Spatial transcriptome technologies are being widely used, however some spatial analyses of transcriptomes currently fall short of the single-cell level of resolution. For the purpose of placing gene expression in a spatial context and delineating the spatial distribution of cell types within a tissue, theA team of researchers from Sweden has proposed a model-based probabilistic approach: stereoscope, which uses single-cell data to resolve mixtures of cells in spatial data.
What is a stereoscope?
ought tomouldThe framework utilizes the single-cell data to infer estimates of the proportion of each cell type at each capture location in the spatial data, thereby eliminating the need for spatialdata analysisThe need for any interpretation or annotation of abstract entities such as elements or clusters at the time.
STEREOSCOPE OVERVIEW: Single-cell data is first used to characterize the expression profile of each cell type, and then combinations of these types are found within each capture location to best interpret the spatial data.
The research team has implemented this method in code as a stereoscope namedexpand one's financial resourcespython packagereleased, it performs a deconvolution process and spatial mapping of cell types that is seamless, translatable by a variety of techniques, and does not require any preprocessing of the data.
Evaluation and application of the stereoscope
/ Technical evaluation /
To demonstrate the utility of the stereoscope, the team used data from different experimental platforms and spatially mapped cell types from the mouse brain and the developing heart, which aligned as expected.
Overview of mouse brain results
Summary of estimated cell type proportions in the developing heart, all from dh-B section
To illustrate how the stereoscope can be used in conjunction with other spatial techniques, the team analyzed Slide-seq data from the hippocampus and cerebellum that successfully reproduced the technique's originally published results.
In addition, the research team designed a program to collect synthetic data from real single-cell data similar to those obtained from space-based technologies, comparing the stereoscope with two recently published methods (DWLS and deconvSeq), which confirmed that the implementation of the stereoscope is superior to the other two methods.
Results of one-sided paired Wilcoxon signed rank test
/ technical application/
By design, the stereoscope is suitable for any type of spatial data and its applications are widespread:
Evaluate the presence and properties of tumor-infiltrating immune cells in cancer or depict the cell types that make up the tumor microenvironment;
Cell type interactions inferred from spatial co-localization patterns;
The abundance of cell types within the relevant anatomical region was determined by examining the distribution of proportional values in the tissue;
...
Information about the spatial distribution of cell types can be used as the basis for many different analyses.
stereoscopehardwareThe package is available at the following link:/almaan/stereoscope.
bibliography
Andersson A, Bergenstråhle J, Asp M, et al. Single-cell and spatial transcriptomics enables probabilistic inference of cell type topography[J]. Communications biology, 2020, 3(1): 1-8.
Images from Communications biology official website and references, please contact to remove any infringement.