Spatial EcoTyper is a framework for identifying spatially distinct multicellular ecosystems–termed spatial ecotypes (SEs)–from single-cell spatial transcriptomics and for recovering them from spatial, single-cell, and bulk expression data. Liquid EcoTyper is an interpretable AI framework that quantifies spatial ecotypes from cfDNA methylomes, allowing noninvasive inference of spatial biology in tumor microenvironment.
We have defined nine spatial ecotypes conserved across melanoma and carcinomas (Zhang/Brown/Usmani et al., Nature 2026). This website allows users to recover these SEs from user-provided data through the following modules:
- Recovery of SEs from single-cell tumor profiles: Recovery of spatial ecotypes from single-cell gene expression data, including scRNA-seq and high-resolution spatial transcriptomics such as MERSCOPE, Xenium, and Visium HD
- Deconvolution of SEs from bulk tumor profiles: Inference of spatial ecotype abundance from bulk tumor gene expression data, including bulk RNA-seq and spatial transcriptomics platforms (e.g., Visium, legacy ST)
- Inference of SEs from plasma cfDNA methylomes: Inference of spatial ecotype levels from plasma cfDNA methylation profiles
- Download: data associated with the Spatial EcoTyper publications.
To note, this website is designed for SE recovery and inference only. Discovery of new spatial ecotypes from single-cell spatial transcriptomics data is supported through the Spatial EcoTyper R package available on our GitHub repository.
Please send questions, issues, and/or licensing requests to: spatialecotyper@gmail.com