MultiNanopolish

MultiNanopolish accelerates signal-level analysis of Oxford Nanopore sequencing data to improve consensus polishing and the detection of genetic variants and DNA methylation.

Key Features:

• Improved Consensus Sequencing: Refines consensus sequences from draft genomes using signal-level analysis to increase assembly accuracy.
• Variant and Methylation Detection: Detects genetic variants and DNA methylation patterns using a hidden Markov model (HMM) framework applied to raw nanopore signals.
• Parallel Processing with GroupTasks: Decomposes iterative calculations into independent "GroupTasks" that are distributed across a thread pool for concurrent multi-threaded computation.
• Performance Improvements: Reduces running times by approximately 50% with read-uncorrected assemblers such as Miniasm and by about 20% with read-corrected assemblers such as Canu and Flye under a 40-thread configuration.

Scientific Applications:

• Genomics: Enhances genome assembly polishing and accuracy for long-read sequencing projects.
• Epigenetics: Enables detection and analysis of DNA methylation patterns from nanopore signal data.
• Personalized Medicine: Supports variant detection from long-read datasets for applications in clinical genomics and precision medicine.

Methodology:

Performs signal-level analysis using an HMM framework and an iterative calculation strategy that breaks computations into independent "GroupTasks" which are processed in parallel via a thread pool to reduce redundant calculations and execution time.