Managing Complex Workflows in Bioinformatics: An Interactive Toolkit With GPU Acceleration.

Area of Science:

• Bioinformatics and Computational Biology
• High-Performance Computing (HPC)
• Genomics and Big Data Analytics

Background:

• Bioinformatics research is rapidly expanding, generating massive biological datasets.
• Processing these large datasets requires high-performance and high-throughput computing.
• Hardware accelerators like GPUs and distributed computing offer potential for faster data processing.

Purpose of the Study:

• To introduce BioWorkflow, an interactive workflow management system.
• To automate bioinformatics analyses using parallel task scheduling.
• To evaluate the performance of BioWorkflow with GPU-accelerated and distributed computing.

Main Methods:

• Development of BioWorkflow, an interactive workflow management system.
• Implementation of parallel task scheduling utilizing GPU-acceleration and distributed computing.
• Evaluation through a case study of a complex, branching workflow for multiple sequence alignment.

Main Results:

• Significant speed-ups achieved: x2.89 with GPUs and x2.832 average speed-up from parallel node execution.
• Combined speed-up of x1.71 for complex workflows, demonstrating benefits over sequential execution.
• BioWorkflow achieved a System Usability Scale score of 82.9, indicating high usability.

Conclusions:

• BioWorkflow effectively accelerates bioinformatics analyses through GPU-acceleration and concurrent node execution.
• The system demonstrates substantial performance gains for big data processing in bioinformatics.
• The interactive user interface enhances usability for managing complex bioinformatics workflows.