1 of 10

VITESSCE

Gehlenborg Lab,

Harvard Medical School

2 of 10

Background

Multi-modal single-cell resolution datasets

We lack tools for integrative visualization of multi-modal datasets

Transcriptomics

Epigenomics,

chromatin accessibility

Imaging, microscopy

Cell segmentations

Dimensionality reductions

Clusters, cell types

3 of 10

4 of 10

Goals for integrative visualization with Vitessce

  • Interactive visualization of multi-modal (spatial) single-cell data
  • Coordination of common entities across visualizations
  • Lack of specialized server infrastructure
  • In progress: support for data analysis environments
    • Python (Jupyter / JupyterLab widget)
    • R (Shiny / htmlwidget)
  • Support for single-cell data structures
    • AnnData
    • Seurat

5 of 10

Supported environments

6 of 10

Demo: co-detection by indexing (CODEX)

https://portal.hubmapconsortium.org/browse/dataset/2f8d43199ca3167e991b320038024b77

7 of 10

Demo: scATAC-seq

http://beta.vitessce.io/index.html?dataset=sn-atac-seq-hubmap-2020

8 of 10

Demo: cell type annotations

Azimuth (Hao et al., bioRxiv 2020)

https://deploy-preview-3--trusting-perlman-9a8667.netlify.app/references/

9 of 10

Technical details

  • Genomic visualization: HiGlass (Kerpedjiev et al. Genome Biology 2018)
    • Genome browser tracks
  • Bioimaging visualization: Viv (Manz et al. OSF Preprints 2020)
    • Multi-scale
    • Multi-channel
    • OME-TIFF and OME-Zarr formats supported
  • Statistical visualization: Vega-Lite (Satyanarayan et al. IEEE TVCG 2017)
    • Histograms
    • Violin plots

10 of 10

Learn more at http://vitessce.io

Acknowledgements

Gehlenborg Lab

  • Nils Gehlenborg
  • Chuck McCallum
  • Ilan Gold
  • Trevor Manz
  • Tos Chan
  • Jennifer Marx