MERFISH+, a large-scale, multi-omics spatial technology resolves the transcriptomic holograms of the 3D human developing heart
Data files
Dec 18, 2025 version files 68.90 GB
-
Fig1_FirstImaging.h5ad
324.46 MB
-
Fig1_ReImage.h5ad
394.57 MB
-
Fig2_1835genes.h5ad
2.47 GB
-
Fig3_multimodal_v2.h5ad
2.61 GB
-
Fig4to6_3DHeart.h5ad
10.41 GB
-
Fig5_Pulmonary_valve_vessel.h5ad
259.05 MB
-
Fig7_3DHeart_imputed.h5ad
51.78 GB
-
FigS2_Codebook1.h5ad
325.82 MB
-
FigS2_Codebook2.h5ad
326.77 MB
-
README.md
5.27 KB
Abstract
Hybridization-based spatial transcriptomics technologies have advanced our ability to map cellular and subcellular organization in complex tissues. However, existing methods remain constrained in gene coverage, multimodal compatibility, and scalability. Here, we present MERFISH+, an enhanced version of Multiplexed Error-Robust Fluorescence in Situ Hybridization (MERFISH), which integrates covalent probe anchoring in protective hydrogels with automated high-throughput microfluidics and microscopy. This optimized design supports >100 hybridization cycles and centimeter-scale 3D serial imaging. MERFISH+ quantified over 1,800 genes and resolved the 3D organization of chromatin loci and epigenomic marks in developing human hearts. To unify these molecular features, we developed a generative integration framework for spatial multimodal data (Spateo-VI), generating a comprehensive 3D spatial multi-omic atlas encompassing ~3.1 million cells across 34 cellular populations. This 3D atlas provides a holistic view of an entire human organ enabling the characterization of 3D cellular neighborhoods and transcriptional variation within substructures such as valves and major artery branches. Thus, MERFISH+ offers a robust, large-scale platform for spatial multi-omics that enables high resolution mapping of gene expression at subcellular resolution and the characterization of cellular organization within 3D organs.
Dataset DOI: 10.5061/dryad.k98sf7mkx
Description of the data and file structure
Gene expression data measured using MERFISH in the developing human heart. Each distinct experiment is contained in a separate AnnData file, a standard format for storing single-cell gene expression data. Each AnnData file contains a ".X" matrix with the normalized expression values as well as a ".layers['counts']" matrix with the raw un-normalized expression values. Additionally, spatial coordinates of each cell are in the ".obsm['X_spatial']" matrix. Additional file-specific contents are described below.
Files and variables
File: Fig1_FirstImaging.h5ad
Description: Imaging of 238 genes using MERFISH in a single 2D heart section. AnnData object containing the following information for each cell: The number of genes ('n_genes') and transcripts ('total_counts') detected, and the cell type identity of each cell ('celltype3').
File: Fig1_ReImage.h5ad
Description: Re-imaging of 238 genes using MERFISH in a single 2D heart section. AnnData object containing the following information for each cell: The number of genes ('n_genes') and transcripts ('total_counts') detected, and the cell type identity of each cell ('celltype3').
File: FigS2_Codebook1.h5ad
Description: Imaging of 251 genes using MERFISH in a 2D heart section. AnnData object containing the following information for each cell: The number of genes ('n_genes_by_counts') and transcripts ('total_counts') detected, and cell type annotation ('celltype').
File: FigS2_Codebook2.h5ad
Description: Imaging of 251 genes using MERFISH in a 2D heart section, with a re-organized codebook. AnnData object containing the following information for each cell: The number of genes ('n_genes_by_counts') and transcripts ('total_counts') detected, and cell type annotation ('celltype').
File: Fig2_1835genes.h5ad
Description: Imaging of 1835 genes using MERFISH in a 2D heart section. AnnData object containing the following information for each cell: The number of genes ('n_genes_by_counts') and transcripts ('total_counts') detected, and cell type annotation ('celltype').
File: Fig3_multimodal_v2.h5ad
Description: Imaging of 251 genes using MERFISH and 6 proteins with immunostaining in a 2D heart section. AnnData object containing the following information for each cell: The number of genes ('n_genes_by_counts') and transcripts ('total_counts') detected, cell type annotation ('celltype'), and protein quantification measured by immunofluorescence ('H3K4me3', 'PolSer2', 'H3K27ac', 'Pol2Se5', 'H3K9me3', 'H3K27me3'). Chromatin tracing data is stored in the .obsm['X_h_score'] matrix, which is an Nx4x50x5 matrix where N is the number of cells, 4 is the maximum number of chromosomes per cell, 50 is each of the chromatin regions imaged, and 5 are the 5 properties measured for each region: The z, x, and y coordinates of the region, the brightness, and the score (see methods).
File: Fig4to6_3DHeart.h5ad
Description: Imaging of 238 genes using MERFISH in multiple 2D sections from the same heart, which was then reconstructed into its 3D conformation. AnnData object containing the following information for each cell: The number of genes ('n_genes_by_counts') and transcripts ('total_counts') detected, the high-level cell class ('class'), cell type annotation ('celltype'), spatially defined sub-type ('celltype-extra'). The MYH11+ cells assigned to coronary artery vessels are annotated in the 'vessel_name' column. The 'RCA/LCX/LAD/LCA_community' columns indicate whether a cell is part of the specified blood vessel or surrounding cells. The 'celltype_extra' column contains additional sub-clusters for the valves and blood vessels which were defined spatially. The cell coordinates contained in the .obsm["X_spatial"] are 2D coordinates of each section, while the .obsm["X_spateo"] matrix are the 3D reconstructed coordinates.
File: Fig5_Pulmonary_valve_vessel.h5ad
Description: Subset of 3D heart including the cells consisting of the pulmonary artery, pulmonary valve, and outflow tract. AnnData object containing the following information for each cell: The number of genes ('n_genes_by_counts') and transcripts ('total_counts') detected, the high-level cell class ('class'), cell type annotation ('celltype'), spatially defined sub-type ('celltype-extra'), endothelial subcluster ('Endo_subcluster'), and VIC subcluster ('VIC_subcluster')
File: Fig7_3DHeart_imputed.h5ad
Description: Imputed expression of 1825 genes and 5 antibodies in a reconstructed 3D heart. AnnData object containing 1,825 genes and 5 protein abundances in the .X matrix imputed computationally in the same cells as Fig4to6_3DHeart.h5ad. Cell names match those in Fig4to6_3DHeart.h5ad and can be linked to the additional information stored there.
Code/software
These files are in anndata format, which can be loaded in python using the Scanpy package:
!pip install scanpy
import scanpy as sc
adata = sc.read("Fig1_FirstImaging.h5ad")