Spatial EcoTyper

method
spatial-omics
Published

May 18, 2026

Definition

Spatial EcoTyper is a machine-learning framework for identifying and profiling spatial ecotypes from spatially resolved cell states in tumour tissues.

What It Gives

  • Spatially informed cell-state embeddings
  • Conserved spatial ecotypes across samples
  • SE-specific cell states and marker programs
  • Recovery of SEs from spatial, single-cell, or bulk expression data

Core Idea

The method fuses cell-type-specific gene-expression variation across spatial neighbourhoods, then identifies multicellular ecotypes that recur across tumour samples.

Why It Was Needed

Earlier methods either used limited predefined protein markers, ignored spatial information, or struggled to integrate diverse samples, cancer types, and platforms. Spatial EcoTyper is designed to combine spatial information, cell-state resolution, and cross-sample integration.

In Papers

  • Zhang et al. 2026 introduces Spatial EcoTyper to define nine spatial ecotypes in human carcinomas and melanomas.