Gene function hypotheses for the Campylobacter jejuni glycome generated by a logic-based approach

Area of Science:

• Computational biology
• Genomics
• Systems biology

Background:

• Biological systems generate vast experimental data requiring computational integration for modeling.
• Understanding gene function is crucial for deciphering biological pathways.
• The capsular polysaccharide biosynthetic pathway in Campylobacter jejuni is a key area of study.

Purpose of the Study:

• To demonstrate a logic-based machine learning approach for proposing gene function hypotheses.
• To integrate diverse experimental data for biological system modeling.
• To investigate gene function in the Campylobacter jejuni capsular polysaccharide pathway.

Main Methods:

• Utilizing inductive logic programming (ILP) for automated hypothesis generation.
• Integrating data from gene knockout experiments and gene presence/absence across strains.
• Employing pathway structure as background knowledge for the machine learning model.

Main Results:

• Successfully proposed gene assignments for five previously unassigned reaction steps in the pathway.
• Achieved unambiguous gene-to-reaction assignments for four steps.
• Identified three candidate genes for the fifth unassigned step, with some assignments supported by further experimental data.

Conclusions:

• Logic-based machine learning offers a robust strategy for integrating diverse experimental data.
• This methodology effectively proposes testable hypotheses for biological system behavior.
• The study advances understanding of the Campylobacter jejuni capsular polysaccharide pathway.