Crosstalk between RNA secondary and three-dimensional structure prediction: a comprehensive study.

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Computational methods predict RNA 3D structures.
• RNA secondary (2D) structure aids 3D prediction accuracy and efficiency.
• The impact of 2D structure accuracy on 3D prediction is understudied.

Purpose of the Study:

• Benchmark RNA 3D structure prediction models.
• Investigate the effect of 2D structure accuracy on 3D prediction.
• Explore modification of base-pairing interactions during 3D modeling.

Main Methods:

• Benchmarking six RNA 3D structure prediction models.
• Utilizing datasets with 2D structures of varying accuracies.
• Analyzing the relationship between 2D input accuracy and 3D model performance.

Main Results:

• Significant crosstalk exists between RNA 2D and 3D structure predictions.
• 3D prediction accuracy depends on the model's ability to adjust input base pairs.
• Models are more sensitive to false positive base pairs in 2D structures.

Conclusions:

• RNA 3D structure prediction is influenced by 2D structure accuracy and model flexibility.
• Improving 2D structure prediction, particularly reducing false positives, can enhance 3D modeling.
• Understanding base-pair modification is key to advancing RNA structure prediction.