kmtricks

kmtricks performs joint k-mer counting and constructs Bloom filters from large sequencing data collections to enable efficient indexing and improved representation of metagenomic and high-throughput sequencing datasets.

Key Features:

• Efficient k-mer counting across multiple samples: Performs joint counting of k-mers across multiple samples by counting, partitioning, and sorting hash values.
• Enhanced preservation of low-abundance k-mers: Leverages joint k-mer counting to retain low-abundance, non-erroneous k-mers across samples and reduce their loss during Bloom filter construction.
• Improved accuracy in metagenomics: Reduces inclusion of erroneous k-mers while preserving more non-erroneous ones, improving the fidelity of Bloom filter-based representations for metagenomic datasets.
• Accelerated Bloom filter construction: Streamlines hash-based counting and partitioning to construct Bloom filters approximately four times faster than existing state-of-the-art tools.

Scientific Applications:

• Metagenomics: Indexing and representation of complex microbial communities from metagenomic sequencing data using Bloom filters.
• Large-scale sequencing data indexing: Building scalable Bloom filter indexes for large collections of high-throughput sequencing datasets.

Methodology:

Jointly counts k-mers across multiple samples by counting, partitioning, and sorting hash values, and uses these hash partitions to construct Bloom filters.