Bioimage Analysis in Deep Visual Proteomics: Advancing Transparency, Reproducibility, and FAIR Principles

Deep Visual Proteomics (DVP) leverages high-resolution microscopy and computational instance segmentation for ultrasensitive mass spectrometry to enable spatially resolved cellular and molecular analysis. Within this workflow, laser capture microdissection (LMD) plays a pivotal role by physically isolating image-defined regions of interest. Yet, the accuracy of LMD remains limited by the optical resolution of the imaging system, tissue preparation quality, laser cutting parameters and lacking automatisation. These constraints can result in cross-contamination from adjacent cells. DVP provides a conceptual pipeline to streamline the entire process, whereas the practical implementation in a fully reproducible way remains challenging for each individual problem. Here, we describe an optimised and automatised methodological framework that integrates image-guided region definition to enhance the reproducibility and precision of neutrophil excision as an exemplar use case. These advances strengthen the interface between microscopy and downstream proteomics, thereby improving the DVP workflow for the analysis of spatially defined immune cell populations.