UNCALLED

UNCALLED performs real-time mapping of streaming nanopore raw electrical signals to reference sequences to enable targeted sequencing and enrichment without basecalling.

Key Features:

• Real-Time Signal Mapping: Directly matches streaming nanopore current signals to a reference sequence without prior basecalling.
• Probabilistic k-mer Analysis: Employs a probabilistic model to consider candidate k-mers represented by raw signals and narrow candidate sequences.
• FM-index Integration: Uses an FM-index encoded with reference sequences to prune unlikely candidates and accelerate mapping.

Scientific Applications:

• Targeted Sequencing in Metagenomics: Enables selective depletion of known bacterial genomes from community samples, producing a 4.46-fold enrichment of other species.
• Enrichment and Detection in Human Genomics: Has been applied to enrich genes associated with hereditary cancers, achieving 29.6x coverage of target regions on a single MinION flowcell and enhancing detection of SNPs, indels, structural variants (SVs), and methylation.
• Superior Variant Detection: Identified twice as many structural variants compared to Illumina at 50x coverage, with findings corroborated by whole-genome sequencing using nanopore and PacBio HiFi data.

Methodology:

Processes raw nanopore electrical signals in real time, applies probabilistic k-mer analysis to propose candidate k-mers, and prunes candidates using an FM-index-encoded reference to map signals to reference sequences without basecalling.