RNA World Models: A New Frontier in Biological AI
RNA World Models: A New Frontier in Biological AI
At Eigen Bio, we believe the transcriptome holds vastly more information than traditional analysis methods can unlock. Our approach centers on building RNA World Models—foundational AI systems that model RNA as a dynamic biological system rather than a static list of genes.
Why RNA?
Cell-free RNA (cfRNA) in human plasma provides a minimally invasive, real-time readout of tissue physiology across the entire body. Unlike DNA, which is largely static, RNA expression patterns shift dynamically in response to disease, treatment, and environmental changes. This makes cfRNA an ideal substrate for inferring biological state.
From Signal to Insight
Traditional bioinformatics pipelines reduce the transcriptome to a handful of differentially expressed genes, discarding the rich correlational structure that encodes true biological state. Our foundation models learn directly from the raw transcriptomic data, capturing coordinated gene expression programs that reflect cell-type compositions, pathway activities, and tissue-level physiology.
The ORIGIN™ Platform
Our ORIGIN platform operationalizes RNA World Models across three key dimensions:
- Infer biological state from cfRNA as a dynamic system
- Simulate biological change under perturbations
- Enable decision-grade outcomes for diagnostics and therapeutics
What’s Next
We are actively expanding our foundation models to incorporate multi-modal data, improve zero-shot generalization across cohorts, and partner with clinical teams to validate cfRNA-based signatures in prospective studies. Stay tuned for upcoming publications and partnership announcements.
This is the first in a series of posts exploring the science behind Eigen Bio’s platform. Follow our Science Blog for more.