CEAR: Automatic construction of a knowledge graph of chemical entities and roles from scientific literature

TL;DR

CEAR addresses the challenge of expanding ChEBI and grounding chemical knowledge in literature by automatically constructing a knowledge graph of chemical entities and roles (CEAR). It combines ontology-guided augmentation with transformer-based named entity recognition and a second LLM for relation validation across a corpus of 8,000 ChemRxiv papers, producing a literature-backed KG linked to ChEBI. The results demonstrate strong precision/recall in entity and role identification and reveal new CEAR entries not present in ChEBI (e.g., PBS as a buffer), with RDF/Turtle representations and plans for RDF-star integration. This approach offers a scalable path to extending chemical ontologies and enabling enhanced search, exploration, and knowledge integration in chemistry research.

Abstract

Ontologies are formal representations of knowledge in specific domains that provide a structured framework for organizing and understanding complex information. Creating ontologies, however, is a complex and time-consuming endeavor. ChEBI is a well-known ontology in the field of chemistry, which provides a comprehensive resource for defining chemical entities and their properties. However, it covers only a small fraction of the rapidly growing knowledge in chemistry and does not provide references to the scientific literature. To address this, we propose a methodology that involves augmenting existing annotated text corpora with knowledge from Chebi and fine-tuning a large language model (LLM) to recognize chemical entities and their roles in scientific text. Our experiments demonstrate the effectiveness of our approach. By combining ontological knowledge and the language understanding capabilities of LLMs, we achieve high precision and recall rates in identifying both the chemical entities and roles in scientific literature. Furthermore, we extract them from a set of 8,000 ChemRxiv articles, and apply a second LLM to create a knowledge graph (KG) of chemical entities and roles (CEAR), which provides complementary information to ChEBI, and can help to extend it.

Figures (6)

• Figure 1: Working steps (yellow) and resources (green) used to create the KG (blue and red)
• Figure 2: Information types provided by our approach.
• Figure 3: Sample sentences with inferred chemical entities (red) and roles (blue).
• Figure 4: Question answering using LLAMA-2-CHAT
• Figure 5: While solvent is annotated in CRAFT (blue), dissolved and redissolved are not.