A public virtual spatial proteomics resource for TCGA-COAD

We're releasing predicted spatial protein maps for 442 diagnostic H&E slides from the TCGA colon adenocarcinoma cohort (TCGA-COAD). We used Lattice, our virtual staining model, to generate the maps. They're available as a public resource for research.

Explore spatial patterns across the cohort by browsing the original tissue alongside the predictions. Pan and zoom through whole slides to compare predicted markers over the same field of view. You can browse without signing in.

We've already seen that these predictions carry a prognostic signal in this cohort. More on that soon.

Explore a slide

Drag to pan, scroll to zoom, and toggle the predicted channels. Split view shows H&E and predictions side by side over the same tissue.

For this cohort, we ran a subset of seven protein markers plus DAPI: CD3e, CD8, CD20, CD68, Ki67, CD45, PanCK, and DAPI.

Use the resource

Browse the slides, find metadata and downloadable examples on Hugging Face, or use the Python SDK in your research. For programmatic access, sign up for Strand AI and create an API key.

Predictions and metadata are released under CC BY-NC 4.0: attribution is required and commercial use is excluded. For commercial use, contact Strand AI. Source H&E slides remain available from the NCI Genomic Data Commons under GDC data policies.

Research use only (RUO). Not for use in clinical or diagnostic procedures.