Detalle de: Mapping shared and unique features in spatial transcriptomics through multivariate curve resolution

• MCR-ALS delivers interpretable insights into spatial transcriptomics data. • The proposed workflow uncovers both shared and tissue-specific attributes. • MCR-ALS results align closely with standard analytical pipelines. • Multiple tissue analysis using MCR-ALS reveals common biological features. Spatial biology is poised to play a pivotal role in enhancing our understanding of biological systems. Recent advancements have led to the development of a number of analytical pipelines, particularly within the framework of spatial transcriptomics. However, the analysis of spatial transcriptomic data remains computationally challenging. The Multivariate Curve Resolution Alternating Least Squares (MCR-ALS) model has proven to be a powerful chemometric approach, offering a more interpretable representation of complex spatial data compared to other exploratory approaches such as principal component analysis for not imposing an orthogonality constraint. Despite the growing use of similar models, MCR-ALS has yet to be tested for analyzing spatial transcriptomics data. In this study, a critical evaluation of the potential of MCR-ALS-based approaches to complement this key step in spatial transcriptomics analysis is conducted. Specifically, the MCR-ALS evaluation is performed on four samples of European sea bass testis at different early-maturation stages. Our results demonstrate that MCR-ALS is able to provide an accurate interpretation of the data when analyzing tissues both individually and simultaneously. The bilinear resolution effectively identified key spatial regions, which were putatively assigned to specific gonad compartments, in agreement with histological analysis. Furthermore, MCR-ALS models yielded results consistent with those from a standard spatial transcriptomics pipeline, particularly in examining gene expression profiles in specific gonadal regions. Therefore, integrating this chemometric tool into spatial transcriptomics workflow offers significant advantages for unraveling complex biological processes.