Lighter

Lighter corrects sequencing errors in high-throughput DNA sequencing reads using a sampling-based k-mer approach and dual Bloom filters to improve downstream analyses such as genome assembly and variant calling.

Key Features:

• Sampling-based k-mer error detection: Samples a representative subset of k-mers instead of counting all k-mers to identify candidate errors while reducing computational overhead.
• Dual Bloom filters: Uses two Bloom filters, one to store sampled input k-mers and a second to retain k-mers judged likely correct.
• Memory efficiency: Achieves an approximately constant memory footprint independent of sequencing depth when the sampling fraction is appropriately adjusted.
• Parallelization and speed: Is designed for parallel execution to increase processing speed on large sequencing datasets.
• Accuracy maintenance: Maintains near-constant error-correction accuracy across varying sequencing depths by adjusting the sampling fraction.

Scientific Applications:

• Whole-genome sequencing error correction: Reduces sequencing errors in whole-genome and other high-throughput sequencing reads.
• Genome assembly: Improves assembly accuracy by supplying error-reduced reads to assemblers.
• Variant calling: Enhances variant-calling reliability by lowering false positives caused by sequencing errors.
• Comparative genomics: Provides higher-quality read data to support comparative analyses across genomes.
• Evolutionary studies: Supplies cleaner sequence data for evolutionary inference and downstream analyses.

Methodology:

Lighter samples k-mers from input reads and stores them in one Bloom filter, uses a second Bloom filter to retain k-mers likely to be correct based on prevalence and consistency across samples, identifies erroneous k-mers via the dual-filter comparison and corrects reads without exhaustive k-mer counting, and adjusts the sampling fraction relative to sequencing depth; the method is amenable to parallel execution.