Search and Match across Spatial Omics Samples at Single-cell Resolution

Abstract

Spatial omics technologies characterize tissue molecular properties with spatial information, but integrating and comparing spatial data across different technologies and modalities is challenging. A comparative analysis tool that can search, match, and visualize both similarities and differences of molecular features in space across multiple samples is lacking. To address this, we introduce CAST (Cross-sample Alignment of SpaTial omics), a deep graph neural network (GNN)-based method enabling spatial-to-spatial searching and matching at the single-cell level. CAST aligns tissues based on intrinsic similarities of spatial molecular features and reconstructs spatially resolved single-cell multi-omic profiles. CAST enables spatially resolved differential analysis (Delta Analysis) to pinpoint and visualize disease-associated molecular pathways and cell-cell interactions, and single-cell relative translational efficiency (scRTE) profiling to reveal variations in translational control across cell types and regions. CAST serves as an integrative framework for seamless single-cell spatial data searching and matching across technologies, modalities, and disease conditions, analogous to BLAST in sequence alignment.

VISTA Uncovers Missing Gene Expression and Spatial-induced Information for Spatial Transcriptomic Data Analysis

“…To simulate the perturbation case, we decoded the gene expression values using latent space from spatial data and decoder part for scRNA-seq data. Firstly, we explored the ability of VISTA to transfer the gene expression levels with diseased information as the special case of perturbation from reference scRNA-seq [61] to target spatial data [62]. We treated the scRNA-seq dataset and the spatial dataset as paired datasets.…”

TissueMosaic enables cross-sample differential analysis of spatial transcriptomics datasets through self-supervised representation learning

“…Previous work has applied SSL to spatial transcriptomics using graph-based architectures. 25 We reasoned that representing ST data as images provides a more expressive encoding of spatial relationships compared to graphs. The latter only captures local connectivity information but cannot capture higher order spatial features, such as disorder.…”

Joint imputation and deconvolution of gene expression across spatial transcriptomics platforms

“…We further assume that there exists an alignment between ( A X , S X ) and ( A V , S V ); i.e. there is a spot-spot correspondence between two slices as computed from any method that aligns two SRT datasets [50, 26, 40, 20, 11, 24]. Given the spot-spot correspondence, the coordinates S X and S V can be transformed to represent locations in a shared coordinate system between the SRT datasets.…”