REINDEER

REINDEER: K-mer Indexing and Abundance Analysis Across Sequencing Datasets

REINDEER indexes k-mers and their abundances across large collections of sequencing datasets, including raw RNA-seq and metagenomic reads, enabling both presence/absence queries and quantitative abundance retrieval at scale.

Key Features:

• Compacted de Bruijn Graph Indexing: Constructs a compacted de Bruijn graph (DBG) for each dataset and merges them into a global DBG to index k-mers across multiple datasets.
• Abundance Recording: Stores k-mer abundance information in addition to presence/absence data for quantitative sequence analysis.
• Monotig Construction: Builds and indexes monotigs, defined as groups of k-mers with similar abundances, to improve query efficiency and accuracy.
• Scalability: Indexed 2,585 human RNA-seq experiments comprising 4 billion distinct k-mers in 45 hours using 56 GB of RAM.
• Resource Efficiency: Maintains a compact index and low memory footprint during indexing and querying.
• Flexible Querying: Supports querying directly from disk or from RAM for optimized performance based on system resources.

Scientific Applications:

• Large-Scale Sequence Analysis: Enables comparative RNA-seq and metagenomic studies by supporting efficient cross-dataset k-mer abundance queries.

Methodology:

Constructs a compacted de Bruijn graph for each sequencing dataset, merges these into a unified global DBG, and organizes k-mers into monotigs based on shared abundance profiles to enable efficient indexing and rapid quantitative queries.