Association algorithm to mine the rules that govern enzyme definition and to classify protein sequences

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Genome projects generate vast amounts of uncharacterized sequences.
• A need exists for automated annotation schemes to bridge the gap between sequence production and functional annotation.
• This study focuses on developing rules for automatic fungus gene classification.

Purpose of the Study:

• To elucidate enzyme classification rules from the UniProt protein knowledgebase.
• To apply these rules for automatic classification of fungus genes.
• To evaluate the efficacy of the derived rules for functional annotation.

Main Methods:

• Utilized the Apriori association algorithm to mine relationships between enzyme classes and InterPro entries.
• Employed Swiss-Prot datasets as training sets for rule establishment.
• Applied derived rules to TrEMBL entries for testing and classification.

Main Results:

• Achieved a 70% correct enzyme classification rate for prokaryote datasets and 80% for eukaryote datasets.
• Successfully matched 88 out of 5085 poorly annotated fungus entries using the developed rule set.
• Demonstrated the capacity of candidate rules for accurate classification.

Conclusions:

• The presented method is feasible for classifying enzyme classes based on enzyme domain rules.
• The derived rules can assist protein annotators in manual annotation processes.
• The rules are suitable for implementation in automated annotation workflows.