DeepSpot-M

DeepSpot-M is the molecular prediction model behind Aurora's virtual spatial transcriptomics. It turns tissue representations into predicted gene expression across the slide.

DeepSpot-M was introduced in a preprint. Read the paper.

How Aurora works

1. Read tissue morphology. DeepSpot-H represents the tissue structure in each part of the slide.

2. Predict spatial gene expression. DeepSpot-M uses those tissue representations to predict expression.

3. Explore molecular patterns. The result places predicted expression on your slide's own coordinates, to open in Aurora Intelligence or your own tools. What you get back says what it holds.

Explore the molecular patterns

DeepSpot-M is the model that predicts spatial gene expression.

The result places predicted expression on the tissue's coordinates. You can explore individual genes, compare regions and focus the next experiment. What you get back explains the molecular view and its values.

The API catalogue lists the available releases, and each release's model card states its gene coverage.