RabbitMash

RabbitMash accelerates hash-based genome analysis by optimizing Mash's sketching and distance operations to compute genomic sketches and distances for large-scale datasets.

Key Features:

• Multi-threading: Distributes computational tasks across multiple CPU cores to reduce processing time.
• Vectorization: Employs vectorized operations to accelerate data handling and computation.
• Fast I/O: Optimizes input/output operations to minimize bottlenecks when processing large genomic datasets.
• Sketch operation: Achieves a minimum speedup of 1.3 times relative to Mash.
• Dist operation: Exhibits a speedup of at least 9.8 times relative to Mash.
• Triangle operation: Provides a speedup of 8.5 times relative to Mash.
• Screen operation: Offers a speedup of 4.4 times relative to Mash.
• All-versus-all distances: Computes all-versus-all distances for 100,321 genomes in under five minutes on a 40-core workstation, compared to over forty minutes for Mash.

Scientific Applications:

• Clustering: Enables rapid clustering of genomes using computed distance matrices.
• Assembly-related analyses: Provides fast similarity and distance estimates to support assembly workflows.
• Large-scale dataset analysis: Facilitates analysis of extensive genomic datasets by accelerating distance computations.

Methodology:

Implements hash-based genome analysis by providing optimized implementations of Mash's sketch, dist, triangle, and screen operations using multi-threading, vectorization, and fast I/O.