SMART: spatial transcriptomics deconvolution using marker-gene-assisted topic model

Genome Biol. 2024 Dec 2;25(1):304. doi: 10.1186/s13059-024-03441-1.

Abstract

While spatial transcriptomics offer valuable insights into gene expression patterns within the spatial context of tissue, many technologies do not have a single-cell resolution. Here, we present SMART, a marker gene-assisted deconvolution method that simultaneously infers the cell type-specific gene expression profile and the cellular composition at each spot. Using multiple datasets, we show that SMART outperforms the existing methods in realistic settings. It also provides a two-stage approach to enhance its performance on cell subtypes. The covariate model of SMART enables the identification of cell type-specific differentially expressed genes across conditions, elucidating biological changes at a single-cell-type resolution.

Keywords: Deconvolution; Semi-supervised; Spatial transcriptomics; Topic models.

MeSH terms

  • Gene Expression Profiling* / methods
  • Genetic Markers
  • Humans
  • Single-Cell Analysis / methods
  • Transcriptome*

Substances

  • Genetic Markers