Figure 6 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Integrative spatial analysis of multimodal spatial data. A. Spatial analysis of T-cell receptor (TCR) data in brain metastasis: TCR counts per spot; spots with more than five TCR UMI counts (overlayed on the original H&E image); count ratio for TCR with unknown cognate antigen (blue) vs. viral ones (red). B. Spatial analysis of tertiary lymphoid structures (TLS) in renal cancer: pathology annotation on the presence of TLS (in yellow); smoothed spatial plot of the expression-based TLS score; smoothed spatial plot of B-cell fractions estimated with quanTIseq. C. Scaling of deconvolution results by total cell counts per spot in a breast cancer slide. Left to right: total cell counts in the whole
Source publication
Investigating tissue architecture is key to understanding tissue function in health and disease. While spatial omics technologies enable the study of cell transcriptomes within their native context, they often lack single-cell resolution. Deconvolution methods can computationally infer tissue composition from spatial transcriptomics data, but diffe...
Contexts in source publication
Context 1
... analyzed this TCR data using MiXCR [42] and Scirpy [43] to quantify spatially-resolved TCR clonotypes and annotate them based on their antigen specificity according to the VDJdb database [44] (Table S6, details in Methods). After integrating spot-based TCR data, we used spacedeconv to visualize the spatial distribution of the TCR unique molecular identifier (UMI) counts and identify spots with more than 5 TCR UMI ( Figure 6A). The distribution of the latter nicely recapitulates the localization of T cells reconstructed via deconvolution analysis ( Figure 3A). ...
Context 2
... used spacedeconv (plot_comparison function) to visualize the prevalence of unknown TCR (vs. viral/bystander ones), revealing that they were confined at the margin of the tumor rather than being infiltrated in the tumor core ( Figure 6A). ...
Context 3
... TLS region was enriched for B cells, CAFs, macrophages, and T cells, and presented a preferential activity of MAPK, Trail, and EGFR pathways, as well as of TF involved in T helper-cell differentiation and polarization (HLX and STAT6). We further corroborated pathology annotations by visualizing a TLS score derived from the spot-based transcriptomic data (gene_set_score function), as well as the B cell fractions inferred via spatial deconvolution ( Figure 6B). ...
Context 4
... adopted a similar approach for the analysis of the 10x Visium breast cancer slide: we quantified total cell counts per spot from the associated hematoxylin and eosin (H&E) slide (Table S8) and used them to scale the deconvolved fractions per spot via the scale_cell_counts function (details in Methods). Cell counts varied greatly across the slide (up to 36 cells per spot) and revealed a scarcely populated region in the top-right corner of this slide ( Figure 6C). Most importantly, the scaling of the deconvolved cell fractions to absolute cell counts resulted in the elimination of some small areas enriched in luminal B cancer cells and myCAFs, which were likely artifacts, i.e. inflated cell fractions due to to a low overall cell content in the corresponding spots. ...
Similar publications
This article reviews the transformative impact of single-cell whole genome sequencing (scWGS) on cancer biology research. ScWGS has revolutionized our understanding of tumor heterogeneity, clonal evolution, and the tumor microenvironment by enabling high-resolution analysis of individual cells' genomic, transcriptomic, and epigenomic profiles. The...
Citations
Cell-cell interaction (CCI) networks are key to understanding disease progression and treatment response. However, existing methods for inferring these networks often aggregate data across patients or focus on cell-type level interactions, providing a generalized overview but overlooking patient heterogeneity and local network structures. To address this, we introduce “random cell-cell interaction generator” (RaCInG), a model based on random graphs to derive personalized networks leveraging prior knowledge on ligand-receptor interactions and bulk RNA sequencing data. We applied RaCInG to 8,683 cancer patients to extract 643 network features related to the tumor microenvironment and unveiled associations with immune response and subtypes, enabling prediction and explanation of immunotherapy responses. RaCInG demonstrated robustness and showed consistencies with state-of-the-art methods. Our findings highlight RaCInG’s potential to elucidate patient-specific network dynamics, offering insights into cancer biology and treatment responses. RaCInG is poised to advance our understanding of complex CCI s in cancer and other biomedical domains.
