The SpatialData framework: from raw data to interactive exploration (workshop introduction)

BIOINFO 2024, Gyeongju

LucaMarconato

Luca Marconato (Oliver Stegle lab)

LucaMarconato2

Workshop

notebooks

Documentation

notebooks

scverse/spatialdata

October 22nd, 2024

lmarconato

Joint work with Giovanni Palla, Kevin Yamauchi, Tim Treis, Wouter Vierdag and more

The SpatialData framework: from raw data to interactive exploration (workshop introduction)

BIOINFO 2024, Gyeongju

LucaMarconato

Luca Marconato (Oliver Stegle lab)

LucaMarconato2

Workshop

notebooks

Documentation

notebooks

scverse/spatialdata

October 22nd, 2024

workshop

docs

lmarconato

Joint work with Giovanni Palla, Kevin Yamauchi, Tim Treis, Wouter Vierdag and more

Spatial organizations are crucial to explaining biological mechanisms

Oral Pole

Drosophila developing embryos show spatial pattern of gene expression

Spatial organizations are crucial to explaining biological mechanisms

Oral Pole

Tumor microenvironments are organized in complex spatial structures

Adapted from Balkwill et al. (J. Cell Sci, 2018)

Spatial organizations are crucial to explaining biological mechanisms

Oral Pole

Spatial pattern of expression at the subcellular level

Spatial omics data could be, in principle, grouped into three classes

Introduction 1/

Annotated regions

Visium

Example 1/3: Visium 10x technology

55 µm

Visium capture

regions (“spots”)

Introduction 1/

Example 1/3: storage of a Visium dataset

Introduction 1/

Example 2/3: the smFISH and MERFISH techniques

Haimovich et al. (Bio-protocol, 2018)

Introduction 1/

Example 2/3: storage of a MERFISH dataset

Allen: VISp from SpaceTx SpaceJam data (Github)

Introduction 1/

Example 3/3: imaging mass cytometry (IMC) / Hyperion

Aboral Pole

Oral Pole

Adapted from Bodenmiller et al. (Cell Systems, 2018)

Simultaneous staining for:

• RNA
• Proteins

Introduction 1/

Example 3/3: storage of an IMC dataset

Aboral Pole

Jackson, Fisher et al. (Nature,2020)

Introduction 1/

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

• Non-goal: not an analysis library
• Interoperable infrastructure

• Simple read/write
• Flexible representation
• Object manipulation

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

• Simple read/write

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

• Simple read/write
• Flexible representation

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

• Simple read/write
• Flexible representation
• Object manipulation

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

• Non-goal: not an analysis library

• Simple read/write
• Flexible representation
• Object manipulation

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Resolution: 55µm

Transcriptome-wide

​

• Non-goal: not an analysis library
• Interoperable infrastructure

• Simple read/write
• Flexible representation
• Object manipulation

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Reading the data from disk:

Xenium

Visium

Visium HD

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

workshop

workshop

workshop

Data is represented with a modular combination of reusable elements

Microscopy images, …

Segmentation mask

Raster geometries

Data is represented with a modular combination of reusable elements

Microscopy images, …

Segmentation mask

Raster geometries

Vector geometries

Transcripts locations

Cells, ROIs,

…

Data is represented with a modular combination of reusable elements

Microscopy images, …

Segmentation mask

Raster geometries

Vector geometries

Transcripts locations

Cells, ROIs,

…

Gene expression, cell types, …

Annotations

Annotates

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

SpatialData unifies the representation of spatial omics across technologies

a.

Visium

Visium HD

Xenium

Preview Data: FFPE Human Lung Cancer with Xenium Multimodal Cell Segmentation

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Data: https://www.10xgenomics.com/products/xenium-in-situ/preview-dataset-human-breast

