Spatial Multi-omics: Combining RNA, ATAC, and Protein in the Same Tissue Section

By Lociven · SpatiaBio · August 2026

Spatial transcriptomics tells you where genes are expressed. Spatial multi-omics adds a second layer — chromatin accessibility, protein, or DNA — measured in the same tissue section. The result is a much richer picture of cell state that neither modality alone can give you.

---

Why spatial multi-omics

Gene expression tells you what a cell is doing right now. Chromatin accessibility tells you what it could do — which regulatory regions are open and which transcription factors are active. Combining both in the same spatial context reveals regulatory programs that are spatially organized: not just "macrophage" but "macrophage in an open chromatin state consistent with inflammatory activation, in the tumor margin."

Protein (CITE-seq spatial / CosMx protein) adds surface markers that are still the clinical standard for cell identity — CD3, CD8, PD-L1 — with precise spatial coordinates.

---

Current platforms

Platform	Modalities	Resolution	Status
10x Visium HD + CytAssist	RNA + H&E image	2 µm	Commercial
10x Multiome Spatial (ATAC+RNA)	RNA + ATAC	Spot-level	Beta / early access
DBiT-seq	RNA + protein (CITE)	10–50 µm	Academic protocol
CosMx (NanoString)	RNA + protein (same slide)	Single cell	Commercial
MIBI-TOF / CODEX	Protein only (40–60 markers)	Single cell	Commercial

---

What multi-omics buys you: three different measurements indexed by the same spatial coordinates, so a question that spans modalities — does open chromatin here predict expression here — becomes answerable. The caveat in the platform table still applies: most fully-integrated options are early access or academic protocols today.

Spatial ATAC + RNA: the workflow

The most analytically powerful combination is simultaneous spatial ATAC-seq and RNA-seq on the same section. 10x Genomics has a spatial Multiome protocol in early access. The analysis workflow in Squidpy/Muon follows the same pattern as single-cell Multiome — with spatial coordinates added.

import muon as mu
import squidpy as sq
import scanpy as sc

# Load the paired RNA + ATAC AnnData objects
rna = sc.read_h5ad("spatial_rna.h5ad")
atac = sc.read_h5ad("spatial_atac.h5ad")

# Combine into MuData — spots are shared
mdata = mu.MuData({"rna": rna, "atac": atac})

# Standard RNA preprocessing
sc.pp.normalize_total(mdata["rna"])
sc.pp.log1p(mdata["rna"])
sc.pp.highly_variable_genes(mdata["rna"], n_top_genes=3000)

# ATAC: TF-IDF + LSI
mu.atac.pp.tfidf(mdata["atac"])
mu.atac.tl.lsi(mdata["atac"])

# Multi-modal embedding (WNN)
mu.pp.neighbors(mdata, key_added="wnn")
mu.tl.umap(mdata, neighbors_key="wnn")
mu.tl.leiden(mdata, neighbors_key="wnn", key_added="wnn_leiden")

---

Linking open chromatin to spatial gene expression

The key analysis in spatial ATAC+RNA is linking regulatory elements to their target genes — and mapping those links spatially. Peak-to-gene correlations tell you which ATAC peaks are correlated with nearby gene expression across spatial locations.

# Peak-to-gene correlation across spatial spots
mu.atac.tl.rank_genes_groups_df(mdata, groupby="wnn_leiden", mod="atac")

# Motif enrichment in spatially variable peaks
mu.atac.tl.motifs(mdata["atac"])

# Visualize chromatin accessibility overlaid on tissue
sq.pl.spatial_scatter(
    mdata["rna"],
    color="wnn_leiden",
    title="WNN clusters (RNA + ATAC)"
)

---

Practical advice

Start with RNA only. Spatial multi-omics protocols are more technically demanding and more expensive. Do a Visium run first to understand the biology, then add ATAC or protein if the question requires it.

H&E is already a modality. Every Visium slide comes with an H&E image. Integrating the histology image with gene expression using tools like Tangram or BayesSpace is spatial multi-omics in practice — and it's free with your existing data.

Data integration is the bottleneck. The biology is interesting; the analysis is hard. WNN, MOFA+, and totalVI are the current standards for integrating spatial multi-modal data, but none of them are one-click.

---

From the NeoantigenLab sister blog

Profiling the tumor immune microenvironment?

NeoantigenLab covers neoantigen biology, TIL therapy, checkpoint inhibitors, and cancer immunotherapy — the biology behind your spatial data.

Visit NeoantigenLab →

Tags: spatial multi-omics, spatial ATAC-seq, Visium, Multiome, CosMx, CITE-seq, Muon, WNN, chromatin accessibility, tumor microenvironment, Squidpy