Hierarchical String Indexing for Fast Genome Sequence Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for analyzing human genomes, such as BWA-MEM, require significant memory resources and suffer from slow processing speeds due to frequent random memory access, limiting the ability to shorten analysis time despite advancements in sequencer technology and the desire for cost-effective high-performance computing.
Innovation Solution
A hierarchical index is created based on a reference character string using a hash function to efficiently identify matching character strings, allowing for faster approximate string comparisons and alignment processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If BWA-MEM algorithm is used for mapping, then mapping speed is improved, but memory resource consumption increases significantly
Solution Approach 1:
The patent segments the reference genome into multiple blocks, creating separate index structures for each block. This divides the large memory-intensive index into smaller, more manageable segments that can be processed and accessed more efficiently, reducing peak memory consumption while maintaining mapping speed.
Solution Approach 2:
The patent introduces a two-dimensional indexing approach with primary indices for block-level access and secondary indices for within-block access. This dimensional organization allows the system to handle large genomes by distributing memory access across multiple levels, reducing the memory footprint required for any single index structure.
2Speed
If BWA-MEM algorithm is used for mapping, then mapping speed is improved, but processing speed is limited by frequent random memory access
Solution Approach 1:
The patent performs preliminary organization of the reference genome into blocked segments with pre-computed indices before the actual mapping process. This preliminary structuring enables more predictable and efficient memory access patterns during mapping, reducing random access overhead and improving processing speed.
Solution Approach 2:
By segmenting the genome into blocks with organized indices, the patent enables more sequential and predictable memory access patterns compared to the random access required by traditional single-index methods. This segmentation reduces cache misses and memory access time.
3Measurement precision
If redundant data is used for statistical processing to reduce errors, then accuracy is improved, but calculation load increases
Solution Approach 1:
The patent segments both the reference genome and query sequences into manageable blocks, enabling parallel processing of redundant data. This segmentation allows statistical processing to be distributed across multiple computational units, maintaining accuracy through comprehensive analysis while improving overall calculation speed through parallelization.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a method of retrieving an approximate character string in a reference character string based on a query character string. The method includes creating a hierarchical index based on the reference character string; performing mapping of a query character string with respect to the reference character string by referring to the hierarchical index to identify a partial character string in the reference character string that matches at least a part of the query character string; and deriving the approximate character string based on at least one or more of the partial character strings identified by the mapping. The hierarchical index is created by repeating adding and sorting additional keys to respective keys cut out from the reference character string in accordance with the number of times of appearance of the cut-out keys.