.png){kind=icon}
.png){kind=icon}
Return
Enhancing RNA 3D Structure Prediction: A Hybrid Approach Combining Expert Knowledge and Computational Tools in CASP16
.png){kind=icon}
PRE
AI
B
Y
L
DOI：10.1002/prot.70034.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
RNA three-dimensional structures are critical for their roles in gene expression and regulation. However, predicting RNA structures remains challenging due to complex tertiary interactions, ion dependency, molecular flexibility, and the limited availability of known 3D structures. To address these challenges, our team (GuangzhouRNA-human) employed a hybrid strategy combining computational tools with expert refinement in the CASP16 RNA structure prediction challenge, achieving second place based on the sum Z-score. Our approach integrates multiple techniques through modular workflows, including template-based modeling for targets with homologous templates and ab initio prediction using deep learning tools (e.g., AlphaFold3 and DeepFoldRNA) for novel sequences. Additionally, we incorporate experimental constraints and iterative optimization to enhance prediction accuracy. For targets shorter than 200 nucleotides (nt) with homologous templates, our method demonstrated exceptional performance, achieving 75% of predictions with root-mean-square deviations (RMSD) below 5 Å, and all predictions falling under 10 Å. Furthermore, our strategy demonstrated promising results for targets without homologous templates, such as R1209, through comprehensive literature reviews and structural selection. Despite these advances, RNA structure prediction continues to face challenges, particularly in predicting complex topologies like pseudoknots and coaxial stacking. Future improvements in integrating computational tools with expert knowledge are essential to enhance the accuracy and applicability of RNA tertiary structure prediction.
Keywords:
conformational diversity
de novo prediction
homology modeling
RNA structure
structure prediction
tertiary interactions
Journal
P
IF:
2.8
Papers:
6.6K
Citations:
1.4W
Organization
Cited Papers