Spatial biology holds the promise of characterizing complex biological systems to understand diseases and develop new therapies.

Cost and complexity are large risk barriers to generating this data.

Our proprietary AI reduces the time and cost of performing in situ single-cell whole transcriptome studies by more than 90%.

About us

Cell-level resolution of every transcript without doing the work.

Our proprietary iStar™ technology uses generative AI to infer a whole transcriptome profile at single-cell resolution without performing the expensive molecular imaging assays required to capture this data.

What is iStar™ ?

iSTAR™ integrates Visium™ spatial transcriptomic (ST) data sets and histology images to provide complete single-cell gene expression profiles from incomplete ST datasets using our hierarchical vision transformer (HViT) super-resolution technology.

Why use iStar™?

Although Visium™ only captures a small, incomplete data set from a tissue, our AI approach infers single-cell spatial gene expression across the whole tissue. This method has been demonstrated to out-perform all other existing methods in a recent Nature Biotechnology paper.

What can I do with iStar™?

Having whole-tissue, single-cell resolution on tissue sample analyzed with Visium™ is just the beginning. With a small training set, the method can be applied to new H&E images to predict spatial transcriptomic profiles from untested samples, eliminating the need to perform Visium™ to gain the ST profile of untested tissues.

Discovery Research applications: Extend your data reach

-Create super resolution data set to characterize the whole tissue.

-Create a virtual ST profile to test in a larger cohort without performing additional testing.

Translational Research applications: Elucidate a biomarker signature

-Identify a new biomarker in the clinical samples based on a virtual ST profile created from an outside set.

-Confirm the expected presence of the virtual biomarker in the clinical samples prior to analytical testing.

Clinical applications: Enhance your multi-omic data profile

-Use an orthogonal ST dataset to create an inferred signature and apply to entire dataset to strengthen the integrated diagnostic signature, and/or create new predictions.

Why partner with OmicPathAI ?

OmicPathAI is the inventor and continual developer of this enabling technology (patent pending).

OmicPathAI is led by recognized leaders in the field of spatial transcriptomic data prediction and integration.

We partner our technology with you to:

Create tools to understand tissue (macro) level to cell (micro) level gene expression patterns
Classify spatial biology functional phenotypes using ST data
Create novel tissue scoring systems that reflect the target biology
Discover novel virtual biomarkers that increase the efficacy of an integrated diagnostic strategy

iStar is an AI based method which uses Visium™ datasets from the small area of incongruent spots characteristic of the approach to predict gene expression resolution to single-cell levels across the whole tissue.
The method provides a tissue-wide, high resolution data file that can be use to create tissue classifiers, cell phenotypes, or identify tissue structures (TLS example).

The HViT technology extracts histology features at multiple scales to capture both gross and fine tissue features. These features are used to predict gene expression levels using a feed-forward neural network which is trained through weakly supervised learning. This model then assigns multiple values of gene expression within each spot, facilitated by the histology features Additionally, the model predicts l gene expressions outside the spots within the tissue, or in external tissue sections, using an H&E image as reference to assign a gene expression value against the histology features.

Build from your existing work. Use our services to:

Create cell-level data from a sample cohort already tested using Visium™
Model training/development and validation to create a tool which can provide cell-level data for untested samples.
Workflow customization and systems integration for your specific needs.
Development of proprietary IP licensed to you.

“The journey for spatial biology data from data acquisition to useful data is arduous, involving complex and expensive molecular and data workflows. We reduce the cost barrier to entry and accelerate the time-to-data window, enabling you to test large cohorts of samples to validate hypothesis and conclusions.”

Joseph Krueger

CEO, PathOmicAI

Watch, Read, Listen

No posts were found.

Join 900+ subscribers

Stay in the loop with everything you need to know.

Sign up