New Tool Links 125 Cell Types Across 700 Million Years
Researchers at KAUST introduced Unify, an AI system that maps cellular similarities across distant species to improve human health research.
Key points
- Unify compares 125 cell types across seven species spanning 700 million years of evolution.
- The tool groups genes by biological function rather than direct sequence matches to find similarities.
- Unify predicted human cell responses based on mouse data with higher accuracy than existing methods.
A team led by King Abdullah University of Science and Technology has released a new artificial intelligence tool designed to bridge the gap between animal research and human medicine. The system, named Unify, compares 125 distinct cell types from seven different species, covering a span of more than 700 million years of evolution. According to Phys.org, the study published in Nature Communications suggests this approach can better identify which findings from animal models are actually relevant to human biology.
Traditional methods for comparing cells often rely on finding direct one-to-one matches between genes in different species. However, this strategy becomes increasingly difficult as species diverge over millions of years, often causing researchers to miss important biological similarities. By shifting the focus from specific gene names to the biological functions those genes perform, the new tool aims to capture shared cellular jobs that persist even when the underlying genetic code has changed significantly.
Focus on function over gene labels
Conventional comparison tools operate somewhat like dictionaries looking for word-for-word translations, which fails when languages evolve differently. Unify instead uses AI models to analyze protein sequences and scientific descriptions of gene functions, grouping them into broader units called macrogenes. This allows the system to recognize cells performing similar roles, such as immune defense, even when the specific genes involved no longer match directly between species like humans and mice.
Testing predictions across distant species
In practical tests, the tool successfully reconstructed relationships among cell types from species separated by vast evolutionary distances. It distinguished between identical genes, those that evolved new jobs, and different genes that independently evolved to do the same task. When analyzing immune cells, Unify identified shared defense tactics that direct gene comparison methods missed, highlighting its ability to detect deeper biological patterns.
The tool also demonstrated improved predictive power in a specific experiment involving blood cells. By analyzing the response of mouse lymph node immune cells to a signaling protein, Unify predicted the corresponding response in human blood cells with greater accuracy than existing methods. This capability is crucial for translating animal study results into human health applications, as it helps filter out findings that may not carry over effectively.
Limitations and future expansions
Despite its improvements, the tool is not yet complete. The KAUST team is currently working to extend Unify to include data on gene regulation and the physical positions of cells within tissues. These additions are expected to provide a fuller picture of biological principles shared across life. Until these features are integrated, the tool remains limited in its ability to account for spatial and regulatory contexts, which are critical for a complete understanding of cell behavior.
