parMatt

parMatt performs parallel multiple alignment of protein three-dimensional (3D) structures to enable comparative structural analyses of homologous proteins.

Key Features:

• Parallelization: Hybrid parallel re-implementation of the MATT algorithm using MPI, pthreads, and OpenMP to accelerate alignment on distributed-memory systems and multi-core processors.
• Translation and Twist Adjustments: Permits translations and twists during alignment to accommodate conformational and evolutionary structural changes.
• Scalability: Distributes computational tasks across multiple nodes to handle large datasets, including extensive protein families and superfamilies.

Scientific Applications:

• Structure–function analysis: Enables accurate structural alignments to study structure–function relationships of proteins.
• Evolutionary biology and functional genomics: Supports comparative analyses that reveal structural variation among evolutionarily distant homologues for evolutionary and functional inference.
• Large-scale protein family comparison: Facilitates analysis of extensive protein families and superfamilies by leveraging distributed-memory parallelism.

Methodology:

Computational steps explicitly include initial construction of pairwise alignments parallelized with MPI and pthreads, iterative progression and refinement of the multiple alignment benefiting from parallelization, and a final refinement step optimized with OpenMP.