Contact-Assisted Threading in Low-Homology Protein Modeling

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein structure prediction

Background:

• Protein structure prediction from amino acid sequence is a significant challenge in bioinformatics.
• Advancements in deep learning and large-scale protein sequence databases have accelerated progress.
• Accurate prediction of inter-residue contacts is crucial for determining protein 3D structure.

Purpose of the Study:

• To provide an overview of contact-assisted threading methods for protein structure prediction.
• To highlight recent advancements in contact-assisted threading techniques.
• To discuss limitations and future directions in applying contact-assisted threading for low-homology protein modeling.

Main Methods:

• Utilizing deep learning-based inter-residue contact map predictors.
• Employing contact map threading for structure prediction.
• Analyzing existing contact-assisted threading methodologies.

Main Results:

• Deep learning contact map predictors have significantly improved accuracy.
• Contact-assisted threading enables accurate structure prediction, even for proteins with low homology.
• The integration of contact information enhances protein modeling capabilities.

Conclusions:

• Contact-assisted threading is a powerful approach for improving protein structure prediction accuracy.
• Further research is needed to address current limitations and explore future prospects.
• This method is particularly valuable for proteins not amenable to direct homology modeling.