The rapid growth of publicly available single-cell RNA sequencing data has opened the door to answering many new biological questions. Researchers can now explore how individual cells behave in health and disease across many tissues and conditions. However, actually reusing these datasets is often slow and difficult. Scientists usually need to read the original publication, extract metadata, locate the correct repository, download the data, and then run a series of analysis steps that often require custom scripts and computational expertise.

Researchers at AstraZeneca have developed a new system called CellAtria to simplify this entire process. CellAtria is designed to act like an intelligent assistant that guides users from published studies all the way to analysis-ready single-cell data, using natural language interactions rather than manual coding.

CellAtria combines two key components. The first is an AI agent that can read scientific articles, identify relevant datasets, retrieve them from public repositories, and decide what processing steps are needed. Users interact with this agent through a chatbot style interface, asking questions or requesting analyses in plain language. Behind the scenes, the AI coordinates each step of the workflow.

The second component is CellExpress, an automated analysis pipeline built specifically for single-cell RNA sequencing data. Once raw count matrices are retrieved, CellExpress applies widely accepted processing steps, such as quality control, normalization, and basic downstream analyses, to generate standardized and analysis-ready single-cell profiles.

Together, these components remove many of the barriers that currently limit data reuse. Researchers no longer need to write custom scripts or stitch together fragmented tools. Instead, CellAtria provides a consistent, reproducible workflow that reduces turnaround time and makes single-cell data analysis accessible to scientists with little or no computational background.

By automating dataset ingestion and standardizing downstream processing, CellAtria has the potential to accelerate discovery and improve reproducibility across the single-cell field. As the volume of RNA sequencing data continues to grow, tools like this may play an important role in ensuring that valuable public datasets are fully and efficiently reused.

Availability – The source code, documentation, and all materials required to execute the agent are publicly available on the GitHub at https://github.com/AstraZeneca/cellatria.