Functional coherence metrics in protein families

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Biological sequences like proteins are annotated with functional information.
• Incomplete protein annotations limit the assessment of functional coherence within protein sets.
• Existing semantic similarity measures capture only limited functional aspects due to annotation incompleteness.

Purpose of the Study:

• To propose a comprehensive approach for assessing functional coherence in protein sets.
• To evaluate two novel functional coherence metrics, mUI and mGIC.
• To improve the understanding of functional annotation aspects: completeness, agreement, and coherence.

Main Methods:

• Developing and applying novel functional coherence metrics (mUI and mGIC).
• Combining semantic similarity measures with term enrichment techniques.
• Utilizing visualization tools anchored in domain-specific knowledge (e.g., protein families).

Main Results:

• The proposed metrics effectively capture and measure local similarity cores within protein sets.
• The metrics provide an improved grasp on functional annotation completeness, agreement, and coherence.
• Visualization tools enhance the interpretation of functional coherence analysis.

Conclusions:

• Measuring functional similarity in proteins is complex due to inherent data characteristics and annotation processes.
• Single metrics are insufficient for comprehensive functional coherence assessment of protein sets.
• Understanding the behavior and limitations of different similarity metrics is crucial for robust analysis.