CURC

CURC compresses high-throughput sequencing FASTQ files containing nucleotide sequences and quality scores using a CUDA-accelerated, reference-free, lossless pseudogenome approach to improve storage and transfer efficiency.

Key Features:

• CUDA acceleration: Uses the CUDA library to perform compression computations on GPUs.
• Heterogeneous parallel scheme: Combines GPU and CPU resources for parallelized processing.
• Reference-free compression: Operates without requiring an external reference genome.
• Lossless pseudogenome approach: Implements a pseudogenome-based method to ensure no data loss during compression.
• FASTQ-focused: Targets FASTQ files that store nucleotide sequences and quality scores from high-throughput sequencing (HTS).
• Speedup versus CPU-based compressors: Reports a 2–6-fold acceleration in compression times compared to traditional CPU-only compressors.
• Competitive compression rates: Achieves compression ratios comparable to state-of-the-art reference-free read compressors.

Scientific Applications:

• HTS data storage and transfer: Reduces storage footprint and bandwidth requirements for high-throughput sequencing datasets.
• Large-scale sequencing projects: Enables more efficient handling and archiving of large sequencing runs and cohort datasets.
• Preservation of sequencing information: Maintains nucleotide sequences and quality scores for downstream genomic analyses through lossless compression.

Methodology:

CURC implements a reference-free, lossless pseudogenome-based compression of FASTQ reads using the CUDA library for GPU acceleration within a heterogeneous parallel scheme that combines GPU and CPU resources.