Benchmarking Methods of Protein Structure Alignment

Area of Science:

• Structural Bioinformatics
• Computational Biology
• Biochemistry

Background:

• Protein function is intrinsically linked to its structure and amino acid sequence.
• Accurate protein structure alignment is crucial for understanding protein domain relationships and biological function.
• Numerous protein structure alignment (PSA) methods exist, employing diverse techniques.

Purpose of the Study:

• To evaluate the efficacy of various PSA methods in identifying protein domains with varying structural similarities.
• To assess the utility of PSA methods for clustering protein domains into known structural groups.
• To provide insights for researchers selecting or developing PSA tools.

Main Methods:

• Comprehensive investigation of eighteen distinct PSA methods.
• Separation of alignment algorithms from similarity scoring functions to assess their independent effectiveness.
• Evaluation of method performance in both classification (identifying similar pairs) and clustering (grouping domains).

Main Results:

• SP-AlignNS (non-sequential) demonstrated superior performance in protein domain classification.
• SP-AlignNS was also among the top performers for domain clustering.
• Hybrid combinations of algorithms and scores, even when mismatched, sometimes outperformed native methods in classification and clustering.

Conclusions:

• The study offers a large-scale independent assessment of PSA method capabilities.
• SP-AlignNS is recommended for protein domain classification tasks.
• The effectiveness of PSA methods can be influenced by the interplay between alignment algorithms and scoring functions, suggesting potential for novel hybrid approaches.