Title:	Contextual modelling of IDPs in cancer hubs - from proxiomes to direct interactions
DNr:	Berzelius-2026-304
Project Type:	LiU Berzelius
Principal Investigator:	Maria Sunnerhagen <marsu@ifm.liu.se>
Affiliation:	Linköpings universitet
Duration:	2026-10-25 – 2027-05-01
Classification:	10601
Keywords:

Abstract

Highly dynamic intrinsically disordered regions (IDRs) are present in ~70% of proteins in the human proteome, and adopt conformational ensembles rather than fixed 3D structures, both alone and in the context of complexes. IDRs are particularly overrepresented among oncoproteins, where their dynamic properties facilitate interactions with multiple partners and/or binding sites in so-called “hubs” that facilitate rapid response to cellular signals. However, the structure-functional understanding of how IDP hub proteins coordinate large numbers of interactions in cellular time and space remains significantly underexplored. While the >230,000 high-resolutions structures in the Protein Data Bank (PDB) have been instrumental in the development of revolutionary structure-prediction tools such as AlphaFold and its variants (2,3), they do not provide reliable information on IDPs. Integrated approaches have to date contributed only »900 IDPs/IDRs to the Protein Ensemble Data base (PED) and the integrative/hybrid method database (PDB-IHM) – a small dataset that is insufficient to understand, or predict, IDP function or train new AI models for this purpose. In this grant, we aim to enhance state-of-the-art predictive modelling of IDP interactions by actively integrating proximity profiling data from biotinylated interactomes assayed by mass spectrometry (BioID). Our approach will leverage data on regulatory protein hubs in cancer where we have direct access to experimental validation and will be directly applicable to other proteins with relevant deposited data. The experimental BioID technique is high-throughput and uniquely identifies both structural and transient/dynamic protein interactions within cells, providing insights into dynamic biology in-vivo. However, BioID alone cannot distinguish direct interactions from indirect, or proxymediated, nor can it resolve structural details of these interactions. To address these limitations, we will develop novel pipelines for predictive interaction modelling with BioID data, complemented by validation through a comprehensive set of cellular and biophysical approaches accessible to us. This integrated strategy will enable structural and dynamic characterisation of IDP interactions based on their native cellular context, opening novel modes to target critical IDP interactions in cancer. Specific aims: AIM 1: Develop a computational 2D modelling pipeline tool to obtain structural resolution within BioID proxiomes and evaluate whether this can be used to discriminate between direct and indirect binding. AIM 2: Extend the pipeline to resolve IDP interactomes at 3D structural resolution for MYC and other cancer targets and thereby resolve IDP dynamic regulation of multidomain protein assemblies, AIM 3: Explore spatiotemporal 4D description of IDP binding events as a basis for hypothesis-based invitro and cellular validation and discovery as well as for interruptive in-silico targeted protein design, with critical implications for cancer biology and therapeutics. The project is a collaboration with researchers at LiU, KI and University of Toronto who are eager to use BioID as hypothesis-driving also on a molecular level, with DDLS funding as described below.