LIGAND-AI: artificial intelligence and small molecule drug discovery - an experimental and data roadmap

• Grant number: 25/12632-9
• Support Opportunities: Research Projects - Thematic Grants
• Start date: November 01, 2025
• End date: October 31, 2030
• Field of knowledge: Physical Sciences and Mathematics - Chemistry - Organic Chemistry
• Agreement: European Commission (Horizon Europe)
• Principal Investigator: Ronaldo Aloise Pilli
• Grantee: Ronaldo Aloise Pilli
• Principal researcher abroad: Aled Morgan Edwards
• Institution abroad: Structural Genomics Consortium Toronto , Canada
• Host Institution: Instituto de Química (IQ). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• City of the host institution: Campinas
• Principal investigators: Claudia Maria Bauzer Medeiros ; Katlin Brauer Massirer
• Associated researchers: Bruno Sérgio do Amaral ; João Marcos Batista Junior ; Lucas Rodrigo de Souza ; Mário Henrique Bengtson ; Micael Rodrigues Cunha ; Priscila Zonzini Ramos ; Rafael Junqueira Borges ; Rafael Lemos Miguez Couñago

Abstract

Potent and selective small molecules targeting specific proteins are essential tools for investigating biological processes and for the development of new therapeutics. Despite their importance, current small molecules only target approximately 10% of the human proteome. This limited coverage persists because the development of new chemical inhibitors remains time-consuming, expensive, and technically challenging, even with current modern technologies. Artificial intelligence (AI) offers the potential to accelerate drug discovery; however, its progress is constrained by the scarcity of large-scale, high-quality experimental datasets needed to train robust predictive models. To address this limitation, the LIGAND-AI initiative proposed and coordinated by the Structural Genomics Consortium, the SGC, (Innovative Heath Initiative/Horizon global initiative) aims to generate an unprecedented experimental dataset describing the interactions of billions of small molecules with 2,000 diverse human proteins. The project integrates complementary screening platforms, including chemical fragment screening using Weak Affinity Chromatography coupled to Mass Spectrometry (WAC-MS) developed at CQMED. In addition, the large project will use screenings of Enantioselective Affinity Selection Mass Spectrometry (EAS-MS) and DNA-Encoded Libraries (DEL). The screening will generate high quality experimental data for training IA algorithms and will be followed by validation with biophysical methods such as Surface Plasmon Resonance (SPR), and biochemical assays. As a key data-generation node within LIGAND-AI, CQMED-Unicamp will lead the production of 100 target proteins selected by Brazilian research groups and perform local screenings using DEL and WAC-MS. WAC-MS fragment data will also be used to evaluate if DEL data can enhance hit confirmation rates and prioritization. In addition, our group will contribute to the chemical synthesis of AI/Machine Learning predicted ligands for lead molecules generated in both, local and collaborative manner with the LIGAND-AI consortium, to supporting model validation and iterative improvement. For four selected targets, we will advance validated leads into high-quality chemical probes, enabling functional studies. By contributing to this effort CQMED-Unicamp will play a strategic role in overcoming current limitations in chemical probe discovery, strengthening local expertise on the therapeutic potential of the human proteome and for global breakthroughs in biological research and drug discovery. (AU)