Protein's cavities Bestiario: an atlas of proteins cavities

ANA JULIA VELEZ RUEDA; FRANCO LEONARDO BULGARELLI; GUSTAVO PARISI; BUGNON, LEANDRO A

Montevideo

Encuentro; Khipu 2023; 2023

On their surface, proteins are shaped into numerous cavities and protrusions that create unique microenvironments for ligand binding, catalysis, or other biologically relevant processes. Despite their biological and biotechnological relevance and their potential impact on various research areas in medicine, drug design, and evolutionary biology, there is still no exhaustive and comprehensive research on protein cavities combining sequential, structural, and evolutionary perspectives. We propose a representative set of features that enable machine learning, such as clustering and classifying the different types of protein cavities into functional categories. To this end, we combined classical bioinformatics techniques for molecular characterization with unsupervised learning techniques. We created and curated a dataset containing all proteins with known structures and their features, which is currently available in CaviDB. In this dataset, we computationally predicted the cavities present in each of the proteins and characterized them structurally and sequentially using a set of 47 features. We explored the data using nonlinear projections with UMAP and Self-Organized Maps. Using biological information such as volume, drug content, and polarity, we were able to cluster representative cavities and use them to reconstruct protein families.