Year: 2026
Type: preprint
Abstract: Spatial omics technologies can now profile finely resolved cell populations in patient tissue at scale, offering a way to link their arrangement to outcomes. Yet images often differ in size, shape and cell content, and existing tests between patient groups mistake these differences for co-localisation or weight every image equally. We present spicyR v2, which avoids both problems with a fast, permutation-free test that accounts for these differences and weights each image by how precisely it measures co-localisation. In simulations and 74 analyses of 42 public datasets, it controlled false discoveries, had more power than spicyR v1 and was robust to tears, holes and differences in cell number that misled other methods. These analyses also revealed that many studies relating co-localisation to clinical outcomes are underpowered, and spicyR v2 can estimate the number of patients a new study needs. spicyR v2 analyses 100 million cells in minutes, making robust, well-powered tests of co-localisation practical for large spatial omics cohorts.
Institution: The University of Sydney
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7236312019
