High-definition spatial transcriptomics for in situ tissue profiling

Abstract

Spatial and molecular characteristics determine tissue function, yet high-resolution methods to capture both concurrently are lacking. Here, we developed high-definition spatial transcriptomics, which captures RNA from histological tissue sections on a dense, spatially barcoded bead array. Each experiment recovers several hundred thousand transcript-coupled spatial barcodes at 2-μm resolution, as demonstrated in mouse brain and primary breast cancer. This opens the way to high-resolution spatial analysis of cells and tissues.

Figure 1.. High-definition spatial transcriptomics (HDST).

(a) HDST workflow. (b) Labeling of morphological layers. HDST H&E image of a main olfactory bulb and matching HDST (x,y) barcodes annotated into 9 morphological areas. (c) Layer-specific DE patterns in HDST. Shown is the summed normalized expression of positively enriched signature genes significantly (FDR<0.1, two-sided t-test) associated with each layer as annotated in (b). (d, e) Nuclei segmentation and binning of HDST as in (b). (d) Segmented nuclei (“sn-like”) and lightly binned (“sc-like”) spatial barcodes assigned (black) to each of two cell types as in (b). (e) Enrichment of sn- and sc-like spatial barcodes with assigned cell types (columns) to morphological layers (rows) as in (b). −log₁₀(p-value) (one-sided Fisher’s exact test, Bonferroni adjusted, p-value<0.01) represents the color bar and grey tiles non-significant values.

Figure 2.. HDST distinguishes cell types and niches in a breast cancer resection.

(a) Labeling of morphological layers. HDST H&E image (left) of a breast cancer section and matching HDST (x,y) barcodes annotated into 13 morphological areas (right, color code). (b) Layer-specific spatial DE patterns in HDST. Summed normalized expression of positively enriched signature genes significantly (FDR<0.1, two-sided t-test) associated with each layer as in (a). (c) Cell type assignments by single nuclei as in (a). Two zoomed in regions (black and red squares) with H&E and color coded segments.