Spatial Transcriptomics: The Basics

Spatial Transcriptomics: The Basics

What is Spatial Transcriptomics?

Spatial transcriptomics (ST) technology allows you to map gene expression for tissues and cells and see how those patterns vary from region to region. ST offers a snapshot of gene expression and gives you context, so you know where that expression is happening within the overall tissue’s architecture.

From this information, you can glean more details about cell types (i.e., their characteristics, interactions, etc.). For example, using ST, you can:

  • identify which genes are active in the tumor and neighboring cells.
  • compare the gene expression of immune cells near a tumor compared to those farther away.
  • discover new therapy targets based on cell-to-cell interactions.
  • determine how cellular components communicate and adapt to (or resist) treatment.
  • gain a more complete view of the tumor microenvironment—beyond traditional bulk and single cell RNA sequencing.

Why is ST Important for Cancer Research?

Testing a Hypothesis

You might have a hypothesis about the mechanisms underlying a certain cell or the tumor environment. ST can help test your hypothesis, filling in missing pieces in your research puzzle. Using ST, you can leverage both sequencing-based technologies and imaging-based technologies to see gene expression within the context of the full tissue microenvironment. You can further refine your hypotheses and generate new questions to explore. From there, you can confirm your findings in the lab using other techniques, such as single cell RNA sequencing or functional experimentation.

Identifying Biomarkers

ST results also have clinical applications. You can use ST to identify molecular targets for potential medications and to assess new drug candidates to see how well they perform. ST can also help you in personalizing treatment, enabling you to identify patients who might be most likely to respond to certain medications, as well as for detecting drug resistance.

What Do I Need to Know?

Begin by defining your research question. Answering some basic questions will enable you to select the best ST technology and tool. For instance, ask yourself the following:

  • “What is my tissue type?” (e.g., fresh frozen tissues or formalin-fixed, paraffin-embedded [FFPE])
  • “How large is my sample?” (e.g., area, number of cells per tissue slice, etc.)
  • “Do I want a high resolution that gives details on specific cell types or niches, or would a broader spatial pattern or trend suffice?”
  • “Do I have a high-quality scRNAseq reference containing my cell populations of interest?”

Imaging-Based ST (i.e., Fluorescent Signatures or Probes)

Use this to map gene expression to features such as cellular components or cell boundaries, or to see how cells interact with neighboring cells. Image-based technologies also can give you sub-cellular information, letting you see details on certain compartments, such as the nucleus or area outside the cell. This technology works best when you need detailed information about what’s occurring in a very defined space, such as within a tumor. Image-based ST is available in panels of expression (hundreds to a few thousands of genes). To get the best results, you’ll want to customize your panels to answer your specific question. Your choice of profiling panel also can impact the level of detail you get from cell subpopulations.

Sequencing-Based ST (i.e., An Array with Spatial Barcodes)

Use this to see details on gene expression, especially within a large sample, such as an entire tumor section. This approach is ideal when you’re looking at the expression levels of a certain gene. You can see the position of your targeted gene in the tissue. Then, you can use next generation sequencing to determine the gene’s expression for a more detailed profile. Sequence-based ST varies in cellular resolution, from region-of-interest (ROI) and “spot,” to near cellular-level profiling. The higher the cellular resolution, the less likely you’ll be able to identify lower hierarchy cell subpopulations. This is the result of high gene drop-out (i.e., abundance of zeroes).

Tip: In general, use imaging-based ST if you need a lot of detail for a small area; use sequencing-based ST if you’re more interested in regional results (domain or niche-level analysis). In some cases, combining both sequencing and imaging approaches can help you overcome the limitations of ST, including a lack of resolution power that can make it difficult to capture the nuances of every cell and cell interaction in your sample.

Choosing a Tool for ST

You can select from a number of commercial tools. For imaging-based ST, some popular tools are MERFISH, seqFISH, CosMxTM, and Xenium. For sequencing-based ST, popular options include 10x Visium and Slide-seq, to name a few. Although some license-free, user-friendly tools are already available for exploratory data analysis (e.g., spatialGE, Galaxy, 10X Loupe), most studies require specialized data science training.

See NCI’s Center for Cancer Research’s Spatial Biology webpage for more information on ST tools.

SpatialGE

If you’re new to bioinformatics and looking for a simpler option, there’s a new NCI-funded tool, called spatialGE, that can help you get started.

A few quick facts about spatialGE:

  • It’s an open-access, point-and-click platform that guides you through the analysis pipeline, helping you organize data, conduct quality control, and identify and compare spatial gene expression patterns across multiple samples.
  • It’s designed to help you with data normalization, clustering (“spatial domain detection”), cell phenotyping/deconvolution, and spatial hotspot and gradient detection.
  • It’s available as a web-based application or download it as an R package if you want to perform more advanced data analysis and visualization.

What Can spatialGE Do?

Using spatialGE, you’ll have access to common types of ST analyses, including SpaGCN, STdeconvolve, and InSituType, which are particularly useful for cancer researchers, without needing additional programming expertise or specialized data science skills. Users can also download results in tabular form or as high quality images. You can store your data and analyses in a secure user account. The spatialGE tool also logs analysis tasks and their parameters to make it easier to reproduce your results.

How Do I Get Started?

You can analyze data from commercial ST platforms, such as 10x Visium or CosMx, for use in spatialGE. Or you can import your own data using a generic format. You can use text files, containing gene counts and spatial coordinates, or tissue image data. Select your parameters for analysis using slide bars and drop-down menus. You can conduct exploratory analysis using algorithms such as Principal Component Analysis (PCA), Uniform Manifold Approximation and Projection (UMAP), or heatmaps. Such spatially informed analysis allows you to detect where and how genes are expressed. You then can explore specific gene sets that exhibit “hot-spots,” as well as expression gradients (i.e., expression of gene X is higher in cells closer to tumor cells).

Tips to Keep in Mind

Complete each step. Because the application is modular, you won’t be able to jump ahead until you’ve completed each sequential task. For example, once you import your data, the next step is “QC and Data Transformation.” Once you complete that step, you’ll unlock the next step, the “Visualization” module. This stepwise progression helps guide you through the pipeline.

  • Find help when you need it. You can access documentation and “info-tags” along each stage of your analysis. These resources can help you learn about the user interface as you actually use it.
  • Remember that some analyses take time. It’s normal for some spatialGE methods to take time. You can instantly visualize gene expression in spatial context. However, for advanced spatial statistics, such as spatial gene set enrichment, it may take longer, depending on the number of samples and spots/cells in the data set. You have the option to receive an email notification once an analysis is completed (even if the application is closed).
  • Keep your data secure. Even though your data are fully protected by your password and private account, you should not upload protected health information data (PHI) to spatialGE, as per HIPAA’s designations of PHI.
  • Save your file. You can download your results in multiple formats, including tabular results and graphics. The “parameter files” also give you a history of the parameters for each module, so you can easily reproduce your results.

NCI ST Resources and Initiatives

Resources and Tools

Publications

Updated:
Vote below about this page’s helpfulness.