The introduction of multiplexing in cellular imaging enabled the identification and spatial localization of thousands of different RNA transcripts per cell in situ within tissues comprising hundreds of cells. All these new methods generate millions of datapoint with spatial coordinates: the significant challenge is now to make some sense out of those datasets. Here are three tools that could help you identify patterns from your in situ imaging spatial transcriptomics data.
Curious about the advances in spatial transcriptomics techonlogies? Check out a list of the available methods here.
Cytomap: the in situ imaging multidimensional toolbox
Cytomap is a toolbox whose aim is to regroup existing statistical analyses for single-cell spatial data under one roof, reducing the constraint to juggle between different software. The underlying principle behind Cytomap is to group individual cells into local “neighborhoods” based on their morphology, genotype, and spatial localization. These neighborhoods help define the basic blocks of different tissues. Once clustered into “regions”, these neighborhoods then reveal the overarching tissue structure, quantifies its composition, and evaluate its heterogeneity. On top of this 2D/3D reconstruction, Cytomap is also able to evaluate the interaction networks at the cell- and tissue- levels.
You can find Cytomap on GitLab, and here is a step-by-step demo on how to use it to phenotype your single-cell data.
TissUUmap: straightforward and interactive whole slide images for in situ imaging
TissUUmaps aims at displaying relevant information rather than heavy, raw spatial in situ transcriptomics data to have “visual representations of data leading to better human cognition and interpretation”.
TissUUmaps uses high-resolution 2D slide images, on which are overlayed additional spatially-resolved data such as transcriptomic profiles. By manually selecting regions of interest on the slide image, you can directly compare gene expression profiles between similar, or different anatomical structures. Advanced analysis is also available such as constellation of spatial gene expression profiles.
TissUUMaps is on Github, with tutorials available on the University of Uppsala’s website.
Hoping for tools to interpret spatial sequencing data rather than in situ imaging? You’re in luck with our other toolbox on visualizing spatial scRNA-seq data!
InSituNet: extract networks of spatially co-expressed gene profiles
If you’d like to go one step further and bypass manually exploring high-res imaging to move straight to the visualization of networks within your in situ spatial transcriptomic data, take a look at InSituNet. An addition for the Cytoscape 3 open-source software, InSituNet attempts to reduce the complexity of your dataset to emphasize on the main networks it can identify from the sequencing data.
InSituNet relies on the identification of spatially co-expressed transcripts. Each unique transcript sequence is displayed as a node, the size of which is proportional to the number of transcripts found. To simplify, when two different transcripts are spatially co-expressed in the tissue or region of interest, InSituNet then create a link between the relevant nodes. The frequency in which those two transcripts are found to be spatially co-expressed determines the distance between the nodes, thus creating a visually clear display of a network in a dataset rich in gene expression profiles.
InSituNet can be found in an app form here.
