Noetik Announces Foundation Models of Spatial Biology for Precision Oncology
After securing $40 million in August 2024, Noetik—an AI-native biotechnology company—announced in a recently released technical report the launch of OCTO-VirtualCell, a next-generation AI model designed to reveal how cells behave in both healthy and disease contexts. Noetik leverages self-supervised machine learning and high-throughput spatial data to advance precision therapies.
See also: 19 Companies Pioneering AI Foundation Models in Pharma and Biotech
Complementing this, Noetik introduced Celleporter, an interactive tool enabling researchers to simulate virtual cell behaviors across spatial tissue locations. OCTO-vc leverages a multi-scale, multi-modal transformer architecture trained via masked token modeling to predict spatial gene expression patterns, trained on nearly 40 million cells from over 1,000 patient tumor samples.
We place virtual cells in different regions of real tissue samples to probe specific biology in specific patients. For example, we can drop a virtual T cell into various parts of a tumor to infer whether the T cell will be activated – capable of killing cancer cells – or not. By running virtual cell simulations at many locations, we create a holistic picture of the underlying biology.
OCTO-vc
Each OCTO-vc simulation places a virtual cell within a local patch of a real patient tissue sample (left). Based on the gene expression of neighboring cells (top middle) and a prompt specifying which genes are expressed or not expressed in the virtual cell (bottom middle, green bars), OCTO-vc makes predictions about other genes in the virtual cell (bottom middle, gray bars). Simulating virtual cells at thousands of positions in the tissue creates a map of gene expression for a given prompt (right). In this and all following simulated expression maps, red indicates highest expression and blue indicates lowest expression.
Ron Alfa, CEO & Co-Founder of Noetik, said:
“OCTO-vc is a transformative tool for understanding the hidden biology driving disease progression. Leveraging its generative capabilities, we can simulate how virtual cells behave in different tissue environments, shedding light on mechanisms that were previously inaccessible. By scaling these simulations to cohorts of patients within our dataset, we can run virtual experiments to discover patient and therapeutic biology.”
The model allows for virtual experiments, simulating cell behaviors in different tissue environments. For instance, researchers can simulate the behavior of T cells across tumor regions to determine their activation potential and ability to kill cancer cells.
Jacob Rinaldi, CSO & Co-Founder of Noetik, added:
“We think of this technology as analogous to a new kind of microscope, a new way to see and understand biological systems. We think these methods could usher in a new era of self-supervised science, and allow us to redefine the taxonomy of disease at the systems level.”
The accompanying Celleporter visualization tool provides an interactive interface for exploring these virtual simulations, enabling researchers to uncover hidden biological insights within tissue samples. Together, OCTO-vc and Celleporter represent a step forward in spatial biology, allowing scalable experimentation on real patient-derived data.
Celleporter
While oncology is Noetik’s initial focus, the technology’s applications extend beyond cancer, serving as a general interface for spatial biology and enabling foundational research and drug discovery capabilities across a broad range of therapeutic areas.
To explore OCTO-vc and interact with Celleporter, visit https://www.noetik.ai/octo-vc and https://celleporter.noetik.ai.
Topics: AI & Digital