Data at different scales

workshop

Coordinate transformations enable alignment to common coordinate systems

​

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

docs

docs

workshop

Coordinate transformations enable alignment to common coordinate systems

​

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

Coordinate transformations enable alignment to common coordinate systems

​

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

Coordinate transformations enable alignment to common coordinate systems

​

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

docs

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

docs

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

docs

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

docs

workshop

Coordinate transformations enable alignment to common coordinate systems

​

docs

{“name" : “pixel-space",

"axes" : [

{"name": "j", "type": "space", "discrete": true },

{"name": "i", "type": "space", "discrete": true }]}

pixel space

docs

docs

workshop

Interactive visualization and annotation with napari-spatialdata

workshop

Marcela Toth

Rahul B. Shrestha

Interactive visualization and annotation with napari-spatialdata

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Save annotations*

workshop

*after user feedback we are revisiting the UX

Marcela Toth

Rahul B. Shrestha

Interactive visualization and annotation with napari-spatialdata

workshop

Marcela Toth

Rahul B. Shrestha

Interactive visualization and annotation with napari-spatialdata

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

workshop

Marcela Toth

Rahul B. Shrestha

Interactive visualization and annotation with napari-spatialdata

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Save annotations*

workshop

*after user feedback we are revisiting the UX

Marcela Toth

Rahul B. Shrestha

Pathology annotation with napari-spatialdata

Pathology annotation with napari-spatialdata

Interactive visualization and annotations for custom embeddings

Kasia Kedziora

Interactive visualization and annotations for custom embeddings

Kasia Kedziora

Interactive visualization and annotations for custom embeddings

Kasia Kedziora

Static, composable visualization with spatialdata-plot

workshop

workshop

Harald Vöringer

Sonja Stockhaus

Static, composable visualization with spatialdata-plot

workshop

Harald Vöringer

Sonja Stockhaus

Static, composable visualization with spatialdata-plot

workshop

workshop

Harald Vöringer

Sonja Stockhaus

The modular structure of SpatialData objects integrates into existing analysis workflows

workshop

workshop

The modular structure of SpatialData objects integrates into existing analysis workflows

The modular structure of SpatialData objects integrates into existing analysis workflows

workshop

The modular structure of SpatialData objects integrates into existing analysis workflows

workshop

workshop

The modular structure of SpatialData objects integrates into existing analysis workflows

workshop

workshop

Generalized, reusable operations are defined for SpatialData objects

https://spatialdata.scverse.org/en/stable/

Documentation:

e.g. rotate an image

e.g. crop the data

e.g. create image tiles around cells

e.g. count transcripts inside cells

docs

docs

docs

docs

Generalized, reusable operations are defined for SpatialData objects

https://spatialdata.scverse.org/en/stable/

Documentation:

e.g. rotate an image

docs

Generalized, reusable operations are defined for SpatialData objects

https://spatialdata.scverse.org/en/stable/

Documentation:

e.g. rotate an image

e.g. crop the data

docs

Generalized, reusable operations are defined for SpatialData objects

https://spatialdata.scverse.org/en/stable/

Documentation:

e.g. rotate an image

e.g. crop the data

e.g. create image tiles around cells

docs

docs

Generalized, reusable operations are defined for SpatialData objects

https://spatialdata.scverse.org/en/stable/

Documentation:

e.g. rotate an image

e.g. crop the data

e.g. create image tiles around cells

e.g. count transcripts inside cells

docs

docs

Generalized, reusable operations are defined for SpatialData objects

https://spatialdata.scverse.org/en/stable/

Documentation:

e.g. rotate an image

e.g. crop the data

e.g. create image tiles around cells

e.g. count transcripts inside cells

docs

docs

docs

docs

Aligning multiple modalities with a landmark-based workflow

Before alignment (raw data)

docs

Aligning multiple modalities with a landmark-based workflow

Before alignment (raw data)

Aligning multiple modalities with a landmark-based workflow

Before alignment (raw data)

Aligning multiple modalities with a landmark-based workflow

Before alignment (raw data)

Aligning multiple modalities with a landmark-based workflow

Before alignment (raw data)

