Conserved residue clustering and protein structure prediction.

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Conserved residues are vital for protein structure and function.
• Understanding their spatial distribution is key to predicting protein structures.
• Previous analyses of residue clustering had limitations due to sequence and composition biases.

Purpose of the Study:

• To investigate the 3D clustering of evolutionarily conserved protein residues.
• To differentiate structural clustering from sequence and compositional effects.
• To assess the utility of conserved residue clustering in de novo protein structure prediction.

Main Methods:

• Analysis of conserved positions in multiple sequence alignments for 79 proteins.
• Comparison of residue clustering in native structures versus de novo predicted conformations (Rosetta).
• Evaluation of clustering in native structures against alternative conformations to isolate structural effects.

Main Results:

• 92% of proteins showed significantly higher clustering of conserved residues than random sets.
• 65% of proteins exhibited greater conserved residue clustering in native structures compared to Rosetta-generated conformations.
• 79% of proteins benefited from selecting Rosetta conformations with maximal conserved residue clustering, enriching near-native structures.

Conclusions:

• Conserved residues display significant 3D clustering in native protein structures, exceeding expectations from sequence locality and composition.
• This structural clustering provides valuable information for improving de novo protein structure prediction.
• The spatial organization of conserved residues is a predictable feature exploitable in computational protein design.