Frequently Asked Questions

Virtual spatial transcriptomics predicts spatial gene expression from H&E images, without requiring a new spatial assay to generate the prediction. Researchers can use the results to form hypotheses and decide what to validate experimentally. Predictions remain computational estimates, not assay measurements.

DeepSpot-M predicts the entire protein-coding transcriptome: one fixed set of 19,338 protein-coding genes, the same whatever tissue or disease you select. A submission returns this set. DeepSpot-M can also predict non-coding RNAs, such as tRNAs, lncRNAs, rRNAs and miRNAs, which poly(A)-based assays such as standard 10x Visium capture only incompletely. These are not part of a standard result; contact us if your project needs them.

Timing varies by submission and current demand. We'll email you when your results are ready. You can also follow the status of your submission on the Track page.

Yes. After quality filtering, each image is limited to 50,000 tiles. If your image produces more tiles, the system automatically selects the densest contiguous square region to preserve spatial context for downstream analysis. The original tile count and subsampling status are displayed on your results page. For custom limits, please contact us via the support form.

Each result includes a predicted expression matrix (.h5ad.gz), a spatial analysis report and a quality-control report. Spatial gene maps are also available in the results view.

Uploaded images, generated results and account information are stored on infrastructure in Switzerland, and analyses run there. Cloudflare may process website traffic outside Switzerland; see the Privacy Policy for details.

Uploaded data and generated results are confidential and are not made publicly available. You can retrieve your submission with its unique ID and verified email address. Members of our research team can access the data as reasonably necessary to operate the platform, generate results, carry out directly related research and develop Aurora's models and services, as described in the Terms.

Under the academic terms, yes. Academic access is offered at academic rates, and in return Aurora may use the data you upload, and the results generated from it, to develop, train, validate and improve its models and services, including those it offers commercially. Your data is never made public, and deleting it does not withdraw a model already developed with it. Deletion does not remove copies already held in backups. Commercial access is governed by a written agreement with Aurora, and a commercial customer's data is not used for model development without written permission.

Yes. The API and the Python package let you submit samples and read predictions without the browser. Programmatic access is arranged separately from an account on this site: get in touch, tell us who you are and what you intend to do, and we will take it from there.

Academic Standard starts with 3 images. Rating a result adds one image to your allowance. If you would like to analyse more, get in touch and tell us a little about your project. We can discuss the appropriate route for your work.

If you encounter any issues on the platform or need additional support, please use our contact form to submit a detailed description of the problem or request.