Computational modelling of intrinsically disordered proteins – interaction, regulation, and disease

Recent discoveries suggest that intrinsically disordered proteins (IDPs) or regions (IDRs) may be just as critical as the structured proteins or regions.

Given the important and often coordinating roles of IDPs in central cellular processes, it is not surprising that numerous IDPs are associated with human diseases including cancer, autoimmunity, cardiovascular and neurological diseases However, due to lack of stable structures our understanding of how this biologically important class of proteins functions and governs disease is still remarkably restricted.In this proposal, we aim to develop a protein-protein interaction analysis pipeline that starting from proximity data can predict direct interactions (2D), build 3D models of the interactions and give clues on the timing and localization of binding events.

The pipeline will integrate of a number of methods we have previously developed to solve these tasks, with novel graph-representations of protein structure to make use of the latest developments in machine learning to improve and expand prediction accuracy.

All development will be performed in an iterative manner in  close collaboration with experimental groups, from which have access to relevant data setsIn the applied part of the project we aim to analyze the interaction networks of  four IDPs and their role in human health and disease: the oncoprotein MYC, the recently identified DIORA1 in autoimmune disease, TP53BP1 in breast cancer, and Apollo in DNA repair.