Functional annotation by sequence-weighted structure alignments: statistical analysis and case studies from the...

Area of Science:

• Structural biology
• Bioinformatics
• Computational biology

Background:

• Functional annotation of proteins is crucial for understanding biological processes.
• Structural genomics projects generate large datasets of protein structures requiring efficient annotation methods.

Purpose of the Study:

• To develop and validate a novel structural alignment scoring function for accurate functional annotation of protein structures.
• To apply the developed method to a large set of uncharacterized proteins from a structural genomics project.

Main Methods:

• Development of a novel scoring function weighting structurally aligned residue pairs using position-specific scoring matrices.
• Optimization of the scoring function using Pfam, CATH, and SCOP databases.
• Application of the optimized score to functionally annotate 230 query proteins from the Protein 3000 structural genomics project.

Main Results:

• The optimized score achieved an area of 0.87 under the receiver-operating characteristic curve for identifying Pfam families.
• The method successfully annotated 78 query proteins with high confidence (same Pfam family or >30% sequence identity).
• An additional 49 proteins were matched to more distantly related templates, with the predicted closest functional relative sometimes differing from the most structurally similar template.

Conclusions:

• The novel scoring function effectively enhances the functional annotation of protein structures, particularly for identifying conserved evolutionary motifs.
• The method demonstrates utility in structural genomics by assigning functions to previously uncharacterized proteins, including those with distant relationships.
• A significant number of proteins remain uncharacterized at the family or superfamily level, highlighting the need for further refinement of annotation strategies.