Glencoe Software recently presented the latest use cases for OMERO Plus at SLAS in Boston, and here we highlight in detail new solutions for Spatial Transcriptomics data management and mining.
Spatial Transcriptomics is a growing but tumultuous domain, where now more than ever the reliance on a singular vendor’s data management system can be limiting. As a vendor and domain agnostic platform, OMERO Plus allows scientists to not only manage but also integrate spatial data across techniques and vendors of choice.
The dimensionality produced (targeted panel of genes or whole transcriptome) and questions asked (discovery-based or informed by single-cell or nucleus bulk RNA sequencing) by various spatial transcriptomics techniques are quite varied, but one challenge remains: to explore and link gene expression changes with spatial tissue features. OMERO Plus has developed a suite of tools to manage feature data alongside image data, not only for greater traceability but also to enable easily sharable browser-based visualizations and data mining. These tools and how they apply to Spatial Transcriptomics data are detailed below:
PathViewer is a browser-based viewer designed specifically for Digital Pathology data, both in the smooth navigation (including zoom, panning and rotation) of these large datasets and the control of brightfield or multiplexed fluorescence visualization settings. Spatial data representing segmented single cells or spots can be overlaid with the full resolution image data, and these objects can be filtered or color coded based on custom features managed in tabular data stores within OMERO Plus. Use PathViewer to see the heatmap of a particular gene’s expression across the tissue, and zoom in to hot spots to understand the biology in that region. Because PathViewer URLs are parametric, they can be used to easily share this data with colleagues, including the indication of a particular location of interest and the maintenance of specific visualization settings.
Pageant is another browser-based tool which maintains a reference to the image view like PathViewer, but instead centers the particular objects (cells, nuclei, spots, etc.) of interest, both as a tabular view of per-object features and as a grid of object thumbnails. Pageant supports dynamic filtering of this object-level data in numerous ways: based on a region of the image, thresholding of one or multiple features, and interactive filtering on plots. Use Pageant to view the distribution of fold changes of a particular gene across cells, to filter down to the cells which show high expression of a single or set of genes, and to connect the filtered cell population to their location(s) within the tissue, even for rare cell types.
Tabular data stored in OMERO Plus can be integrated into custom AI routines using tools like omero2pandas. This package allows users to download and upload OMERO.tables data as pandas dataframes, with additional integration aimed towards Jupyter Notebooks. This provides access to the full suite of Python data science and machine learning packages when working with OMERO data. Omero2pandas is another step towards making OMERO Plus the data engine of choice for data analytics and AI in bioimaging. Use omero2pandas to classify cells or spots in a custom way and write this data back to OMERO Plus for further exploration and sharing in PathViewer and Pageant.
Most importantly, we know that each data type and question motivates purpose-built tooling. OMERO Plus is a platform with an open source core, which means data managed within it is accessible to other third party and custom software of choice. We are committed to open data models and file formats, which enable interoperable workflows throughout the data lifecycle.
Want to transition your Spatial Transcriptomics data to an open platform today? Contact the Glencoe Software team to learn more.
Visium datasets were acquired from 10x Genomics (https://www.10xgenomics.com/datasets) and are licensed under the Creative Commons Attribution license.
Orion datasets were providing by RareCyte.