From Drug Discovery to Peptide Nanofibers: AI and Cheminformatics Tools for Bioactivity Prediction

The rational design of bioactive molecules, whether small-molecule drugs or functional biomaterials, requires computational frameworks capable of translating molecular structure into reliable predictions of biological activity. Our group at LabMol (Laboratory of Molecular Modeling and Drug Design, Universidade Federal de Goiás, Brazil) develops machine learning and cheminformatics approaches, broadly known as Quantitative Structure–Activity Relationships (QSAR), to build predictive models that accelerate drug discovery and provide ethical alternatives to animal-based toxicity testing. Over the past decade, we have applied these tools to neglected tropical diseases, including Malaria, Leishmaniasis, Zika and COVID-19, and to predictive toxicology, resulting in a suite of freely accessible open-science AI platforms: Pred-hERG, Pred-Skin, STopTox, and LabMol InsightAI, which are actively used by the global research community.

In this seminar, I will present our current research portfolio and discuss an emerging direction at the interface between cheminformatics and supramolecular chemistry. Epitope-functionalized peptide nanofibers represent a compelling class of bioactive materials whose function depends critically on the conformational dynamics of surface-displayed epitopes and their interactions with target proteins—a structure–activity problem that is conceptually analogous to those we routinely address in drug discovery. I will explore how QSAR-based ML models can be extended to incorporate conformational ensembles derived from molecular dynamics simulations of supramolecular peptide assemblies, providing a principled framework for connecting nanofiber architecture to bioactivity prediction. This conceptual bridge, from small-molecule drug design to functional nanomaterials, motivates an ongoing collaboration with the Sasselli group at CFM-MPC and opens broader opportunities for interdisciplinary work at the chemistry–physics–biology interface.