NTTMUNSW BioC modules for recognizing and normalizing species and gene/protein mentions

Area of Science:

• Biomedical Informatics
• Computational Biology
• Bioinformatics

Background:

• The rapid growth of biomedical literature presents challenges in accurately identifying and normalizing gene and species information.
• Ambiguity in gene and protein nomenclature complicates data curation for molecular interaction databases.

Purpose of the Study:

• To develop and evaluate a species normalization module that identifies species names and maps them to NCBI Taxonomy IDs.
• To create gene normalization modules for recognizing gene mentions and linking them to Entrez Gene IDs using a multistage algorithm.
• To ensure all developed modules are BioC-compatible and publicly available.

Main Methods:

• Developed a species normalization module that recognizes species names, including prefixes indicating gene origin, and normalizes them to NCBI Taxonomy IDs.
• Employed two separate modules for gene normalization, utilizing a multistage algorithm for processing full-text articles to identify gene mentions and map them to Entrez Gene IDs.
• Ensured modules are implemented as BioC-compatible .NET framework libraries.

Main Results:

• The species normalization module achieved a high F-score of 0.954 on an instance-level corpus.
• Successfully developed modules for both species and gene normalization, addressing complexities in biomedical text.
• All modules are publicly available on the NuGet gallery.

Conclusions:

• The developed modules effectively address the ambiguity in gene and species nomenclature within biomedical literature.
• These tools enhance the accuracy and efficiency of data curation for molecular interaction databases.
• The BioC-compatible libraries offer a valuable resource for the research community.