Quantitative dimethyl sulfate mapping for automated RNA secondary structure inference

Area of Science:

• Molecular Biology
• Biochemistry
• Bioinformatics

Background:

• Dimethyl sulfate (DMS) mapping is a long-established technique for RNA structure analysis.
• Manual modeling of RNA secondary structure has traditionally relied on DMS mapping data.
• Automated RNA structure prediction methods are increasingly important for biological research.

Purpose of the Study:

• To integrate dimethyl sulfate (DMS) mapping data into an automated RNA secondary structure inference framework.
• To evaluate the performance of DMS-guided RNA structure modeling compared to existing methods like SHAPE.
• To establish DMS mapping as a quantitative tool for unbiased RNA secondary structure modeling.

Main Methods:

• Incorporation of DMS reactivity data into an energy minimization framework.
• Development of automated secondary structure inference algorithms.
• Validation using six noncoding RNAs with known crystallographic structures.
• Application of bootstrapping for confidence estimation in structural predictions.

Main Results:

• DMS-guided automated modeling achieved low false negative (9.5%) and false discovery (11.6%) rates.
• Performance was comparable or superior to SHAPE-guided modeling for RNA secondary structure prediction.
• Integration of DMS-SHAPE and CMCT reactivities offered minor improvements in accuracy.
• Bootstrapping provided reliable confidence estimates for the predicted RNA structures.

Conclusions:

• Dimethyl sulfate (DMS) mapping can be effectively utilized for automated, quantitative RNA secondary structure modeling.
• The developed framework offers an unbiased approach to RNA structure prediction.
• DMS mapping is a valuable addition to the toolkit for RNA structure biologists, enhancing routine analysis.