Advances in imaging tools offer spatial biology for the AI age

10x Genomics introduced Atera, a spatial-transcriptomics instrument designed to analyze nearly the full human transcriptome at single-molecule resolution within tissue samples. The device detects hundreds to thousands of transcripts per cell, covering roughly 93% of human protein-coding genes, enabling researchers to map cell types and spatial interactions with unprecedented detail. Serge Saxonov, cofounder and CEO of 10x Genomics, stated that achieving whole transcriptome spatial analysis at high throughput was previously considered impossible, requiring advancements in optics, chemistry, and fluidics. The instrument generates massive datasets—equivalent to streaming Netflix in 4K for 12 years per run—demanding AI-driven solutions for processing and interpretation. On-platform software developed by 10x handles raw data analysis, while vendors like 10x increasingly integrate early-stage AI tools for tasks such as cell segmentation, annotation, and 3D mapping. Michael Snyder, a Stanford University biochemist and co-leader of the Human BioMolecular Atlas Program (HuBMAP), highlighted the necessity of AI for managing complex computational challenges in spatial biology. Spatial biology combines techniques like spatial transcriptomics, light-sheet microscopy, and cryo-electron tomography to study biomolecules in their tissue context. These tools provide high-resolution, high-context data that AI models can leverage to predict biological systems more accurately. The synergy between advanced imaging instruments and machine learning is accelerating research by enabling large-scale, high-resolution data collection and analysis. Saxonov noted that no single breakthrough made Atera possible, but rather a series of optimizations across probe chemistry, fluidics, and superresolution imaging. The instrument’s capabilities reflect a broader trend in bioanalysis, where AI is essential for extracting meaningful insights from exponentially growing datasets. Researchers now have new tools to explore biological complexity, bridging the gap between observation and predictive modeling.