Computational protein design and large-scale assessment by I-TASSER structure assembly simulations

Area of Science:

• Computational biology
• Protein engineering
• Structural bioinformatics

Background:

• Experimental validation of designed proteins is time-consuming and resource-intensive.
• Advances in protein structure prediction offer potential for automated validation.
• Developing efficient protein design protocols is crucial for creating novel protein functionalities.

Purpose of the Study:

• To present a novel computational protocol for protein design and validation.
• To automate the assessment of designed protein sequences using folding simulations.
• To improve the efficiency and success rate of de novo protein design.

Main Methods:

• Utilized Monte Carlo sampling guided by an atomic potential for sequence space exploration.
• Employed clustering of sequence decoys for candidate sequence selection.
• Assessed designed sequences using I-TASSER folding simulations to generate structural models.

Main Results:

• Tested on 52 nonhomologous proteins with an average sequence identity of 24% to native sequences.
• Predicted 3D models for designed sequences achieved <2 Å RMSD to target structures in 62% of cases (77% for top 10).
• Designed sequences exhibited lower free energy (0.39 kcal/mol/residue) than native sequences, suggesting enhanced stability.

Conclusions:

• The developed protocol effectively captures essential features for global protein folding.
• Automated validation via folding simulations significantly reduces experimental effort.
• The designed sequences show potential for greater stability and accurate structural prediction.