RLSFmode: A deep learning approach for predicting RNA-small molecule binding modes via molecular surface modeling.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• RNA is an emerging therapeutic target, making RNA-small molecule binding prediction crucial for drug development.
• Accurate binding mode and affinity prediction are essential for rational drug design and efficient candidate screening.
• Existing computational models face challenges due to RNA binding pocket complexity and limited structural data.

Purpose of the Study:

• To develop a novel deep learning model for predicting RNA-small molecule binding modes and affinity.
• To address the limitations of existing computational methods in RNA-ligand interaction prediction.

Main Methods:

• Developed a deep learning model named RLSFmode.
• Utilized molecular surface modeling, RNA-ligand surface fingerprinting, sequence information, and energy scoring functions.
• Combined these features to predict binding modes and estimate binding affinity.

Main Results:

• RLSFmode demonstrated significantly improved performance in binding mode prediction.
• The model integrates multiple data types for comprehensive prediction.
• Rigorous testing validated the model's effectiveness.

Conclusions:

• RLSFmode offers a powerful computational tool for predicting RNA-small molecule interactions.
• The model has the potential to accelerate drug discovery by reducing experimental costs and timelines.
• This approach advances structure-based rational drug design for RNA targets.