UM academic presents at international structural biology and AI conference

Malta recently hosted the international conference Synergic Entanglements Between Structural Biology & Machine Learning: Looking into the Future, bringing together leading scientists working at the interface of structural biology, artificial intelligence, molecular modelling and biomedical discovery.

Held from 13 to 15 May, the meeting placed Malta at the centre of an important scientific conversation on how AI is reshaping the study of proteins, nucleic acids, molecular assemblies and disease.

The conference was organised by Annalisa Pastore (King’s College London), Gian Gaetano Tartaglia (IIT and Sapienza University of Rome), Silvia Onesti (Elettra Sincrotrone Trieste), and Elena Sultana (University of Malta, ERA-Shuttle project). Together, they brought to Malta a high-level international meeting connecting structural biology, machine learning and research infrastructure.

The programme included several internationally recognised scientists whose work has shaped modern structural and computational biology. Michele Parrinello is known for the Car-Parrinello molecular dynamics method and major contributions to molecular simulation. Christine Orengo is a leading figure in protein structure classification and evolution through the CATH database. Michele Vendruscolo has made major contributions to understanding protein misfolding, aggregation, and their links to neurodegenerative disease and therapeutic strategies. Gian Gaetano Tartaglia is known for computational work on protein-RNA networks, phase separation, and RNA biology.

Other prominent contributors included Franca Fraternali, Marta Carroni, Jianyi Yang, Trevor Sewell, Richard Garratt, and Vinothkumar Kutti Ragunath, reflecting the breadth of the field from cryo-EM and RNA structure to biomolecular dynamics.

Prof. Jean Paul Ebejer from the University of Malta contributed a talk entitled Learning Molecular Interactions: AI Methods for Predicting and Interpreting Protein-Ligand Binding.

His presentation focused on how AI methods can support drug discovery by predicting and interpreting protein-ligand binding, while stressing that binding is not a single problem but a set of linked questions: where a ligand binds, which molecules bind, how they bind, how strongly they bind and why a model makes a particular prediction. The talk also emphasised the need for chemical realism, robust benchmarks, uncertainty, explainability and experimental validation.

A central message from the meeting was that structural biology is entering a new phase. For years, the grand challenge was moving from biological sequence to molecular structure. With AI-driven structure prediction now transforming that landscape, the focus is shifting from structure to function: understanding how molecules interact, how biomolecular complexes assemble and behave, and how AI can help turn structural models into biological and medical insight.

More information about the conference is available online.