The goal of the conference is twofold. First, we aim at highlighting recent developments at the interface between machine learning, mathematical modelling, and biological data. More specifically, protein and nucleic acid sequence and structure data. Second, we aim at fostering exchanges and new collaborations toward a better vision of future challenges, and their practical implications for society. In particular, for medicine and protein engineering. These challenges include the reconstruction of interactomes at large scale at the residue resolution with the ability to sense the impact of sequence variations such as point mutations, and the accurate prediction of protein conformational states and continuous heterogeneity. The combination of machine learning and formal methods holds great promise in addressing these challenges, and we welcome contributions pushing the field forward in these directions.