Predicting RNA Secondary Structure Using In Vitro and In Vivo Data.

Area of Science:

• Computational Biology
• Bioinformatics
• Molecular Biology

Background:

• High-throughput experimental data on RNA secondary structures have advanced computational approaches.
• Predicting in vivo RNA folding is complex due to the crowded cellular environment and interactions with other molecules.

Purpose of the Study:

• To develop and present CROSS and CROSSalive, machine learning algorithms for predicting RNA secondary structure propensity.
• To compare in vitro and in vivo RNA secondary structure prediction using experimental data and Artificial Neural Networks (ANNs).

Main Methods:

• Development of two algorithms, CROSS and CROSSalive, trained on experimental RNA secondary structure data.
• Utilizing Artificial Neural Networks (ANNs) with experimental data as input for prediction.
• Investigating the distinct factors influencing in vitro versus in vivo RNA folding predictions.

Main Results:

• CROSS and CROSSalive algorithms demonstrate the ability to predict RNA secondary structure propensity.
• The study highlights the differences in predicting RNA secondary structure propensity in vitro and in vivo.
• Experimental data successfully trained ANNs for RNA structure prediction.

Conclusions:

• Machine learning algorithms like CROSS and CROSSalive can effectively predict RNA secondary structure propensity.
• Accounting for cellular factors is crucial for accurate in vivo RNA folding predictions.
• Further research into in vivo prediction models is warranted given the complexity of the cellular environment.