GTfold: enabling parallel RNA secondary structure prediction on multi-core desktops

Area of Science:

• Computational molecular biology
• Bioinformatics
• Structural biology

Background:

• Accurate RNA secondary structure prediction is a key challenge in computational molecular biology.
• Previous parallelized prediction tools were often inaccessible on high-performance computers.
• Modern multi-core processors necessitate new parallel prediction program designs.

Purpose of the Study:

• To develop and present the first RNA secondary structure prediction program optimized for modern multi-core computers.
• To leverage thermodynamic optimization for enhanced prediction capabilities.

Main Methods:

• Implementation of a novel parallelized RNA secondary structure prediction algorithm.
• Utilizing thermodynamic optimization principles.
• Testing on multi-core computer architectures.

Main Results:

• GTfold achieves faster RNA secondary structure prediction compared to UNAfold and RNAfold on machines with four or more cores.
• GTfold maintains prediction accuracy while reducing computational time.
• The program demonstrates efficient performance on contemporary multi-core hardware.

Conclusions:

• GTfold advances RNA structural biology by significantly reducing prediction timescales.
• The speed improvements are particularly beneficial for analyzing lengthy RNA sequences, including viral genomes.
• GTfold enhances accessibility of efficient RNA structure prediction for researchers.