DOI

Welcome to the Habenula Atlas project! Here you will find all code used to analyze the data generated as part of our manuscript, which is currently preprinted here.

In this project we studied spatially resolved transcriptomics and single-nucleus multiomic (ATAC + RNA) data from the habenula of postmortem human brain samples from a total of 13 neurotypical controls. We generated data from multiple assays:

This work was a joint effort by the Keri Martinowich, Leonardo Collado-Torres, and Kristen Maynard teams at the Lieber Institute for Brain Development as well as Brion Maher’s group from JHBSPH’s Mental Health Department.

There are 2 GitHub repos associated with this study. They are:

  1. Hb_Multiome
  2. Habenula_Visium

Study Design

Multi-omic mapping of the human Hb. (A) A schematic showing the cutting scheme to allow for smFISH/RNAscope, VisiumHD, and snMultiome. The Hb was scored out to minimize thalamic contamination (N=12). (B) snMultiome schematic showing that the assay allows for profiling of both the transcriptome and epigenome from the same nuclei (n=10). (C) VisiumHD schematic showcasing the subcellular resolution that can be achieved using this spatial transcriptomic technique (n=5). (D) A visual representation of the biological questions this study sought to answer. (D.a.) Cross talk between the medial and lateral portions of the Hb. (D.b.) Transcription factors and which genes are impacted (D.c.) Which cell types are at highest risk in disease states. (D.d.) Locating inhibitory neurons within the Hb, particularly in the LHb. (D.e.) Cell-to-cell communication within the Hb. (D.f.) Cell type distribution in the Hb. (D.g.) Neuron-glia interactions in the Hb. (E) Brain slab illustration that depicts the landmark structures used during dissection (C - caudate, SP - segmented putamen, LGN - lateral geniculate nucleus, SN - substantial nigra, Th - thalamus) and the location of the LHb and MHb and an approximation of the brain blocks that were dissected out. Medial (M), dorsal (D), and anterior (A) directions are shown by the respective arrows. (F) A brain block that contained Hb, confirmed by RNAscope. (F.a.) Brain block with the Hb demarcated in a red dashed line. (F.b-e.) Single channel images of the probe panel used to locate the Hb, TAC3, GPR151, POU4F1, and MBP were used together to identify, generally the Hb (POU4F1) , the MHb (TAC3), LHb (GPR151), and the surrounding white matter (MBP). In (F.e.) the boundaries of the MHb and LHb are drawn. The insets for (E.b-e.) are zoom-ins of the body of the Hb. (G) VisiumHD of the same donor shown in E. (G.a.) The small image to the left shows the whole 6x6 mm array, with the dashed line showing the section that is shown at a larger scale to the right. The right image in (G.a.) is the eosin staining that was completed during VisiumHD with boundaries of the MHb and LHb that were also in (F.e.). (G.b-e.) VisiumHD expression for the same marker genes that were used to confirm the presence of Hb via RNAscope in (E.b-e.). (H) A UMAP of the gene expression/RNA data achieved from snMultiome. (I) A UMAP of the ATAC data from snMultiome. (J) A UMAP of the integrated RNA and ATAC clusters. (K) Broad marker genes for the clusters yielded, where the intensity of the color and size of the spot denote the expression of the respective gene.

Citing Our Work

Please cite our manuscript if you use data from this project.

Below is the citation in BibTeX format.

@article {Montgomery2026.09.30.755667,
    author = {Montgomery, Kelsey D. and Werner, Jonathan M. and Eagles, Nicholas J. and Liu, Chunyu and Cardinault, Cynthia S. and Bach, Svitlana V. and Chandra, Atharv and Zhang, Ruth and Maguire, Sarah E. and Divecha, Heena R. and Deep-Soboslay, Amy and Miller, Ryan A. and Huuki-Myers, Louise A. and Bharadwaj, Rahul A. and Kleinman, Joel E. and Hyde, Thomas M. and Maher, Brion S. and Collado-Torres, Leonardo and Maynard, Kristen R.},
    title = {Multi-omic spatial mapping of the human habenula defines the molecular and anatomical organization of medial and lateral subregions},
    elocation-id = {2026.09.30.755667},
    year = {2026},
    doi = {10.64898/2026.09.30.755667},
    publisher = {Cold Spring Harbor Laboratory},
    URL = {https://www.biorxiv.org/content/early/2026/10/02/2026.09.30.755667},
    eprint = {https://www.biorxiv.org/content/early/2026/10/02/2026.09.30.755667.full.pdf},
    journal = {bioRxiv}
}

Interactive Websites

All of these interactive websites are powered by open source software, namely:

We provide the following interactive websites, organized by dataset with software labeled by emojis:

Data Access

The Zenodo Archive for this project can be found here. Project data was also uploaded to the NeMO and can be found here (collection nemo:col-av55so6).

The major R objects for each analysis are also available through spatialLIBD::fetch_data() as of spatialLIBD version 1.25.5. We provide access to the following R objects, named by the type argument provided to spatialLIBD::fetch_data():

For example, let’s check out the Visium HD cell-level data:

## Check that you have a recent version of spatialLIBD installed
stopifnot(packageVersion("spatialLIBD") >= "1.25.5")

#   Retrieve the SpatialExperiment object
spe <- spatialLIBD::fetch_data(type = "habenula_atlas_HD_spe_cell")

#   This is a SpatialExperiment object
print(spe)
#> class: SpatialExperiment 
#> dim: 17708 287325 
#> metadata(0):
#> assays(2): counts logcounts
#> rownames(17708): ENSG00000187634 ENSG00000188976 ... ENSG00000198695 ENSG00000198727
#> rowData names(7): source type ... gene_type gene_search
#> colnames(287325): 1_H1-W369TJK_D1_9090 2_H1-W369TJK_D1_9090 ... 65739_H1-6FX4YN3_D1_9902
#>   65740_H1-6FX4YN3_D1_9902
#> colData names(18): key sample_id ... cell_type cell_type_colors
#> reducedDimNames(2): PCA UMAP
#> mainExpName: NULL
#> altExpNames(0):
#> spatialCoords names(2) : pxl_col_in_fullres pxl_row_in_fullres
#> imgData names(4): sample_id image_id data scaleFactor

#   Plot the cell types spatially
spatialLIBD::vis_clus(
        spe, sampleid = 'Br9090_1', clustervar = 'cell_type',
        is_stitched = TRUE, spatial = FALSE
    ) +
    #   Make points in legend more visible
    guides(fill = guide_legend(override.aes = list(size = 5)))

Contact

We value public questions, as they allow other users to learn from the answers. If you have any questions, please ask them at LieberInstitute/Hb_multiome/issues and refrain from emailing us. Thank you again for your interest in our work!

Internal

Both GitHub repositories are organized along the R/Bioconductor-powered Team Data Science group guidelines. They aim to follow the LieberInstitute/template_project structure.