SHIMMER Indexing for Fast Genome Assembly
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current genome assembly methods, such as whole genome shotgun sequencing, require significant computational resources and time due to the need for extensive read-to-read comparisons, especially in the presence of long repetitive sequences, which increases computation time and storage requirements.
Innovation Solution
The implementation of sparse hierarchical minimizer (SHIMMER) indexing, which groups DNA sequencing reads likely to overlap based on shared neighboring base pairs, reducing the number of necessary comparisons and eliminating quadratic computation steps, allowing for faster genome assembly and smaller disk storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional overlap-layout-consensus method is used for genome assembly, then assembly accuracy can be maintained, but computation time and computational resources increase significantly due to quadratic comparisons
Solution Approach 1:
The patent segments the genome assembly process into three distinct phases: (1) generating minimizers from reads, (2) constructing a Burrows-Wheeler Transform index from minimizers, and (3) performing assembly operations using the index. This segmentation allows each phase to be optimized independently, with the indexing phase pre-processing data to enable faster subsequent assembly operations without requiring quadratic comparisons.
Solution Approach 2:
The patent performs preliminary actions by constructing a Burrows-Wheeler Transform index from minimizers before the actual assembly process. This pre-computed index contains all necessary information about minimizer occurrences and positions, allowing the assembly algorithm to quickly retrieve and compare relevant sequences without performing exhaustive pairwise comparisons during the assembly phase itself.
2Reliability
If high redundancy (30-fold or higher) is used for successful de novo assembly, then assembly completeness improves, but the number of reads and computational burden increases
Solution Approach 1:
The patent extracts a compressed representation of the genome data by generating minimizers - a sparse set of representative k-mers that capture the essential sequence information. Instead of working with all reads directly, the method extracts only the necessary minimizer elements to construct the assembly, significantly reducing the data volume while maintaining assembly completeness even at lower redundancy levels.
Solution Approach 2:
The patent changes the fundamental parameter of data representation from full sequences to minimizers (subsequences of length k). This parameter transformation allows the system to work with a compressed subset of the data that retains sufficient information for accurate assembly, enabling successful de novo assembly at lower redundancy levels while reducing computational burden.
3Productivity
If minimizer indexing is used instead of BWT-FM-index, then indexing effectiveness for long DNA fragments improves, but implementation complexity increases
Solution Approach 1:
The patent creates a universal indexing system using Burrows-Wheeler Transform that can handle both short and long DNA fragments effectively. The BWT index constructed from minimizers serves multiple functions: it enables fast exact matching, supports gapped alignment, and works across different read lengths and sequencing technologies, providing a single solution that replaces multiple specialized indexing approaches.
Data Source
AI summary
Methods, systems, and devices are contemplated for assembling a genome from duplicate segments of the genome. Sequences with a first level of common neighboring base pairs are identified and organized into first level groups. Groups are then identified from the first level groups that have a second level of common neighboring base pairs and organized into a number of second level groups. A third level of groups can further be organized in some embodiments. Typically the second level groups are assembled into spans having contiguous base pair sequences, which are then assembled into the broader genome sequence. The inventive subject matter is preferably used for whole genome sequencing.