OT-knn reconstructs each spatial spot from its k nearest neighbors and uses those neighborhood-aware representations inside an optimal transport framework to compute probabilistic correspondences across slices. The method improves alignment when expression is sparse, tissue geometry differs, donors vary, or platforms change.
Those neighborhood-aware representations feed an optimal transport algorithm that returns probabilistic correspondences between spots across slices.
OT-knn is intended for aligning ST data across sections, individuals, stages, and technologies to facilitate comparative spatial analyses.
# What OT-knn does
# Why neighborhood information matters
# How the alignment works in practice
The probabilistic output allows partial matches (a spot can map with weights to multiple targets) which helps when slices differ in cell composition or resolution.
# Evaluation and datasets used The authors evaluated OT-knn on both simulated data with known ground-truth alignments and multiple real ST datasets across platforms and species:
Across these tasks OT-knn delivered accurate and robust alignments, notably in scenarios with spatial deformation, donor heterogeneity, developmental variation, and cross-platform differences.
# Practical strengths and use cases OT-knn targets comparative ST analyses where aligning spatial measurements is a prerequisite. Use cases include:
# Output and interpretation The method yields probabilistic correspondences rather than one-to-one hard matches. That representation is useful for downstream analyses that expect uncertainty (for example, transferring annotations with confidence weights, aggregating matched regions, or smoothing cross-sample comparisons).
# What to expect when applying OT-knn
# Where data and resources are available The paper lists locations for several datasets used in evaluation, including GEO accessions and public ST data repositories for Visium, MERFISH, Stereo-seq, and seqFISH resources. These datasets can be used for method reproduction and benchmarking.
# Bottom line OT-knn is a neighborhood-aware alignment method that integrates local spatial context with optimal transport to produce probabilistic cross-slice correspondences. It is designed to improve robustness when single-spot expression is noisy and when samples differ in shape, biology, or technology.
by Zi Huai Huang, Ziyang Xu, Pingzhao Hu Spatial transcriptomics (ST) enhances single-cell RNA sequencing by revealing transcript distribution, offering critical insights into heterogeneous diseases such as breast cancer. However, the high cost and lengthy processes of generating high-quality ST data limit clinical app
by Jinxia Wang, Yuying Huo, Rui Zhao, Yan Pan, Jianqiang Wu, Han Wang, Xiangyu Li Recent advances in spatial multi-omics technologies have opened new avenues for characterizing tissue architecture and function in situ, by simultaneously providing multimodal and complementary information—such as spatially resolved trans
by Nagomi Kurogi, Koki Shimbara, Tatsuya Koreeda, Koki Tsuyuzaki Spatial transcriptomics (ST) enables genome-wide gene expression profiling while retaining spatial context within tissue sections. Since the foundational work by Ståhl et al . in 2016, the field has expanded rapidly, with diverse platforms now spanning se
ResolVI improves spatial RNA sequencing analysis by correcting RNA assignment errors and technical variation, helping researchers distinguish cell states and detect subtle gene expression differences...
Open the app view to save this story, compare related coverage, and continue from the same source.