Proteomic Meta-Study Harmonization, Mechanotyping and Drug Repurposing Candidate Prediction with ProHarMeD

Abstract

Mass spectrometry-based proteomics plays a vital role in biomarker identification and understanding disease mechanisms through mechanotyping. However, utilizing publicly available proteomics data for these purposes poses challenges due to data diversity involving various sources, model organisms, and assay systems. Furthermore, the lack of identifier standardization hinders effective data integration across publications. To tackle these issues, we introduce ProHarMeD, a versatile solution available as a web tool, Python library, and R package. It harmonizes proteomics biomarkers from multiple studies, facilitating ID and name conversions between protein and gene levels and organisms through ortholog mapping. The tool identifies shared IDs among studies, proposes disease mechanisms, and interactively suggests potential drug repurposing candidates. ProHarMeD's efficacy is demonstrated using four bone regeneration studies. For those, the intersection size of shared biomarkers was increased by 50 % through ID harmonization compared to the non-harmonized biomarker lists. It also reveals a potential disease mechanism with corresponding drug targets, including Fondaparinux, a licensed drug candidate to treat thrombosis. However, it is also known to improve bone healing, which provides validation for our ID harmonization approach. With ProHarMeD, researchers have a one-stop proteomics suite for meta-analyses on proteomics data, ID conversion and remapping evaluation, and for bridging gaps between proteomics, disease mechanism exploration, and drug repurposing. The tool is accessible at https://apps.cosy.bio/proharmed/.