Automatic gene annotation using GO terms from cellular component domain.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene Ontology (GO) provides structured information on gene product functions across Cellular Component (CC), Molecular Function (MF), and Biological Process (BP) domains.
• Manual GO annotation, assigning GO terms to genes from literature, is time-consuming and labor-intensive for Model Organism Databases (MODs).
• Current manual annotation efforts can only cover a fraction of relevant scientific articles due to resource limitations.

Purpose of the Study:

• To develop an automated approach for gene annotation using GO terms, specifically within the Cellular Component (CC) domain.
• To address the challenge of efficiently assigning subcellular location and protein complex information to genes.

Main Methods:

• The study frames gene annotation with CC domain GO terms as two relation extraction subtasks: identifying protein subcellular locations and protein complex subunits.
• A relation extraction approach utilizing triggers and syntactic dependencies was employed for both subtasks.
• The method was evaluated on the publicly available BC4GO test corpus.

Main Results:

• The proposed approach achieved a F1-score of 71% for predicting GO terms within the CC domain for given genes.
• The system demonstrated a precision of 91% and a recall of 58% in these predictions.
• These results indicate a significant improvement in the efficiency and accuracy of automated GO annotation.

Conclusions:

• A novel method was developed to automate gene annotation with CC domain GO terms by treating it as two relation extraction tasks.
• The approach successfully accelerates the gene annotation process for bio-annotators.
• The evaluation results confirm the effectiveness of this automated strategy for GO annotation.