Pratheepa Jeganathan: A Post-Constrained Clustering Framework for Detecting Repeated Spatial Patterns

A Post-Constrained Clustering Framework for Detecting Repeated Spatial Patterns

Identifying spatially contiguous clusters and repeated spatial patterns, characterized by spatially distant regions that share similar underlying distributions, is one of the challenges in modern spatial statistics, particularly in spatial omics applications. We propose a post-clustering multiple hypothesis testing procedure (repStat) to detect repeated spatial patterns from constrained clustering results. The repStat includes a structure-preserving block permutation method to approximate the distribution of the test statistic. We show that the maximum mean discrepancy statistic is asymptotically consistent under second-order stationarity and spatial mixing conditions present in the constrained clustering results. Through simulation studies and real-world data from spatial proteomics, we demonstrate the robustness of repStat to varying spatial dependence, cluster sizes, shapes, multivariate dimensionality, and data types to identify repeated spatial patterns in geospatial or lattice processes. This talk is based on the preprint at https://arxiv.org/pdf/2506.14103.

BIO: Pratheepa Jeganathan is an Assistant Professor in the Department of Mathematics and Statistics and an Associate Member of the School of Computational Science and Engineering at McMaster University. Prior to this, she completed a postdoctoral fellowship in the Department of Statistics at Stanford University. Her research focuses on statistical methodologies for dependence modeling with applications to molecular microbiology, spatial omics, sensor-based traffic data, and loss reserving. Her research is supported by the Natural Sciences and Engineering Research Council (NSERC) of Canada and Canadian Statistical Sciences Institute (CANSSI) Collaborative Research Team Project.