Bayesian Nonparametric Clustering with Feature Selection for Spatially Resolved Transcriptomics Data

TL;DR

This paper introduces BNPSpace, a Bayesian nonparametric framework that effectively clusters spatial transcriptomics data by modeling count data directly, incorporating spatial information, and identifying key discriminating genes, overcoming limitations of existing methods.

Contribution

BNPSpace is a novel Bayesian nonparametric method that models SRT count data directly and integrates spatial information for improved clustering and gene selection.

Findings

BNPSpace accurately identifies homogeneous spatial domains.

It effectively handles high-dimensional, zero-inflated, and heterogeneous data.

The method improves downstream analysis by reducing information loss.

Abstract

The advent of next-generation sequencing-based spatially resolved transcriptomics (SRT) techniques has reshaped genomic studies by enabling high-throughput gene expression profiling while preserving spatial and morphological context. Nevertheless, there are inherent challenges associated with these new high-dimensional spatial data, such as zero-inflation, over-dispersion, and heterogeneity. These challenges pose obstacles to effective clustering, which is a fundamental problem in SRT data analysis. Current computational approaches often rely on heuristic data preprocessing and arbitrary cluster number prespecification, leading to considerable information loss and consequently, suboptimal downstream analysis. In response to these challenges, we introduce BNPSpace, a novel Bayesian nonparametric spatial clustering framework that directly models SRT count data. BNPSpace facilitates the partitioning of the whole spatial domain, which is characterized by substantial heterogeneity, into homogeneous spatial domains with similar molecular characteristics while identifying a parsimonious set of discriminating genes among different spatial domains. Moreover, BNPSpace incorporates spatial information through a Markov random field prior model, encouraging a smooth and biologically meaningful partition pattern.