Protein folding rates estimated from contact predictions

Area of Science:

• Computational biology
• Protein folding dynamics
• Structural bioinformatics

Background:

• Protein folding rates are crucial for protein function and stability.
• Existing methods predict folding rates from known 3D structures or sequence-derived secondary structures.
• Predicting folding rates for proteins with unknown structures remains a challenge.

Purpose of the Study:

• To develop a novel method for estimating two-state protein folding rates using predicted internal residue-residue contacts.
• To assess the accuracy of folding rate predictions based on predicted contacts compared to known structures.
• To explore the combination of different predictive methods for improved accuracy.

Main Methods:

• Utilizing predicted internal residue-residue contacts for proteins with unknown structures.
• Correlating folding rates with the number of predicted long-range contacts, normalized by protein length squared.
• Comparing predictions from noisy contact data with those from known contacts and secondary structure predictions.

Main Results:

• Folding rates can be accurately estimated from predicted long-range contacts, even with noisy predictions.
• The accuracy of estimates based on predicted contacts is comparable to estimates from known structures.
• Combining predicted contacts with secondary structure predictions improved performance in distinguishing fast and slow two-state folders.

Conclusions:

• Predicted internal residue-residue contacts offer a viable approach for estimating protein folding rates.
• This method significantly advances the ability to predict folding kinetics for proteins with unknown structures.
• The integration of diverse predictive strategies enhances the reliability of protein folding rate estimations.