A computational Grid framework for immunological applications.

Area of Science:

• Computational Biology
• Immunoinformatics
• Grid Computing

Background:

• The European Union ImmunoGrid project aims to simulate the immune system across multiple levels.
• Consortium members possess distributed computational resources across the Atlantic.
• Existing Grid middleware is leveraged to unify diverse computing environments.

Purpose of the Study:

• To develop a unified computational Grid for accessing distributed resources.
• To facilitate large-scale simulations of the immune system.
• To enhance T-cell epitope prediction and molecular dynamics calculations.

Main Methods:

• Developed a computational Grid with a single interface for resource access.
• Utilized existing Grid middleware for unified exploitation of heterogeneous computing resources.
• Employed neural networks for T-cell epitope prediction and performed molecular dynamics free-energy calculations.

Main Results:

• Generated over 14 million high-quality protein-peptide binding predictions for influenza strains.
• Mapped predictions onto 3D protein structures.
• Executed batches of 120 molecular dynamics free-energy calculations for new prediction method development.

Conclusions:

• The developed Grid successfully integrates diverse computational resources for complex biological simulations.
• The platform enables large-scale immunoinformatics analyses, including extensive T-cell epitope prediction.
• The Grid supports both established and novel computational methods for immune system research.