Transcriptional output, cell types densities and normalization in spatial transcriptomics

Abstract

Spatial transcriptomics measures mRNA at hundreds of 100 micrometer-diameter spots evenly spread across 6.5×6.9 mm2 histological slices. Gene expression within each spot is commonly normalized by total read counts. However we show that the transcriptional output of individual spots reflects the number of cells they contain, hence total read counts per spot reflect relevant biology. Although per-spot read-count normalization reveals important enrichment trends, it may heavily distort cell-type-related absolute local expression and conceal important biological information.

Characterizing spatial gene expression heterogeneity in spatially resolved single-cell transcriptomic data with nonuniform cellular densities

“…Likewise, we find MERINGUE to be complementary to expression-based clustering analysis in order to identify additional aspects of spatial heterogeneity within cell clusters or shared spatial gradients across cell clusters. In addition, in analyzing spatially resolved single cell gene expression datasets obtained from different technologies, MERINGUE may also be applied in combination with different normalization and error model schemes such as cell volume-based normalization for imaging data (Moffitt et al 2018), cell density normalization for ST data (Saiselet et al 2020). Furthermore, for zero-inflated transcriptomics measurements, additional drop-out error modeling or imputation of drop-outs may be applied prior to MERINGUE analysis (Kharchenko et al 2014;Hou et al 2020).…”

Single-Cell Transcriptome Analysis in Plants: Advances and Challenges

“…This method uses a tissue slide with thousands of barcoded spots (total 1,007 spots per capture area of 6.2 3 6.6 mm 2 and each spot is 100 mm in diameter with 200-mm center-to-center distance between spots), each containing millions of capture oligonucleotides with spatial barcodes unique to that spot, to profile the transcriptomes of multiple adjacent cells of a tissue section by RNA-seq (Figure 2C). Recently, the Visium Spatial Gene Expression Solution (10X Genomics, Pleasanton, USA) has been made available to provide a step increase in capability, such as with more spots (total 4,992 spots per capture area of 6.5 3 6.5 mm 2 and each spot is 55 mm in diameter with 100-mm center-to-center distance between spots) and, as a result, fewer cells per spot (10-200 cells depending on the tissue context; Saiselet et al, 2020). The resolution of microarray-based spatial transcriptomics can be further boosted by generating scRNA-seq data from the same sample…”

Hypoxia and the phenomenon of immune exclusion

“…5). Each area resolves the transcriptome of 10-200 cells depending on the tissue context [311]. ST has been used to study breast cancer, melanoma, prostate cancer, adult human hearth tissue, pancreatic ductal adenocarcinoma, mouse, human and mouse spinal cord tissue and mouse olfactory bulb [312][313][314][315][316][317][318][319].…”