Academic researchers: select research offerings, derived from our products, free for academic research

H&E Spatial Profiling

H&E Spatial Profiling

OpenTME

4,691 TCGA whole-slide images, fully analyzed with HETA.

Research
Access Program

Apply to have your own H&E slides analyzed at no cost.

PanCAN
SPARK

4,691 TCGA whole-slide images, fully analyzed with HETA.

Learn more

RudolfV 2

4,691 TCGA whole-slide images, fully analyzed with HETA.

Learn more

OpenTME

TCGA is one of the most comprehensively characterized cancer cohorts available, with linked genomic, transcriptomic, proteomic, and clinical data.

OpenTME provides HETA outputs for thousands of TCGA whole-slide images, including pre-computed spatial features in .csv format covering tissue quality control, tissue segmentation, cell detection / classification, and neighborhood analyses.

OpenTME is an ongoing initiative. We will continue to add new TCGA images as HETA coverage expands to additional cancer types.

Research Access Program

Run your own slides, at no cost

The Research Access Program is an application-based initiative that provides free access to HETA for select academic projects.

Selection process

We review applications on a rolling basis and select projects based on scientific merit. There is no cost to apply or to participate in the Program. We aim to respond to all applicants within 2 weeks of submission. The application should take no more than 30 minutes to complete.

The application covers

Basic information about you and your institution

An overview of your proposed project and its scientific impact

Future publication plans

PanCAN SPARK

Pancreatic cancer profiles from PanCAN

HETA profiles and linked multimodal data for 1,400+ pancreatic cancer patients, available through the PanCAN SPARK platform.

RudolfV 2: open-access pathology foundation models

Across a broad set of public benchmarks, RudolfV 2 is the best performing open-access pathology foundation model, approaching the performance and robustness of our flagship model, Atlas 2. Two distilled versions preserve high performance at a fraction of the compute.

Model
Backbone
Parameters
Speed
RudolfV 2
ViT-g
~1.1B
1x
RudolfV 2-B
ViT-B
86M
2.5x faster
RudolfV 2-S
ViT-S
22M
6.5x faster

RudolfV 2 includes a dedicated post-training stage, a first for pathology foundation models.