Benchmarking spatial and single-cell transcriptomics integration methods for transcript distribution prediction and cell type deconvolution

Integrative analysis of spatial transcriptome with single-cell transcriptome and single-cell epigenome in mouse lungs after immunization

“…To better interpret spatial transcriptome data, it is vital to determine the proportions of different cell types within each spot. Using our finely annotated scRNA-seq data as a reference, we carried out deconvolution for each spot in the four slices using the robust cell-type decomposition (RCTD) method ( Cable et al., 2022 ), one of the top-performing methods for cell-type deconvolution ( Li et al., 2022 ). The deconvolution results for slice A3 were shown as proportions of 12 cell types across slice A3 ( Figure S2 A).…”

Tumoroscope: a probabilistic model for mapping cancer clones in tumor tissues

“…Furthermore, studies combining H&E, WES, and ST for large cohorts of patients, could explore the dependencies between patient clinical features and the spatial patterns of clones found using Tumoroscope. Combined with cell-type deconvolution approaches for ST data in the tissue surrounding the tumors [39][40][41], our framework has the potential to bring unprecedented insights into the interactions of specific cancer clones, their phenotypes, and the surrounding microenvironment. In summary, Tumoroscope opens up a new avenue in cancer research with broad applications for a basic understanding of the disease and its clinical applications.…”

Emerging artificial intelligence applications in Spatial Transcriptomics analysis

“…Furthermore, Tangram could visualize the chromatin accessibility information in space by analyzing SHARE-seq [63] data containing matched RNA and chromatin accessibility information from single cells. External benchmark study [64] showed Tangram had decent deconvolution performance across diverse real and synthetic datasets and top performance in predicting spatial distribution of gene expression compared to Seurat [65] , Cell2location [66] , SpatialDWLS [67] , RCTD [68] , Stereoscope [69] , DestVI [60] , STRIDE [70] , SPOTLight [71] , and DSTG [72] .…”

“…The authors also applied DestVI to in-house ST datasets of human lymph node sections and syngeneic mouse tumor tissues profiled by Visium [2] ; DestVI delineated transcriptional states of the cell types and identified spatially resolved multicellular immune responses and hypoxic population of macrophages in the tumor core, respectively. The deconvolution functionality of DestVI was further tested in two recent external benchmark studies that focus on ST deconvolution [64] , [75] , which showed that DestVI had decent performance but was not robust enough across different tissue types.…”

“…Real data analysis on osmFISH [89] , merFISH [90] , and STARmap [7] datasets showed that the predicted spatial gene expression by stPlus helped to reduce technical noise and achieved improved cell type clustering compared to other methods such as SpaGE [91] , Seurat [65] , Liger [92] , and gimVI [93] . However, an independent benchmark study [64] showed that stPlus had low accuracy in predicting spatial gene expression compared to Tangram [59] , gimVI [93] , SpaGE [91] , Seurat [65] , SpaOTsc [94] , novoSpaRc [95] , and LIGER [92] . Therefore, the overall performance of stPlus requires further examination.…”