Poster #65: Julianna Pakstis MI

A Framework for Integrating Clinical Proteomics Data into Multi-Omics Discovery

Julianna Pakstis, MI; Nicole Feldman, MSIS; Surendra Babu Buddi, MHI; Allison Olsen, MA, MS; Ene Belleh, MBA, MLIS; Hossein Fazelinia, Ph.D.; Jennifer Roof; Lynn A. Spruce; Michael Xie, Ph.D.; Nadir Yehya, MD, MSCE

Poteomics enables comprehensive characterization of the protein landscape, offering deeper insights into the complexity of human phenotypes not often captured by genomic or transcriptomic data alone. Combining proteomics data with other omics data types alongside clinical data opens powerful possibilities for phenotypic-genotypic research not otherwise feasible. While pipelines are available to incorporate proteomics data into multi-omics analyses, streamlined documentation of methods – from sample preparation through data processing – is often lacking. This is due to the broad range of proteomic instrumentation and techniques as well as the absence of widely-adopted reporting guidelines. However, missing metadata can complicate integration with clinical and multi-omics data sets and hinder reproducibility. To address these issues, the CHOP Proteomics Core and Arcus Library Science team have collaborated to develop a standardized yet flexible framework for storing and sharing proteomics data that supports diverse technologies and can accommodate future methods. This structure makes CHOP’s proteomics data FAIR (Findable, Accessible, Interoperable, Reusable) and ready for integrated analysis alongside other biomedical datasets. This poster presents an example implementation of the structure in a recent proteomics project, showcasing practical applications and providing guidance on how the framework can be adapted for future studies. It formalizes the expertise of proteomics scientists who carry out the analysis with librarians’ knowledge of organizational best practices, enabling clinicians to systematically access and utilize results. In short, the structure requires proteomics scientists to generate a standard protocol yaml file that documents key metadata about the analysis. Comprehensive csv and txt manifest files match the protocol yaml files to individual participants, biosamples, and downstream data files. Once processed into the Arcus Archives, the manifest files provide a single key to integrate proteomics data with all other omics sequencing types and clinical data already stored there. Thus, researchers can augment even limited datasets with corresponding proteomics information.