AIRCHECK: An open, FAIR platform for AI protein–ligand discovery

MLG Seminar by Dr. Benjamin Haibe-Kains

Abstract

Recent breakthroughs in artificial intelligence (AI) have opened new avenues to accelerate drug discovery. However, progress is fundamentally limited by access to large-scale, high-quality experimental chemical screening data. To address this challenge, the Structural Genomics Consortium is developing AIRCHECK, an open platform that transforms large-scale protein–ligand experiments into reusable, AI‑ready resources.

Our mission is to generate standardized, experimentally curated binding data across thousands of proteins, release these data rapidly under FAIR and open science principles, and use them to accelerate and democratize early-stage drug discovery. To achieve this, we are building a central data hub that supports data ingestion, provenance tracking, and rigorous quality control, along with automated and versioned dataset generation. Broad access is provided through web-based tools, application programming interfaces, and rich, searchable metadata, enabling the creation of high-quality AI-ready datasets and helping to fully realize the potential of AI in early drug discovery.

These data support the development of advanced AI models that fuse chemical structure, protein sequence, and three-dimensional information, as well as generative design methods. Given the complexity of these models, AIRCHECK places strong emphasis on benchmarking and experimental validation. This is achieved through internal comparisons against baseline models and through open community challenges organized with the MAINFRAME network. Together, this work aims to lower the barrier to entry for data-driven drug discovery, improve the reliability of AI predictions, and enable faster, more systematic exploration of the human proteome.

Speaker Biography

Dr. Benjamin Haibe-Kains is the Executive AI Scientific Director at the University Health Network, a Senior Scientist at the Princess Margaret Cancer Centre, University Health Network, and a Professor in the Department of Medical Biophysics at the University of Toronto (Canada). He earned his PhD in Bioinformatics from the Université Libre de Bruxelles (Belgium) and pursued postdoctoral training at the Dana-Farber Cancer Institute and the Harvard School of Public Health (USA) as a Fulbright Scholar.

Dr. Haibe-Kains holds the Canada Research Chair in Computational Pharmacogenomics and serves as Scientific Director of the Cancer Digital Intelligence Program at Princess Margaret and the AI Hub at the University Health Network. He is also Head of Data Science at the Structural Genomics Consortium. His research harnesses artificial intelligence (AI) and machine learning (ML) to integrate large-scale chemical, radiological, and (pharmaco)genomic datasets. By developing predictive models for drug development, cancer progression and treatment response, his team aims to accelerate innovation in precision medicine and ultimately improve outcomes for patients.

WHEN

March 27th 2026
10:00 am

WHERE

ULB Plaine, Building NO, 5th Floor.
Salle Solvay