Cheminformatics

Making AI Work in Discovery Chemistry: Precision, Trust, and Practical Value

Key takeaways from the Optibrium panel discussion featuring Nathan Brown, Charlotte Deane, Pat Walters, Chris Swain, and Paul Czodrowski on AI strategy, data leakage, generative risks, and LLMs in drug discovery.

Jul 30, 2026

Protein Descriptors for Machine Learning in Drug Discovery

A practical map of protein and protein-ligand descriptors for machine learning modeling, from sequence-only features to MD-derived structural and interaction features.

Jul 20, 2026

Reproducible ≠ Robust: Why One UMAP Seed Isn't Enough to Trust a Split

Pinning random_state=42 makes a UMAP-based train/test split perfectly reproducible — and that is exactly why it can lull you into a false sense of rigor. Reproducibility guarantees you get the same answer every run; it says nothing about whether that answer is typical. Here's the distinction, why it matters for evaluating GNNs on chemical-domain shifts, and how to fix it.

Jun 23, 2026

Choosing the Right Partial Charges for Molecular Docking: AM1-BCC, PM6 and Beyond

Partial charges quietly drive the electrostatics behind every docking score. This guide compares the common semi-empirical options — AM1-BCC, PM6, AM1/PM3, Gasteiger and RESP — and gives a practical recommendation for when to use each.

Jun 23, 2026

Building a 3D Pharmacophore Model from PDB Data: A Free Python Workflow

A step-by-step, fully open-source pipeline that turns raw Protein Data Bank structures into a ligand-based 3D pharmacophore — mining the PDB, aligning binding pockets, clustering ligands, fixing bond orders, and distilling a consensus feature map ready for virtual screening.

Jun 23, 2026

Beyond Static Models: Agentic AI and Multi-Agent Systems in Drug Discovery

AI in drug discovery is moving past one-shot predictions toward autonomous agents that plan experiments, write their own analysis code, call docking and QM tools, and critique their results. Here's what the shift means, the design patterns behind it, and five open-source agents worth examining — with an honest look at the caveats.

Jun 23, 2026

Does data leakage really inflate binding-affinity GNNs? A laptop-scale reproduction

I tried to reproduce the well-known PDBbind data-leakage effect with a small 3D GNN on a laptop — and couldn't. Two independent diagnostics show why: leakage only inflates models strong enough to memorize.

Jun 19, 2026

A History of Graph Neural Networks in Drug Discovery

Graphs in drug discovery have gone from a quiet background tool to one of the main ways we think about molecules, proteins, and their interactions. This post walks through that story: how the field moved from fingerprints and QSAR to today’s 3D, attention-based graph neural networks operating directly on protein-ligand complexes.

Jun 8, 2026

A Practical Guide to QSAR Model Validation: Internal, Cross, and External Checks

Building a QSAR model is only half the job. The harder question is: does it actually work? Overfitted models routinely pass internal checks while failing completely on new compounds. The OECD principles and decades of best-practice literature have converged on a three-tier validation framework that separates what a model has memorised from what it can genuinely predict.

Jun 1, 2026

Beyond SMILES: The Evolving Landscape of Molecular Representations

This post summarizes the key ideas from Zhang et al. (2026), “Molecular Knowledge Representations in the Era of Artificial Intelligence,” a preprint published on ChemRxiv (DOI: 10.26434/chemrxiv.15002830/v1). The Core Problem Molecules are quantum-mechanical objects. Their exact description is computationally intractable, and any real sample is a messy mixture of impurities, conformers, and side products. This means every representation of a molecule is, by necessity, an approximation — shaped by the interactions and length scales we care about.

May 23, 2026