The use of grid computing to drive data-intensive genetic research

Area of Science:

• Genetics and Bioinformatics
• Computational Biology
• Genomic Analysis

Background:

• Increasing genetic data and complex algorithms necessitate greater computational capacity.
• Traditional linkage analysis methods, like the Lander-Green hidden Markov model (HMM), face scalability challenges with larger pedigrees.
• Whole genome linkage analysis requires extensive genotype simulations, often too time-consuming for single computers.

Purpose of the Study:

• To develop and evaluate "Grid-Allegro", a Grid-enabled implementation of the Allegro software.
• To enable parallel execution of thousands of genotype simulations for efficient genome-wide linkage analysis.
• To demonstrate the cost-effectiveness and scalability of distributed computing for bioinformatics tasks.

Main Methods:

• Developed "Grid-Allegro" by adapting the Allegro software for Grid computing environments.
• Implemented temporary installations of executables and datasets on remote nodes to bypass the need for predefined Grid run-time environments.
• Evaluated performance, efficiency, and scalability using genome scans on Swedish multiplex Alzheimer's disease families.

Main Results:

• "Grid-Allegro" successfully performed thousands of genotype simulations in parallel, significantly reducing computation time.
• The implementation demonstrated full exploitation of Allegro's features for genome-wide linkage analysis.
• The study confirmed the efficiency and scalability of Grid-Allegro for complex genetic studies.

Conclusions:

• "Grid-Allegro" provides a cost-effective and scalable solution for computationally intensive genetic analyses.
• Distributed computing on Grids offers a viable alternative to in-house cluster computing for most geneticists.
• This approach facilitates advanced genomic research, particularly for complex diseases.