MASTR: multiple alignment and structure prediction of non-coding RNAs using simulated annealing

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• The discovery of non-coding RNAs (ncRNAs) highlights the need for advanced RNA analysis methods.
• RNA structure is critical for function and often more conserved than sequence, necessitating accurate prediction tools.
• Existing methods often struggle to integrate multiple sequence alignment with RNA structure prediction.

Purpose of the Study:

• To develop a novel computational method for simultaneous prediction of RNA structure and multiple sequence alignment.
• To address the limitations of current tools in analyzing conserved RNA structural elements.

Main Methods:

• Implementation of a novel algorithm, MASTR (Multiple Alignment of STructural RNAs).
• Utilizes Markov chain Monte Carlo within a simulated annealing framework.
• Minimizes a combined cost function incorporating sequence conservation, covariation, and basepairing probabilities.

Main Results:

• MASTR iteratively refines both RNA sequence alignment and structure prediction.
• The algorithm demonstrates competitive accuracy and computational efficiency compared to existing programs.
• Successfully integrates multiple sequence alignment with RNA structure prediction.

Conclusions:

• MASTR offers a powerful new tool for analyzing RNA sequences and structures.
• The method advances the study of ncRNAs by enabling simultaneous alignment and structure prediction.
• MASTR provides a valuable resource for researchers in RNA biology and bioinformatics.