Functional knowledge transfer for high-accuracy prediction of under-studied biological processes

Area of Science:

• Genetics
• Bioinformatics
• Systems Biology

Background:

• Identifying gene functions in biological pathways is a major challenge in genetics.
• Existing methods often fail to identify genes in understudied processes or transfer knowledge between organisms effectively.
• Sequence similarity methods like BLAST can transfer annotations but have limitations.

Purpose of the Study:

• To develop and validate a method combining functional genomics and sequence similarity for improved gene annotation transfer between organisms.
• To enhance the accuracy of predicting gene-pathway membership, particularly for poorly characterized biological processes.
• To expand the coverage of accurate gene function predictions across diverse organisms.

Main Methods:

• Developed a functional knowledge transfer (FKT) method that integrates genomic data with machine learning algorithms.
• Applied FKT to combine transferred gene function knowledge with organism-specific high-throughput data.
• Compared FKT performance against traditional sequence similarity methods and state-of-the-art machine learning algorithms.

Main Results:

• FKT significantly improves the accuracy of gene-pathway membership prediction, especially for processes with limited experimental data.
• The FKT method demonstrates superior performance compared to annotation transfer solely based on sequence similarity.
• Predicted novel gene functions for 11,000 biological processes across six organisms, increasing coverage for understudied areas.

Conclusions:

• Functional knowledge transfer (FKT) is a powerful approach to enhance gene function prediction and annotation accuracy.
• FKT effectively complements existing bioinformatics techniques, enabling better integration of prior knowledge for targeted experimental validation.
• Experimental validation in Danio rerio confirmed the predicted role of wnt5b in heart development, highlighting the practical utility of FKT.