Enhancing Proteomics Meta-Analysis through Linked Data and LLM-Driven Metadata Extraction

Proteomics meta-analysis has become an essential tool, enabling researchers to combine and compare findings from diverse proteomic studies and gain deeper
insights into protein functions, interactions, and disease mechanisms. However, the full potential of proteomics data can usually not be is exploited due to the
complexity, heterogeneity, and large volume of information involved in such analyses. Here, metadata extraction using large language models (LLMs) play a
crucial role for gathering various study details. By leveraging LLMs, the extraction, normalization, and integration of metadata from various sources can be
automated, thereby enhancing the accessibility and usability of proteomics data in the light of the FAIR principles. Furthermore, linking proteomics data to
broader biological knowledge through linked data principles opens up new possibilities for data interoperability and contextual understanding. By employing
standardized ontologies and semantic web technologies, proteomics datasets can be interconnected with genomic, clinical, and other omics data, fostering a
more holistic approach to biomedical research or toxicological assessment.