AI Framework Automates Extraction of Tissue Unit Data for Human Atlas
A new AI‑driven system, HRAftu‑LM‑RAG, has been developed to automatically harvest functional tissue unit (FTU) information from the scientific literature. The framework combines large language models, large vision models, and retrieval‑augmented generation to parse 244,640 PubMed Central papers, covering 1,389,168 figures across 22 FTUs. It identified 617,237 microscopy or schematic images and…
Key points
- HRAftu‑LM‑RAG processes 244,640 papers and 1,389,168 figures to extract 1.7M entity mentions
- The system pulls 331,189 scale bars and 617,237 microscopy images for 22 functional tissue units
- Data feeds into the Human Reference Atlas, streamlining future illustration design
The extracted data feed directly into the Human Reference Atlas (HRA), enabling more accurate and efficient design, review, and approval of FTU illustrations. The effort involves over 25 international consortia, including HuBMAP, SenNet, KPMP, GUDMAP, and NIDDK, and represents a scalable solution to a previously manual, labor‑intensive task.
By automating the extraction of multiscale anatomical data, the system accelerates the construction of a comprehensive, data‑rich reference atlas that can support biomedical research, drug development, and precision medicine initiatives.
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us.
Comments
via GitHub Discussions