Blogs

Spatial transcriptomics has rapidly emerged as one of the most transformative innovations in modern biology, empowering researchers to map gene expression while preserving the intricate architecture of tissues. Recent advances in the field have resulted in a variety of new platforms including Visium, Visium HD, and Xenium from 10x Genomics, and GeoMx and CosMx from NanoString Technologies (now part of Bruker). Developed by BioInfoRx in collaboration with Biogen, SpaceSequest (Sun, et al., 2025) emerges as a unified and elegant solution designed to harmonize data from all five major spatial transcriptomics platforms (Fig.1).

Figure 1. Components of SpaceSequest. The pipeline integrates five major workflows tailored to perform a series of analysis for each platform, including Visium, Visium HD, Xenium, CosMx, and GeoMx.

SpaceSequest provides a comprehensive framework for spatial transcriptomics data analysis by performing (1) standardized quality control and general data processing, (2) platform-specific analytical workflows tailored to each spatial technology, (3) automated cell type annotation and spatial deconvolution, and (4) generation of high-quality figures and analytical summaries. Furthermore, SpaceSequest seamlessly integrates with cellxgene VIP and Quickomics, enabling intuitive data exploration and interactive visualization.

Each workflow can be initiated by running its corresponding script, which automatically generates the required configuration files. The primary configuration files are written in YAML format, providing both standardized structure and flexible customization. To execute the pipeline, users typically need to complete a sample metadata file that records essential information for each dataset—such as sample names, file paths, and related metadata.

An overview of Visium data analysis using the SpaceSequest pipeline is presented in the blog. Each of the other supported platforms—Visium HD, Xenium, GeoMx, and CosMx—follows a similar structure in implementation and result generation, with adjustments tailored to their respective data formats and resolutions (Sun, et al., 2025).

Launched in 2019, the 10x Genomics Visium platform remains one of the most widely used and versatile commercial technologies for spatial transcriptomics. Each Visium slide is patterned with an array of barcoded 55 µm capture spots, available in two standard tissue capture sizes—6.5 × 6.5 mm and 11 × 11 mm—enabling high-throughput mapping of gene expression across tissue sections with spatial precision.

The Visium workflow in SpaceSequest begins with pre-processed outputs from 10x Genomics Space Ranger, which converts raw FASTQ files into binned count matrices and quality metrics. SpaceSequest first filters out low-quality spots—those with low gene detection, low UMI counts, or other poor-quality metrics—to retain only reliable data for downstream processing.

The pipeline applies UMAP for dimensional reduction followed by FindNeighbors and FindClusters to group transcriptomically similar spots. Two additional spatially aware methods, SpaGCN (graph convolutional network integrating expression, spatial coordinates, and histology) and BayesSpace (Bayesian model incorporating spatial priors), further refine domain detection and enhance spatial resolution. To infer the composition of mixed-cell spots, SpaceSequest integrates Tangram (deep-learning alignment of scRNA-seq and spatial data) and Cell2location (Bayesian cell-type abundance estimation), allowing researchers to map individual cell types within spatial regions. As a final step, the SpaTalk module can be activated to analyze ligand–receptor interactions and visualize intercellular communication networks within tissue domains (Fig. 2A).

A public dataset generated from the brain tissue of an Alzheimer’s disease mouse model was tested in SpaceSequest which has mutations in APP and PSEN1 genes. To illustrate representative results from this workflow, BayesSpace clustering groups transcriptomically similar spots across samples, revealing distinct and comparable spatial domains within the tissue (Fig. 2B). Similarly, Cell2location produces cell-type abundance heatmaps overlaid on the tissue image, illustrating the spatial distribution and relative abundance of different cell types (Fig. 2C).

Figure 2. 10x Visium data analysis. (A) Workflow for Visium data processing. (B) Spatial clustering results generated by BayesSpace. This run contains four samples, and BayesSpace identified shared clusters across them. (C) Predicted cell type abundance by Cell2location in one sample (PSAPP-TAM1), highlighting astrocytes, OPCs, microglia, and endothelial cells.

By bridging five major spatial transcriptomics platforms within a single unified framework, SpaceSequest represents a major step toward standardized, reproducible, and accessible spatial data analysis. The pipeline’s open-source nature encourages transparency and collaboration, empowering researchers to explore tissue biology from discovery to mechanistic insight without being limited by platform boundaries. Detailed methods, benchmark results, and use cases can be found in the full preprint: SpaceSequest: A Unified Pipeline for Spatial Transcriptomics Analysis. The source code, documentation, and tutorials are freely available on GitHub at https://github.com/interactivereport/SpaceSequest, providing an open gateway for the community to adopt, extend, and contribute to this evolving ecosystem of spatial biology tools.

References:

Sun, Y. H., Piya, S., Ouyang, Z., Chen, Y., Gagnon, J., Cao, S., Zhang, H., Song, B., Zhu, J., Chandratre, K., Yu, H., Hu, W., Ryals, M., Casey, F., Huh, D., & Zhang, B. (2025). SpaceSequest: A unified pipeline for spatial transcriptomics data analysis. bioRxiv. https://doi.org/10.1101/2025.09.15.676389