henhaobio
Sign in

bioRxivBioinformaticsPreprint

Scalable and model-free detection of spatial patterns and colocalization

Qi LiuiD, Chih–Yuan HsuiD, Yu ShyriD

PDFDOI

Abstract summary

Claude Opus 5.5 · xhigh effort · 18 Oct 2026 · from the abstract · CC BY-NC 4.0

SpaGene finds spatial expression patterns without assuming a model, reportedly with more power and better scaling than existing methods, and extends to ligand-receptor colocalization.

Spatial omics platforms now profile molecular events genome-wide at molecular and single-cell resolution, which the authors say creates a need for fast and reliable ways to characterise spatial patterns. SpaGene is model-free, aiming at any spatial pattern rather than a prespecified family, and is reported to gain both power and speed over existing spatial-pattern methods.

Key results

  1. SuggestedThe model-free method detected spatial patterns with greater power and ran at larger scale than competing approaches.

    Based on simulations plus a range of spatially resolved transcriptomics datasets; the abstract names no competing method, no dataset, no power or false-positive figures and no runtimes. · Abstract

  2. ClaimedThe spatial expression patterns returned by the method recovered tissue structures that had not been observed directly in the data.

    Stated as an outcome of applying the method; the tissues, the structures recovered and any comparison with histological annotation are not described in the abstract. · Abstract

  3. ClaimedPairs of ligand and receptor genes were identified by the colocalization of their spatial patterns.

    Reported as a successful application of the same pattern comparison; no counts of recovered interactions and no validation against known interaction sets are given. · Abstract

Methods, data and reproducibility

A model-free method, SpaGene, that searches for spatial patterns in spatial omics measurements without fitting a parametric spatial model; it was assessed on simulated data and several spatially resolved transcriptomics datasets, and the same pattern comparison is used to test ligand-receptor colocalization.

  • simulated spatial data used to compare detection power
  • a variety of spatially resolved transcriptomics datasets
Methods & tools
SpaGene
Limitations (2)
  • The claimed gains in power and scalability come with no metric, no named competitor, no dataset identity and no runtime, so the comparison cannot be weighed from the text available. · noted in this summary
  • Neither the reconstructed tissue structures nor the ligand-receptor pairs are checked against independent annotation or a known interaction resource here. · noted in this summary

Read this if you screen spatially resolved transcriptomics for spatially variable genes or colocalized pairs and need a method that scales.

Citations
64

Abstract

The expeditious growth in spatial omics technologies enable profiling genome-wide molecular events at molecular and single-cell resolution, highlighting a need for fast and reliable methods to characterize spatial patterns. We developed SpaGene, a model-free method to discover any spatial patterns rapidly in large scale spatial omics studies. Analyzing simulation and a variety of spatial resolved transcriptomics data demonstrated that SpaGene is more powerful and scalable than existing methods. Spatial expression patterns by SpaGene reconstructed unobserved tissue structures. SpaGene also successfully discovered ligand-receptor interactions through their colocalization.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISAn '26Abdelaal '20Hu '22Cai '26Sun '20Hawinkel '26
63% semantic matchBioinformatics
Shaokun An, Mark S. Keller, Nils Gehlenborg+1 more·bioRxiv·Apr 30, 2026