docs

Spatial query APIs allow for cropping the data based on regions of interest

docs

workshop

Spatial query APIs allow for cropping the data based on regions of interest

Spatial query APIs allow for cropping the data based on regions of interest

Spatial query APIs allow for cropping the data based on regions of interest

docs

workshop

Aggregation example: computing cell types fractions within regions of interest

Xenium cell types

Manually annotated ROI

Immune infiltration

Cell type fraction

Collaborator: Elyas Heidari

Aggregation example: computing cell types fractions within regions of interest

Xenium cell types

Manually annotated ROI

Immune infiltration

Collaborator: Elyas Heidari

Aggregation example: computing cell types fractions within regions of interest

Xenium cell types

Manually annotated ROI

Immune infiltration

Cell type fraction

Collaborator: Elyas Heidari

Aggregation example: computing cell types fractions within regions of interest

Xenium cell types

Manually annotated ROI

Immune infiltration

Cell type fraction

Collaborator: Elyas Heidari

We provide a unified API covering all the aggregation operations

We provide a unified API covering all the aggregation operations

We provide a unified API covering all the aggregation operations

We provide a unified API covering all the aggregation operations

We provide a unified API covering all the aggregation operations

Visium HD bins are in a grid, allowing us to represent them as an image

a.

​

• Visium HD data has millions of bins
• Instead of storing them as vector data, images would be more performant

​

​

​

​

​

​

​

​

​

Visium HD

2µm, 8µm, 16µm bins

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Visium HD bins are in a grid, allowing us to represent them as an image

a.

​

• Visium HD data has millions of bins
• Instead of storing them as vector data, images would be more performant

​

​

​

​

​

​

​

​

​

Visium HD

2µm, 8µm, 16µm bins

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

> rasterization is needed

Visium HD bins are in a grid, allowing us to represent them as an image

a.

​

• Visium HD data has millions of bins
• Instead of storing them as vector data, images would be more performant

​

​

​

​

​

​

​

​

​

Visium HD

2µm, 8µm, 16µm bins

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

(x1, y1)

(x2, y2)

Rasterization

+

Bins as vector data

Bins as rasterized data

Gene 1

Gene 2

Gene 3

> rasterization is needed

1

1

2

2

3

4

4

4

5

Naive rasterization not possible. Solution: on-demand rasterization for binned data

Rasterization is already supported in SpatialData

​

​

​

​

​

​

​

​

Visium HD data is composed of small bins of 2x2 microns

Challenge:

About 6M bins and 30K genes per bin would represent 180GB as a dense image

> Solution: on-demand rasterization via rasterize_bins()

​

​

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Naive rasterization not possible. Solution: on-demand rasterization for binned data

Rasterization is already supported in SpatialData

​

​

​

​

​

​

​

​

Visium HD data is composed of small bins of 2x2 microns

Challenge:

About 6M bins and 30K genes per bin would represent 180GB as a dense image

> Solution: on-demand rasterization via rasterize_bins()

​

​

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Naive rasterization not possible. Solution: on-demand rasterization for binned data

Challenge:

About 6M bins and 30K genes per bin would represent 180GB as a dense image

> Solution: on-demand rasterization via rasterize_bins()

​

​

Visium HD data is composed of small bins of 2x2 microns

rasterize_bins() in action! Displayed on Napari

https://spatialdata.scverse.org/en/latest/tutorials/notebooks/notebooks/examples/technology_visium_hd.html

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

Naive rasterization not possible. Solution: on-demand rasterization for binned data

Challenge:

About 6M bins and 30K genes per bin would represent 180GB as a dense image

> Solution: on-demand rasterization via rasterize_bins()

​

​

Visium HD data is composed of small bins of 2x2 microns

rasterize_bins() in action! Displayed on Napari

https://spatialdata.scverse.org/en/latest/tutorials/notebooks/notebooks/examples/technology_visium_hd.html

Visium HD Spatial Gene Expression Library, Mouse Small Intestine (FFPE)

rasterize_bins() preserves spatial alignment

1a. the pixel size is different for the various bins (2µm, 8µm, …)

> rasterize_bins() handles all of this

1b. the image origin should be aligned with the vector data

Not a problem for alignment, but may lead to Moiré patterns if rasterized

(x1, y1)

(x2, y2)

Small bins as vector data

2 microns bins as rasterized data

Gene 1

Gene 2

Gene 3

1

1

2

2

3

4

4

4

5

(x1, y1)

(x2, y2)

8 microns bins as vector data

…

100

References: [1] Visium HD M.F. Oliveira, S.E.B. Taylor et al. (bioRxiv, 2024); [2] Stereo-seq A. Chen, J. Wang, et al. (Cell, 2022); [3] Open-ST M. Schott, N. Rajewsky (Cell, 2024)

Challenge:

• Visium HD: about 6M bins and 30K genes require 300MB in sparse format, but 180GB as a dense image.
• Stereo-seq: up to 70M bins, 2.1TB as a dense image

​

Solution:

• on-demand rasterization via SpatialData.rasterize_bins()
• The resulting images preserves spatial alignment (pixel sizes, rotations, etc.)

