From networks to mechanisms in health and disease

• Applicant: Dr. Katja Luck
• Subject Area: Bioinformatics and Theoretical Biology ; Biochemistry
• Term: since 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 449991970

Proteins mediate cellular functions by interacting with each other and with other biomolecules such as DNA, RNA, or metabolites. Thus, protein interactions determine the function of a gene and its protein product. Understanding which molecular function is mediated by a protein-protein interaction (PPI), its mechanism and which mutations can perturb that interaction and potentially cause disease requires knowledge about which amino acids contribute to the interaction, i.e. form the PPI interface. Existing experimental strategies to determine the interface of a PPI are laborious and error-prone. Computational methods for PPI interface prediction are unreliable and flawed. Our lack of knowledge about the molecular functions and underlying mechanisms for most human proteins is responsible for the extremely slow pace at which functional effects of variants in protein sequences from patient cohorts are identified. This resulted in an incredible amount of uncharacterized genetic variants, and most patients with rare diseases, i.e. those suffering of neurodevelopmental disorders (NDDs), remain without a genetic diagnosis and treatment despite their genomes or exomes being sequenced. The molecular mechanisms underlying NDDs are poorly understood. Surprisingly, most of the many genes that are considered causal for NDDs when mutated, are expressed throughout the human body at uniform levels. This suggests that studying the cellular context of NDD risk proteins, i.e. their interaction partners within the relevant cell type or brain sub-region, is critical for understanding the molecular mechanisms that mediate the brain specificity of NDDs. However, experimentally generated cellular context-specific PPI data is sparse due to difficulties in manipulating primary cell types, organoids, and tissues. To tackle these challenges, I propose to first, develop an accurate PPI interface predictor; second, develop a cross-linking mass spectrometry protocol for the efficient mapping of PPI interfaces; third, employ computational and experimental strategies to characterize the interfaces of PPIs involving NDD risk proteins; fourth, employ generated PPI interface information to predict the pathogenicity of variants in NDD risk proteins, followed by experimental testing in vitro, and fifth employ a PPI-centric integrative systems approach to predict molecular mechanisms that mediate brain-specific phenotypes of NDDs, followed by in vitro experimental testing of predicted mechanisms to be perturbed by known pathogenic but not benign variants in NDD risk proteins. This study will result in novel computational and experimental methods that enable identification of molecular mechanisms from PPI data, and will generate novel insights into the molecular mechanisms that likely underlie normal and pathological brain development. The combination of structural, computational, and systems biology with experimentation represents a major advancement over previous studies in the field.

• DFG Programme: Independent Junior Research Groups