Skip to content

Helena Crowell — "Colorectal cancer through the lens of whole transcriptome imaging"

Affiliation: Charité — Universitätsmedizin Berlin; scverse community core developer Day / session: Day 1 (Wed Sep 17, 2025) — morning keynote 2 Talk title: "Colorectal cancer through the lens of whole transcriptome imaging" Slides: Zenodo 10.5281/zenodo.17189967 Video: Bioconductor YouTube — plenaries playlist (keynotes recorded) Abstract / schedule: eurobioc2025 schedule

Thesis

Whole-transcriptome imaging platforms (Xenium 5K-plex, MERSCOPE, Stereo-seq, etc.) collapse the long-standing trade-off between gene-panel breadth and subcellular spatial resolution — and colorectal cancer, with its well-mapped niches (tumor epithelium, stromal CAFs, T-cell exclusion zones, mucinous regions), is a productive testbed for what changes when you can ask transcriptome-scale questions in situ. The talk almost certainly argues that the bottleneck has shifted from acquisition to analysis tooling that can handle multi-million-cell, multi-modal objects coherently.

Speaker context

Helena Crowell is a core developer in the scverse / SpatialData community and a methodological voice on differential-state and compositional analysis in single-cell data (muscat co-author, several CyTOF / scRNA-seq method papers). Her appointment at Charité Berlin places her at one of Europe's larger spatial-omics groups. She is an author on the SpatialData framework, which is the Python/Zarr substrate now being mirrored by Bioconductor's SpatialExperiment / SpatialData-R. Inviting her gives the EuroBioC audience a direct line into the scverse interop conversation — which is exactly the bridge Carey's opening keynote is likely setting up.

Connections to the corpus

  • SpatialData — Crowell is a co-author of the underlying framework; her talk is the applied motivation
  • OSTA — the Bioconductor book that codifies the analysis patterns she would be deploying
  • DESpace — differential-spatial-pattern testing across CRC niches is exactly DESpace's DSP mode
  • spatialFDA — functional-data approaches to spatial heterogeneity, complementary to image-based whole-transcriptome readouts
  • AACR axis: single-cell-spatial-omics — this is the canonical EuroBioC contribution to that axis

Open questions

  • Which platform(s) — Xenium 5K, MERSCOPE, Stereo-seq, Visium HD — and at what scale (cells, samples, patients)? CRC studies range from single-block proof-of-concept to atlas-scale cohorts.
  • Does the analysis stack run in Python (SpatialData / squidpy) or R (SpatialExperiment / OSTA), and are the niches called from segmentation-first or transcriptome-first clustering?
  • New methodological contribution vs. application paper? The title hints at biology-forward; the EuroBioC audience would want to know which packages were touched.