Plast

Plast performs parallelized local sequence alignment searches to accelerate sequence similarity comparisons of large genomic banks and next-generation sequencing datasets.

Key Features:

• Parallel algorithm design: Implements a novel algorithm tailored to exploit the parallel capabilities of contemporary microprocessor architectures and targeted at middle-range computers.
• SIMD and SSE utilization: Leverages SIMD parallelism via the SSE instruction set to perform simultaneous operations on multiple data elements.
• Multithreading on multicore CPUs: Uses multithreading to distribute work across multiple CPU cores for concurrent processing.
• Performance benchmarking: Comparative tests on an 8-processor server demonstrated a 3–6× speedup relative to multithreaded BLAST while maintaining comparable accuracy.

Scientific Applications:

• Large genomic bank comparison: Accelerates pairwise and database-wide similarity searches across extensive genomic repositories.
• Next-generation sequencing data processing: Enables rapid similarity searches on high-throughput sequencing datasets generated by next-generation sequencing technologies.
• Gene discovery: Facilitates identification of homologous sequences relevant to gene discovery efforts.
• Evolutionary studies and functional genomics: Supports comparative analyses for evolutionary inference and functional annotation.

Methodology:

Employs a parallel algorithm aligned with microprocessor architecture, implements SIMD via the SSE instruction set and multithreading on multicore processors, and was benchmarked on an 8-processor server against multithreaded BLAST.