Visual analytics for whole-brain spatial transcriptomics

Abstract

During the past decades, advances in sequencing and omics technologies have upgraded the scale and granularity, and accelerated the pace of biological research. However, exponential data growth has simultaneously created a bottleneck in exploration and interpretation. As data volume and complexity exceed the capabilities of standard computational pipelines, it has become increasingly challenging to transform the computational power of the machine to the intuition of the researcher.

To address how we meaningfully explore high-dimensional, spatially organized datasets, this thesis presents visual analytics tools designed for spatially resolved transcriptomics data. The central aim is to facilitate the exploration and interpretation of complex single-cell and spatial data in an integrated manner. By mapping analytical tasks to specific biological questions, these tailored tools address distinct problems, ranging from the exploration of local cell neighborhoods and co-expression modules to multi-modal integration and cross-species comparison. Ultimately, this thesis demonstrates that visual analytics are essential instruments for extracting biological insights from the rapidly expanding field of spatially resolved transcriptomics data.