Introduction: What Is Spatial Transcriptomics?
Spatial transcriptomics is a rapidly growing technology that allows researchers to measure gene expression while preserving the spatial location of the cells or tissues. Unlike traditional RNA-seq that loses all spatial information, spatial methods can tell not just what genes are expressed, but where they are expressed inside tissue structures.
This has opened many exciting possibilities in cancer research, neuroscience, developmental biology, and many other fields. It provides new ways to understand tissue organization, cellular interactions, and disease microenvironments.
Although the concept of spatial transcriptomics is straightforward, analyzing the data is highly complex and requires specialized expertise.
Major Spatial Transcriptomics Platforms
| Platform | Resolution | Strengths | Weaknesses |
|---|---|---|---|
| 10X Visium | ∼55 μm spot size (multiple cells per spot) | High throughput, full-transcriptome profiling | Not single-cell; spot contains mixed cell types |
| 10X Xenium | Subcellular (∼1–2 μm) | High resolution, detects single molecules | Pre-selected target gene panels only |
| NanoString CosMx | Subcellular (∼1 μm) | Highly customizable target panels | Limited throughput compared to 10X |
Each platform has its own advantages and technical challenges. Visium can capture thousands of genes easily, but Xenium and CosMx provide much finer resolution at the expense of panel size and complexity.
Practical Challenges in Spatial Transcriptomics Data Analysis
1. Quality Control of Spots
Especially for 10X Visium, not every spot contains useful biological information. QC steps such as filtering low-UMI spots, high mitochondrial reads, and background noise are necessary.
2. Normalization and Scaling
Spatial data often requires customized normalization methods beyond total count scaling due to regional expression differences and batch effects across slides.
3. Detection of Spatial Domains
Clustering spatial data must incorporate coordinates. Methods like Seurat Spatial, BayesSpace, and SpaGCN help preserve spatial patterns during clustering.
4. Spatially Variable Genes Identification
Tools like SpatialDE, SPARK, and differential spatial analysis are used. Proper multiple-testing corrections and noise filtering are critical.
5. Integration with Single-cell RNA-seq
Deconvolving Visium spots using scRNA-seq (via stereoscope, Cell2location, etc.) significantly improves biological interpretation but requires careful matching and QC.
How We Can Help
At AccuraScience, we are experienced in full spatial transcriptomics analysis, including:
- Raw data processing, image alignment, spot QC
- Detection of spatial domains
- Spatially variable gene analysis
- Deconvolution with matched scRNA-seq
- Advanced visualization and trajectory analysis
We have worked with 10X Visium, 10X Xenium, and NanoString CosMx platforms. We understand practical issues such as capture efficiency, background correction, and customized panel analysis.
Contact Us
If you are working on spatial transcriptomics data or planning a new project, send us an inquiry, chat with us online (during business hours 9-5 Mon-Fri US Central), or reach us other ways!
We look forward to helping you turn spatial information into new biological discoveries.
About the Author: Zack Tu holds a B.S. in Biochemistry, an M.S. in Software Systems, and a Ph.D. in Pharmacology. A computational genomics expert with over 20 years of experience, Zack has provided informatics support to numerous research groups across academia and industry. He led the development and operation of the core bioinformatics infrastructure at the University of Minnesota’s research sequencing center (2001–2012) and contributed to a clinical genome sequencing facility. Since joining AccuraScience in 2013, he has worked extensively with NGS data generated from Illumina, Roche/454, Ion Torrent, PacBio, and Oxford Nanopore platforms. His areas of expertise include genome assembly, bulk and single-cell RNA-seq, 16S rRNA profiling, shotgun metagenomics, variant detection and interpretation, and computational modeling of mutation calling. More recently, he has contributed to projects involving spatial transcriptomics, integrative multi-omics analysis, and clinical bioinformatics applications. More at https://www.accurascience.com/our_team.html.