Building the virtual molecular layer for tissue biology.

Aurora predicts spatial gene expression from routine H&E images, helping researchers explore molecular patterns across samples, cohorts and disease.

Virtual spatial transcriptomics from H&E. Built for the hard questions in biology.

Rooted in research at
  • ETH Zurich
  • University of Zurich
  • University of Basel

Working on biomarkers, targets or therapeutic response? Explore commercial access

Histology is everywhere. Molecular profiling isn't.

Decades of research have created tissue images at enormous scale. Aurora is building a way to explore them molecularly, without a new molecular experiment for every sample.

Virtualmolecular layerYour H&Eslide

What Aurora makes possible

One image. Another dimension.

  1. 01

    One slide

    One image can open a virtual molecular view of one sample.

  2. 02

    Many slides

    A cohort can show virtual molecular patterns across many.

  3. 03

    Entire collections

    At scale, histology becomes a new substrate for molecular exploration.

One sample can suggest a pattern. A cohort can show whether it repeats.

Read the science

The questions Aurora is built for.

  • Why does the same disease behave differently across patients?

  • Which molecular patterns define a meaningful subgroup?

  • Where should we look for the next biomarker, target or mechanism?

  • Why do some patients respond while others do not?

Aurora brings a virtual molecular view to tissue at cohort scale, helping researchers investigate heterogeneity, disease mechanisms, biomarkers, targets and therapeutic response.

Where the value appears

Turn more tissue into better decisions.

  • Find patterns at cohort scale

    Explore virtual molecular variation across samples and biological groups.

  • Understand what drives the biology

    Investigate heterogeneity, subgroups, biomarkers, mechanisms and therapeutic response.

  • Prioritise what to pursue

    Focus validation and experimental effort on the signals, samples and hypotheses that matter most.

Better therapies begin with better biology.

Before a therapy can be developed, the biology has to be understood. The right subgroup has to emerge. A mechanism has to become worth pursuing. A biomarker or target has to survive deeper investigation. Aurora is being built to help researchers find more of those starting points.

From the archive

Past tissue. Future questions.

A cohort assembled years ago can meet a question that emerged yesterday. Aurora gives archived slides a virtual molecular view, so tissue already in the drawer can shape the studies still to come.

A new chapter for your collection.

The virtual molecular layer starts with scale.

Scale matters because many of biology's hardest questions only emerge across patients, samples and cohorts.

tissue slides
28,000+
predicted spatial spots
295M+
cancer types
32+
genes and RNAs
19,338+

Independently benchmarked

State-of-the-art performance, independently tested.

Research models underpinning Aurora ranked first in two independently organised open challenges, ahead of competing methods from other teams.

  • 1st place

    Autoimmune Disease ML Challenge

    Broad Institute of MIT and Harvard

  • 1st place

    MICOS Europe

    BGI Genomics

Presented at
  • 10x Genomics
  • Google Genomics Deep Dives
  • NeurIPS

Built to work where research already lives.

Use Aurora through Swiss-hosted infrastructure, or keep original images within your own environment through our secure API approach.

  • Swiss-based processing
  • Encrypted transfer
  • Local-image options

The tissue is already here. A new layer of biology is within reach.

Aurora is opening a richer virtual molecular view of existing tissue, so the samples collected yesterday can help answer the questions that matter tomorrow.

Working on biomarkers, targets or therapeutic response? Explore commercial access