Structure Modeling Protocols for Protein Multimer and RNA in CASP16 With Enhanced MSAs, Model Ranking, and Deep...

Area of Science:

• Computational biology
• Structural biology
• Bioinformatics

Background:

• Accurate prediction of protein complex and RNA structures is crucial for understanding biological functions.
• CASP16 provides a benchmark for evaluating the performance of structure prediction methods.

Purpose of the Study:

• To present the methods and results of KiharaLab's protein complex and RNA structure predictions at CASP16.
• To evaluate the effectiveness of integrated deep learning models, enhanced multiple sequence alignments (MSAs), and consensus scoring strategies.

Main Methods:

• Integration of multiple state-of-the-art deep learning models with consensus-based scoring.
• Enhancement of MSAs using a large metagenomic sequence database.
• Ensemble approach for RNA structure prediction centered around the NuFold framework.
• Manual refinement of predictions based on literature evidence.

Main Results:

• KiharaLab group ranked first in protein complex prediction.
• KiharaLab group ranked third in RNA structure prediction.
• Analysis revealed strengths in MSA and scoring strategies, alongside areas for future development.

Conclusions:

• The developed computational approach demonstrates high accuracy in protein complex and RNA structure prediction.
• The study highlights the importance of deep learning, comprehensive MSAs, and robust scoring for structural predictions.
• Further improvements are identified for future iterations of the prediction models.