Juan Diego Guarimata presentó un poster en el ICTP-SAIFR en São Paulo, Brasil

Felicitamos a Juan Diego Guarimata por haber presentado el poster titulado:

"AI-Driven Generative Pipeline for Drug Discovery in Neglected and Emerging Infectious Diseases"

en el 4th School on Data Science and Machine Learning -- International Centre for Theoretical Physics - South American Institute for Fundamental Research (ICTP-SAIFR), en São Paulo, Brasil.

Abstract: Neglected tropical diseases such as Chagas disease (Trypanosoma cruzi), leishmaniasis (Leishmania spp.), and arboviruses like dengue remain major global health challenges with limited therapeutic options. Our work develops an integrated computational pipeline combining bioassay-driven data mining, machine learning, and generative deep learning approaches to accelerate drug discovery for these pathogens. Initially, bioactivity data from PubChem assays targeting essential enzymes (e.g., topoisomerases, cytochrome P450s, trypanothione reductase in trypanosomatids, and NS5 methyltransferase/polymerase in dengue), as well as phenotypic screening assays, were curated, prioritizing candidate scaffolds. Building on these scaffolds, we implemented a generative chemistry framework based on MolGAN, which produces novel molecular graphs guided by reward functions incorporating Quantitative Estimate of Drug-likeness (QED) and Synthetic Accessibility Score (SAS). The system integrates structural pocket information from crystallographic data, filtering rules for drug-likeness, and iterative evolutionary strategies. Candidate molecules undergo 3D conformer generation and molecular docking with OpenEye OMEGA/FRED, closing an active learning loop where high-scoring ligands are reincorporated into model training. Selected compounds predicted to show improved binding are subjected to molecular dynamics simulations to validate interaction stability. The top candidates will be prioritized for chemical synthesis and experimental evaluation, bridging computational predictions with translational drug discovery. This pipeline highlights the synergy of AI, structural biology, and cheminformatics in delivering tractable leads for neglected and emerging infectious diseases.