Accurate annotation of protein coding sequences with IDTAXA.

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Protein sequence data grows faster than functional knowledge.
• Current protein annotation relies on homology searches (BLAST, HMMER).
• Existing methods may misclassify novel protein functions.

Purpose of the Study:

• Develop a novel protein classification algorithm.
• Improve accuracy and reliability in genome annotation.
• Identify novel protein-function associations, including antibiotic resistance.

Main Methods:

• Developed IDTAXA, a novel sequence-based protein classification algorithm.
• Compared IDTAXA performance against BLAST and HMMER for KEGG ortholog assignment.
• Applied IDTAXA for eukaryotic and prokaryotic genome annotation and contamination detection.
• Re-annotated microbial genomes to investigate antibiotic resistance phenotypes.

Main Results:

• IDTAXA demonstrated higher accuracy than BLAST and HMMER in assigning KEGG ortholog groups.
• IDTAXA effectively distinguished novel functions, avoiding common misclassification errors.
• Successfully applied to multi-level genome annotation and detection of eukaryotic genome contamination.
• Discovered two novel associations between proteins and antibiotic resistance in microbial genomes.

Conclusions:

• IDTAXA offers a more accurate and reliable method for protein function classification.
• The algorithm enhances genome annotation pipelines and aids in discovering biological insights.
• IDTAXA has practical applications in identifying genome contamination and exploring genotype-phenotype relationships.