Bio+MedVis Challenge @ IEEE VIS 2024

Latest News

Bio+MedVis Challenge 2024 Published 14 Mar 2024

The Bio+MedVis Challenge 2024 has been published. Please visit the challenge page for details.


Redesign Challenge: Redesign an Existing Visualization

Biological Background and Data Description

In biology, structure and function are tightly linked, and the biology of cells in tissues is no exception. A key challenge for cell biologists is understanding the spatial patterns characteristic of health and disease. A cutting-edge area of cell biology is the development and application of technologies that measure molecular states within cells while preserving spatial context (historically, many single-cell profiling techniques have required dissociation of tissues).

A key cellular measurement technique is called transcriptomics, which provides insight into cell types (e.g., immune cell vs. muscle cell), cell states (e.g., response to stimuli), and information about proteins that might be produced downstream (i.e., via translation). Transcriptomics techniques measure gene expression: the abundance of RNA transcripts corresponding to each gene in each cell. As the name implies, spatial transcriptomics methods capture gene expression and spatial coordinates simultaneously. Resolutions of spatial transcriptomics methods can vary from subcellular, to single-cell, to “spots” that represent multiple cells. There is typically a tradeoff between the spatial resolution and the number of genes that can be measured by a given technology.

VISIUM is a spatial transcriptomics technology developed by 10x Genomics. It allows researchers to measure gene expression and spatial organization of tissues simultaneously. For a given tissue slice, VISIUM can measure transcripts genome-wide (e.g., ~20,000 genes for human samples) at a resolution of approximately 5-10 cells per spot. This enables researchers to understand how gene expression patterns vary across different regions of a tissue sample. This technology has applications in various fields such as cancer research, developmental biology, neuroscience, and more, allowing researchers to gain deeper insights into the molecular mechanisms underlying complex biological processes within tissues.

Each VISIUM spot represents the gene expression of multiple cells, but researchers are typically interested in interpreting the data at the cell type level (e.g., linking expression differences to particular cell types that are characterized in the biological literature). As a result, computational methods have been developed to deconvolute each VISIUM spot into a distribution of (predicted) cell type proportions. Biologists can then use the deconvolution results to link their findings to existing cell type knowledge.

These deconvolutional methods thus produce data that, on a per-spot basis, represents proportional cell-type membership (e.g. 30% Cell Type 1, 50% Cell Type 2, 20% Cell Type 3). Visualizing these data in the context of the tissue image provides crucial spatial context, enabling domain experts to explore deconvolution results within the intricate cellular architecture and tissue morphology. One common technique for visualizing these data is to superimpose a pie chart on each spot location representing the cell types represented. Pie charts and their limitations have been well explored by the visualization community. Moreover, superimposing them on the image occludes the underlying tissue.

Example data for Redesign Challenge
Figure 1: Example data for Redesign Challenge. (a) H&E image, (b) spot cell-type proportion pie charts, and (c) Pie charts superimposed on the image. Plots generated w/ STdeconvolve (Miller et al., 2022).

Data and Documentation

Example Dataset

Data Description

Additional Data and Resources

Redesign Challenge Task

For this challenge, we ask participants to propose an alternative visualization approach that links the spot cell-type proportion to the H&E image. Submissions should consider visualization theory and principles.

Sketching and prototyping as well as fully interactive solutions will be accepted. Creativity and novelty will be considered in the evaluation.


Kleshchevnikov, V., Shmatko, A., Dann, E. et al. Cell2location maps fine-grained cell types in spatial transcriptomics. Nat Biotechnol 40, 661–671 (2022).

Miller, B.F., Huang, F., Atta, L. et al. Reference-free cell type deconvolution of multi-cellular pixel-resolution spatially resolved transcriptomics data. Nat Commun 13, 2339 (2022).


Submissions will be considered for talk or poster presentations. Please send a two-page PDF abstract with up to 5 additional figures and a draft of your proposed poster (max 10MB) to the PCS The abstract should include:

Selected submissions will be invited for talk presentations during the Bio+MedVis session at the IEEE VIS 2024 conference.

Evaluation of Submissions

All submissions will be thoroughly reviewed by at least two reviewers, coming from the challenge chairs and selected domain experts. All accepted submissions will be published in the conference proceedings. Winning designs may be invited to become a plugin or extension to Vitessce.

Important dates


Please feel free to ask questions at:

The chairs of the Bio+MedVis Challenge @ IEEE VIS 2024: