Comprehensive analysis of conformational ensembles of disordered proteins using IDPET python package

Intrinsically Disordered Proteins (IDPs) represent a distinct class of biomolecules characterized by their lack of a stable three-dimensional structure, enabling them to perform diverse regula- tory and functional roles in cellular processes. This thesis investigates the conformational com- plexity of IDPs through a large-scale analysis of structural ensembles from the Protein Ensem- ble Database (PED) using the IDPET Python package, a novel computational tool developed to facilitate the study of IDPs. The study begins with a comprehensive global analysis of PED ensembles, extracting features such as radius of gyration (Rg), end-to-end distance, solvent-accessible surface area (SASA), and asphericity. These metrics provide quantitative insights into the compactness, flexibility, and shape variability of IDPs. Correlation analysis revealed significant interdependence among features, highlighting coordinated structural behaviors crucial for IDP function. Dimensionality reduction techniques, including Principal Component Analysis (PCA), t- Distributed Stochastic Neighbor Embedding (t-SNE), and Uniform Manifold Approximation and Projection (UMAP), were employed to visualize high-dimensional data, facilitating the identification of conformational substates. A case study on the SH3 protein demonstrated the utility of IDPET in uncovering structural similarities and variations across ensembles, offering new perspectives on their functional implications. The findings emphasize the adaptability and functional versatility of IDPs, as reflected in their structural heterogeneity and dynamic behavior. This thesis underscores the value of in- tegrating advanced computational tools and systematic analyses to explore the relationship be- tween IDP structure and function. The methodologies and insights presented lay a foundation for future studies aimed at unraveling the biological roles of IDPs and their potential as thera- peutic targets.