SM-Omics is an automated platform for high-throughput spatial multi-omics

Abstract

The spatial organization of cells and molecules plays a key role in tissue function in homeostasis and disease. Spatial transcriptomics has recently emerged as a key technique to capture and positionally barcode RNAs directly in tissues. Here, we advance the application of spatial transcriptomics at scale, by presenting Spatial Multi-Omics (SM-Omics) as a fully automated, high-throughput all-sequencing based platform for combined and spatially resolved transcriptomics and antibody-based protein measurements. SM-Omics uses DNA-barcoded antibodies, immunofluorescence or a combination thereof, to scale and combine spatial transcriptomics and spatial antibody-based multiplex protein detection. SM-Omics allows processing of up to 64 in situ spatial reactions or up to 96 sequencing-ready libraries, of high complexity, in a ~2 days process. We demonstrate SM-Omics in the mouse brain, spleen and colorectal cancer model, showing its broad utility as a high-throughput platform for spatial multi-omics.

Fig. 1. SM-Omics.

Overview of approach. SM-Omics approach combines automated imaging of H&E, IF stained or tissue sections stained with DNA-barcoded antibodies with high-throughput liquid handling to create spatially resolved RNA-Seq and/or antibody-seq libraries. The RNA-Seq protocol consists of three main steps. (I) in situ reactions on a ST slide that include tissue permeabilization, capture of mRNAs on the spatial array followed by a reverse transcription reaction in solution. The transcribed material is then collected and a two-step library preparation protocol (II–III) is run in standard 96-well plates.

Fig. 2. SM-Omics performance.

a, b Sensitivity of spatial gene expression measurements. Mean number of unique molecules detected (y axis) at different proportions of annotated reads (x axis) in a SM-Omics (blue, n = 3) and ST (red, n = 3) and b SM-Omics (blue, n = 3) and Visium (green, n = 3). Shaded areas: 95% confidence intervals. Colored line: mean of summarized library values (n = 3) per condition. c Performance of automated spatial library preparation reactions. Impact of ligation reaction times and adapter concentrations on quantitative concentrations (Cq) values for automated prepared libraries (n = 9). Cq values were measured at Fluorescent unit 10,000. Statistical significance (two-sided Wilcoxon’s rank-sum test) markings are displayed: 0.05 < p ≤ 1 (ns), 0.001 < p ≤ 0.01 (**), 0.0001 < p ≤ 0.001 (***). Center black line, median; color-coded box, interquartile range; error bars, 1.5x interquartile range; black dots; outliers. Individual reaction conditions are detailed in Methods. d, e Spatial gene expression. d Examples of SM-Omics spatial gene expression patterns (normalized expression shown in color scale) detected in each of the major histological regions in the MOB of an adult mouse brain and e corresponding in situ hybridization images from ABA (Image credit: Allen Institute for Brain Science, Methods) for the same genes as in d with illustrated and highlighted region annotation patterns. Annotated region abbreviations: GL (glomerular layer), GR (granular cell layer), MI (mitral layer), OPL (outer plexiform layer) and ONL (olfactory nerve layer).

Fig. 3. Spatial RNA-Seq and immunofluorescence highlights tissue specific expression patterns in the mouse brain cortex.

a Experimental setup. Tissue sections are placed on the spatial array (I), stained for nuclear and corresponding antigen targets, imaged for IF signals (II) and SM-Omics libraries created (III). Spatial gene and antibody expression data are processed and compared to the reference ABA atlas (IV). b Combined antibody IF and spatial transcriptomics in situ measurements. ABA in situ hybridization reference image (left) with NeuN staining neuronal nuclei with marked isocortex area (rectangle). Mouse brain isocortex tissue (n = 3) stained for DAPI (middle; cyan) and NeuN IF (middle; purple) and corresponding fluorescent gene activity cDNA footprint (right; white). Scale bar; 200 µm. c, d Performance of combined antibody IF and spatial transcriptomics measurements. c NeuN immunofluorescence (stained tissue section, left; scale bar 800 µm; and y axis; mean scaled signal, right) and mRNA in situ measurements (x axis, scaled normalized expression, right) per tissue section (n = 5, Methods) in each of seven regions (color code) in SM-Omics. Black line: linear regression with respective standard deviations (gray lines). d Antibody IF signals (normalized and scaled expression shown in color scale, NeuN IF; left) and mRNA expression (normalized and scaled expression shown in color scale, NeuN mRNA; right) aggregated in SM-Omics-like spots. White dashed lines: hypothalamus region. Annotated region abbreviations: CTXsp (cortical subplate), FIB (fiber tracts), HY (hypothalamus), HIP (hippocampal formation), ISOCTX (isocortex), PIR (piriform areas) and TH (thalamus).

Fig. 4. Spatial RNA-Seq and protein profiling with DNA-barcoded antibodies in mouse splenic tissue.

a Experimental setup. SM-Omics approach combines automated imaging of IF antibody stained tissue sections, tagging antigens spatially in situ using DNA-barcoded antibodies and capturing mRNA on a spatially barcoded poly(d)T array. Frozen tissue sections are placed on a SM-Omics array, tissues stained with both IF and DNA-barcoded antibodies, imaged and in situ copying reactions performed and at the same time as cDNA is made (I). Then, both the antibody tags and cDNAs are used in the library preparation reactions and sequenced (II). Finally, spatial IF, antibody tag and gene expression patterns can be evaluated (III). b Performance of combined antibody IF and DNA-barcoded antibody signal measurements. Splenic tissue illustration of red and white pulp structures (top) followed by spatial expression profiles of sequenced antibody tags (middle; normalized expression shown in color scale) as well as IF images (bottom) in splenic tissue for F4/80 staining red pulp macrophages and IgD staining marginal zone B cells in the white pulp (n = 7). Scale bar (bottom) denotes 300 µm. c Performance of combined DNA-barcoded antibody signal and spatial transcriptomics measurements. Spatial expression profiles (normalized expression shown in color scale) for a 6-plex SM-Omics reaction with F4/80, IgD, Cd163, Cd38, Cd4 and Cd8a DNA-barcoded antibody-based expression in the top panel (tags) and respective gene expression shown in the bottom panel (mRNA).