RawHash

RawHash maps raw nanopore sequencing signals to reference genomes using a hash-based similarity search to enable real-time analysis of large genomes.

Key Features:

• Real-Time Signal Analysis: Processes electrical signals produced by nanopore sequencers as they are generated to enable immediate signal-level analysis.
• Read Until Optimization: Leverages the "Read Until" capability of nanopore sequencing to allow early ejection of DNA strands and reduce sequencing of unnecessary fragments.
• Hash-Based Similarity Search: Quantizes raw signal values and computes hashes so that signals from identical DNA yield the same hash despite minor signal variations.
• Scalability for Large Genomes: Maintains high throughput and accuracy for large genomes, reporting a 25.8× average throughput improvement compared to UNCALLED and improved accuracy relative to Sigmap.

Scientific Applications:

• Read Mapping: Aligns raw nanopore signals to a reference genome, achieving a 25.8× average throughput improvement over UNCALLED and improved accuracy relative to Sigmap for large genomes.
• Relative Abundance Estimation: Estimates relative abundances of genomic sequences within samples from raw-signal mappings to support microbial community and gene expression analyses.
• Contamination Analysis: Detects and characterizes contaminant DNA in sequencing data using raw-signal mappings to the reference.

Methodology:

Creates an index from the reference genome, quantizes raw nanopore signals to consistent values, computes hashes from quantized signals, and performs hash-based similarity search to map raw signals to the reference genome.