Domain fishing: a first step in protein comparative modelling

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein modeling

Background:

• Protein comparative modeling relies on identifying suitable structural templates.
• Existing methods may yield suboptimal template identification.
• Accurate template searching is crucial for reliable protein structure prediction.

Purpose of the Study:

• To enhance the efficiency and accuracy of structural template searching in protein comparative modeling.
• To develop a robust method for filtering and ranking potential protein structural templates.

Main Methods:

• Splitting query protein sequences into distinct domains.
• Extracting initial template candidates from databases like PFAM, PDB, and SCOP.
• Ranking templates based on sequence identity (less than 30% triggers further filtering), coverage, and resolution.
• Utilizing secondary structure matching to filter false positive templates when sequence identity is low.

Main Results:

• A refined list of structural templates is generated for protein comparative modeling.
• Improved filtering of irrelevant templates, particularly those with low sequence identity.
• Enhanced reliability of template selection for downstream modeling.

Conclusions:

• The domain-based approach combined with multi-criteria ranking and secondary structure matching improves template identification.
• This method offers a more optimized search strategy for protein comparative modeling.
• The approach contributes to more accurate protein structure prediction through better template selection.