BIORANGE: augmenting BioRED dataset with NIL entity annotation and n-ary relations

BIORANGE is an extension of the BioRED dataset by annotating NIL entities and extracting n-ary relations. This work highlights the potential for reusing existing datasets, in a resourceful manner to improve edge cases.

Our academic paper describing BIORANGE in detail can be found [here] (COMING SOON).

---

Overview

This repository provides:

• Annotated Corpus: Includes NIL entities and n-ary relation annotations:
  (Entities/data/annotations/round_3.tar.gz)
• N-ary Relation Corpora: Full corpora for 3-ary and 4-ary relations:
  (RE/nary_dataset)
• Annotation Guidelines: Detailed instructions for dataset expansion:
  (BioRED_Expansion_Guidelines.pdf)

Each section directory includes a README with detailed task-specific information:

• Entities

  • Annotate NIL entities in the BioRED dataset.
  • Annotate CellTypeOrAnatomical entities using the best available identifiers in target knowledge bases.

• Relations (RE)

  • Expand binary relations in BioRED into n-ary relations (3-ary and 4-ary).

**Note on Round Labels** The datasets labeled **Round 2** and **Round 3** correspond to the paper **Round 1** and **Round 2**, respectively. The original labels were kept to avoid potential errors when renaming multiple files.

---

Citation

To cite BiORANGE, please use the following:
(Coming soon)

---

Acknowledgments

We acknowledge the contributions of:

• Furkan Goz and Filipe Nascimento: Assistance during curation of NER, EL, and RE tasks.
• Authors of BioRED: For making the data and code publicly available.

This work was supported by:

• FCT (Fundação para a Ciência e a Tecnologia) through:
  • PhD Scholarship 2020.05393.BD (awarded to PR).
  • PhD Scholarship UI/BD/153730/2022 (awarded to SIRC).
• LASIGE Research Unit, reference *UID/000408/2025

---

Additional Resources

• BioRED dataset BC8 release
• BioRED paper