Performant representation of “bins”-based technologies

Storage of sparse data is convenient, but images are preferred for visualization. E.g. 6M bins = slow to visualize; 6M pixels = very fast

Visium HD data is composed of small bins of 2x2 microns

Figure adapted from the Visium HD paper [1]

Other popular binned technologies:

• Stereo-seq
• Open-ST

(x1, y1)

(x2, y2)

Rasterization

+

Bins as vector data

Bins as rasterized data

Gene 1

Gene 2

Gene 3

1

1

2

2

3

4

4

4

5

The scverse consortium pursues long-term software maintenance

First scverse conference (September 2024)

scverse hackathon (Innsbruck 2022)

scverse hackathon (Heidelberg 2023)

Core developers and contributors

2025 plan

• Paris hackathon
• Bay Area hackathon

​

Get involved!

(both for academia and industry)

The scverse consortium pursues long-term software maintenance

Core developers and contributors

The scverse consortium pursues long-term software maintenance

Core developers and contributors

The scverse consortium pursues long-term software maintenance

scverse hackathon (Innsbruck 2022)

Core developers and contributors

The scverse consortium pursues long-term software maintenance

scverse hackathon (Innsbruck 2022)

scverse hackathon (Heidelberg 2023)

Core developers and contributors

The scverse consortium pursues long-term software maintenance

First scverse conference (September 2024)

scverse hackathon (Innsbruck 2022)

scverse hackathon (Heidelberg 2023)

Core developers and contributors

The scverse consortium pursues long-term software maintenance

First scverse conference (September 2024)

scverse hackathon (Innsbruck 2022)

scverse hackathon (Heidelberg 2023)

Core developers and contributors

2025 plan

• Paris hackathon
• Bay Area hackathon

​

Get involved!

(both for academia and industry)

Conclusions and acknowledgements

Wolfgang Huber

Josh Moore (OME, GerBi)

Kevin Yamauchi (ETH)

Yvan Saeys (UGhent)

DKFZ

​

​

Ilia Kats

Tobias Graf

Moritz Gerstung

Elyas Heidari

​

• established interoperable format for

spatial omics based on OME-NGFF

EMBL

​

Oliver Stegle

Luca Marconato

Sinem Saka

Wouter-Michiel Vierdag

Wolfgang Huber

Harald Vöheringer

Constantin Ahlmann-Eltze

Mike Smith

Helena Crowell (CNAG)

Nils Eiling (UZH)

Will Moore (OME, UDundee)

Quentin Blampey (UParis Saclay)

Florian Wünnemann (UKHD)

Mark Keller (HMS)

10x Genomics team

CZI cellxgene team

…

spatialdata.scverse.org

Funded by

Giovanni

Palla

Lotte Pollaris (UGhent)

Benjamin Rombaut (UGhent)

Arne Defauw (VIB)

Grzegorz Bokota (UW)

Christian Tischer (EMBL)

Andreas S. Eisenbarth (EMBL)

Omer Bayraktar (Sanger)

Tong Li (Sanger)

Ilan Gold (Helmholtz)

Giovanni Palla

Kevin Yamauchi

Isaac Virshup

First authors are underlined

Helmholtz Munich

​

Fabian Theis

Giovanni Palla

Isaac Virshup

Tim Treis

Sonja Stockhaus

Laurens Lehner

Marcella Toth

Rahul Shrestha

• in-memory multimodal representation
• processing, visualization
• scales to large datasets

Wouter-Michiel Vierdag

Tim Treis

Marcela Toth

Rahul B. Shrestha

Harald Vöringer

Sonja Stockhaus

Laurens Lehner

Quentin

Blampey

Sinem Saka

Josh Moore

Nature Methods, 2024

Elyas Heidari

Luca

Marconato

Moritz

Gerstung

Fabian Theis

Oliver

Stegle

Benjamin Rombaut

Lotte

Pollaris

Yvan Saeys

Kasia Kedziora