Single-cell omics is a rapidly evolving field with numerous fronts of growth. In our earlier blog entries, we examined a variety of open-source single-cell visualization tools, each designed to effectively examine single-cell data through unique user workflows. You can read Part 1 and Part 2 of the blogs here.
With several recently released applications centered around spatial single-cell data and virtual reality, respectively, part 3 of our blog series on single-cell data visualization applications takes a unique turn. While their long-term adoption will become clear only with time, these tools undoubtedly provide a fresh perspective on analyzing single-cell datasets over the 'traditional'(for lack of a better word) applications.
VR over single-cell data is a very fresh perspective.
The single-cell data, even after dimensionality reduction, has nuanced feature exploration requirements. The cells are usually clustered in a 2D or 3D space for exploration. The 2D representation of the data provides a very limited perspective of the data. The visualization of even three-dimensional (3D) projected data faces many limitations, the majority of them being that- these 3D plots often lack interactivity, hindering actions such as selecting cells for further analysis. To address these challenges, virtual reality (VR) is emerging as a powerful solution for visualizing 3D scientific data. In VR, researchers perceive their data as genuine 3D objects, and with their hands operating within the virtual space, they enjoy virtually unlimited interaction possibilities.
Spatial data complements high-dimensional omics data, particularly in single-cell analysis. When combined with expression data, spatial information provides a deeper understanding of the roles cells play within tissue architecture. This integration allows researchers to explore how gene expression, structure, and layout contribute to cellular function more comprehensively. Spatially resolving the cell expression provides valuable insights into tissue architecture and cellular interactions reliably in comparison to the standalone expression data.
Specifically from the context of diseases like cancer, understanding the spatial organization of tumor cells within the tumor microenvironment is crucial for elucidating disease progression, treatment response, and resistance mechanisms. Spatial scRNA-seq data can provide insights into the spatial heterogeneity of tumor cells, the interactions between tumor cells and stromal cells, and the immune cell infiltration patterns within the tumor.
With the fresh perspective of spatial omics and VR, let’s explore and compare a recent set of applications.
StarmapVis as an application allows the exploration of spatial single-cell data using an interactive VR-like interface on your device. The application also supports visualization of pre-calculated cell trajectory data. While the application is very low cost and provides a limited interactive experience of VR, it only supports a few visualizations and analyses. Further, it doesn’t allow exploration of single-cell data beyond some pre-calculated exported data slots from Scanpy or Seurat. If hosting limited pre-calculated data with a low-cost VR-like experience is a requirement, StarmapVis offers a great small-size option.
CellexalVR is the true amalgamation of VR experience meeting single-cell data exploration. While the application doesn’t offer visualization of spatial data in conjunction, it is a solid package delivering what it intends. The great feature of the application is cell selection using dimensionality reduction data, clustering, gene expression, and other mediums in a true VR environment. The cells selected can be explored with downstream features such as heatmap-based expression exploration and transcription factor network analysis. It is a great resource when cost isn’t an issue as it requires dedicated VR gear for exploration.
While CellexalVR is power power-packed visualization application, it requires a dedicated VR gear for use. SinglecellVR is a simple tool that allows basic exploration of pre-calculated single-cell RNASeq and ATACSeq data on existing devices such as mobile phones using simple sub 10$ viewing lenses. The app also allows exploration of pre-calculated trajectory analysis and RNA velocity data generated using different tools. SinglecellVR has dedicated scripts for exporting data out of Seurat, AnnData, Stream, scVelo, and other single-cell toolkits. While it is great for the aforementioned exploration without incurring much cost, it is very limited in its analysis and visualization capabilities.
The great thing about Vitessce is its customizability of layout and flows which can be designed by the user. Features like genomic tracks are a plus for the framework which is not offered by other applications. The applications support the concurrent visualization of data from multiple modalities, cell clustering, spatial coordinates, and other different plots generated from the data. It can work as an ipywidget that can be launched from a python kernel ipython notebook or as a html widget from RStudio. Overall, calling Vitessce an application might be incorrect, as it is a visualization framework working as an HTML widget to visualize spatial single-cell data but it fulfills all the roles when used as a traditional application. We especially aim to explore Vitessce more, to understand how it can be used to build scaleable visualization mediums over a large repository of single-cell data.
As we mentioned at the beginning, the long-term adoption of these applications is unclear but they surely challenge the creative ways to visualize single-cell data. Seeing the individual cells in a virtual space and toying around them was particularly fun for us. What is your opinion on these applications, ping us and we will be happy to discuss more things on the matter.
Elucidata offers a whole suite of solutions for single-cell data analysis and visualization, you can check it out here. Or if you would like to understand the single-cell capabilities in more detail reach out to us or email us at info@elucidata.io.