Genome Read Indexing with Cluster Signatures for Fast Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current genome sequencing technologies lack efficient indexing methods for raw and unmapped sequence reads, leading to high processing power requirements and long data access times when searching for specific nucleotide sequences, as they do not support indexing for these types of data and require decoding and parsing of entire datasets for pattern matching.
Innovation Solution
The method involves clustering reads based on shared nucleotide sequences, encoding cluster signatures as integer numbers, and storing these in ordered or unordered vectors, allowing for pattern matching on clusters rather than individual reads, enabling parallel processing and ordering of reads by cluster signature characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If genome sequencing data is stored as unordered sequence of records without indexing, then data storage is simple, but search for specific patterns requires decoding and parsing the entire dataset resulting in high processing power requirements and long data access times
Solution Approach 1:
The patent segments the genome sequencing data by dividing it into multiple blocks, where each block contains a subset of reads. It further segments each block into sub-blocks organized by cluster signatures. This hierarchical segmentation enables selective access to specific blocks containing relevant patterns without processing the entire dataset, thereby reducing data access time and processing power requirements.
2Productivity
If pattern matching is performed on individual reads, then precision is high, but processing time increases significantly when searching through large datasets
Solution Approach 1:
The patent merges multiple reads that share common subsequences into clusters represented by cluster signatures. Instead of performing pattern matching on each individual read, the system performs matching on the merged cluster signatures. This combining approach maintains matching precision while significantly improving productivity by reducing the number of comparisons needed when searching through large datasets.
Solution Approach 2:
The patent performs preliminary clustering of reads by common subsequences and pre-computes cluster signatures before pattern matching operations. This preliminary organization allows the system to quickly identify and access only those clusters that may contain the target patterns, eliminating the need to process unrelated reads during search operations and thereby reducing search time.
3Measurement precision
If all reads are processed and decoded for gene discovery, then comprehensive analysis is achieved, but processing power requirements become prohibitively high
Solution Approach 1:
The patent extracts and utilizes only the essential common subsequences (cluster signatures) from the reads that are relevant for gene discovery. Instead of processing and decoding all reads in their entirety, the system extracts the discriminative features (signatures) that capture the essential information needed for pattern matching and gene identification. This extraction approach maintains gene discovery accuracy while dramatically reducing processing power requirements.
Data Source
AI summary
Method and apparatus for the indexing of genome sequence data produced by genome sequencing machines. The proposed method can be applied both to raw sequence data produced by sequencing machines and to those sequence reads that cannot be mapped on any reference sequence according to specific matching criteria. This invention describes a method to partition and index unaligned sequence reads to enable browsing and efficient selective access.