Tripling human Phosphorylation Site functional Annotation to enhance Individualized treatment recommendations from Clinical phosphoproteomics data (PhoSAIC)

• Applicant: Dr. Matthew The
• Subject Area: Medical Informatics and Medical Bioinformatics ; Bioinformatics and Theoretical Biology
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 537476536

In the rapidly advancing field of precision oncology, the functionalization of phosphorylation sites (p-sites) emerges as a pivotal strategy. This project aims to functionalize p-sites for application in precision oncology programs through the utilization of both public and in-house data alongside machine learning-based inference methods. To fulfill the project’s main aim, we delineate the following three objectives: 1. Systematically capture public phosphoproteomic data for p-site functionalization: Extend ProteomicsDB with phosphoproteomic data from over 100 new and existing perturbation and clinical studies (>1,000 experiments, >10,000 MS/MS runs) using standardized data analysis workflows and data harmonization steps. 2. Functionalize phosphorylation sites: Put kinases and phosphorylation sites in functional contexts (e.g. drugs, tissues, pathways and kinases) by annotation from direct experimental evidence and inferences based on guilt-by-association using interpretable machine learning methods. 3. Integrate functionalized p-sites in precision oncology programs: Leverage the extended p-site annotations to improve individualized treatment recommendations by providing molecular oncologists easy access to clinically relevant p-site information and by augmenting clinical basket annotations. Based on preliminary data, we conservatively estimate that we can annotate three to four times the number of p-sites currently annotated in PhosphoSitePlus. To ensure accessibility of the research data according to the DFG guidelines, this is accompanied by a web-based phosphoproteomic resource integrated in ProteomicsDB. This gives instant access to the contexts in which phosphorylation sites of interest have been observed before, including information on how we arrived at a drug, pathway or kinase annotation from the data. As part of our ongoing phosphoproteomic patient profiling efforts, these annotations are then integrated and evaluated in generating individualized treatment recommendations for molecular tumor boards across Germany.

• DFG Programme: Research Grants