Molecule Filter

    Abou Hajal A, Bryce RA, Ben Amor B, Atatreh N, Ghattas MA. Journal of Chemical Information and Modeling 2024, 64(13), 4991-5005.

    doi:10.1021/acs.jcim.4c00363 · free accepted manuscript · PubMed 38920403

Cited by

  1. Exploiting colloidal drug aggregation for drug delivery: From promise to prediction using computational tools

    Slaughter, Li and Shoichet · Advanced Drug Delivery Reviews · 2026

  2. Mitigating Molecular Aggregation in Drug Discovery With Predictive Insights From Explainable AI

    Sturm and colleagues · Angewandte Chemie · 2025

  3. From Algorithms to Assets: A Comprehensive Review of AI's Role in Preclinical Drug Discovery and the Hurdles to Clinical Translation

    Cai and Liu · Pharmaceuticals · 2026

For questions about the tool, email Abdallah.