A Compound-Centric Framework for Mapping Plant Chemical Space through Structural Scaffolds and Bioactivity Evidence

Area of Science:

• Natural Products Chemistry
• Cheminformatics
• Bioinformatics

Background:

• Open repositories offer vast plant chemical occurrence data.
• Data challenges include taxonomic inconsistency, structural redundancy, and poor bioactivity annotation, limiting compound prioritization.

Purpose of the Study:

• To present a compound-centric framework for transforming and interpreting plant chemical occurrence data.
• To integrate diverse data sources for evidence-aware compound prioritization in natural products research.

Main Methods:

• Integration of LOTUS species-structure data, World Flora Online taxonomy, InChIKey dereplication, and Murcko scaffold analysis.
• Utilized ChEMBL database for bioactivity evidence.
• Applied the framework to Phyllanthaceae for a case study.

Main Results:

• Organized chemical space across Phyllanthaceae genera and resolved structural redundancy.
• Linked novel scaffolds to family-level rarity and bioactivity records.
• Defined two categories: STARS (compounds with activity) and HIDDEN GEMS (compounds without qualifying activity).

Conclusions:

• The framework provides interpretable, evidence-aware outputs from heterogeneous occurrence data.
• Supports auditable and transparent prioritization in natural products research.
• Taxon-flexible design enables analysis across multiple taxonomic scales.