Genomic Access Units for Compressed Region-Specific Data Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for compressing and accessing genome sequencing data, such as FASTQ, SAM, BAM, and CRAM formats, suffer from poor compression ratios, inefficient data structures, and difficulties in navigating and manipulating large datasets, leading to slow downstream analysis and inefficient selective access to genomic regions.
Innovation Solution
The proposed method organizes genome sequencing data into entropy-coded Access Units, which are classified and partitioned based on their alignment to a reference sequence, allowing for efficient selective access by embedding indexing information directly within the compressed bit streams and using a single alphabet for substitutions, insertions, and deletions, reducing redundancy and improving compression performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If general purpose compression algorithms (gzip, LZ schemes) are applied to genome sequencing data, then data compression is achieved, but compression performance is poor and data access efficiency deteriorates
Solution Approach 1:
The patent transforms the data representation parameters by converting genome sequencing data from text-based formats (FASTQ, SAM) to binary formats with optimized data structures. This parameter change enables both high compression ratios and efficient random access by fundamentally altering how genomic data is encoded and organized in storage.
Solution Approach 2:
The patent segments the compressed genomic data into independently decodable blocks with embedded indexing information. Each block contains self-contained metadata that enables selective access to specific genomic regions without decompressing the entire file, thus maintaining both high compression and fast access performance.
2Ease of operation
If BAM format with external index is used, then selective access to genomic regions is enabled, but device complexity increases and processing efficiency decreases
Solution Approach 1:
The patent merges the indexing information directly into the compressed data blocks themselves, eliminating the need for separate external index files. This consolidation reduces system complexity by removing redundant index management while maintaining efficient selective access capability through embedded block-level indexes.
Solution Approach 2:
Each compressed data block in the patent contains self-contained indexing information that enables it to be independently located and accessed. This self-service approach allows the data structure to manage its own access operations without requiring complex external indexing systems.
3Quantity of substance
If monolithic compressed format is used, then storage space is reduced, but data manipulation and processing become difficult
Solution Approach 1:
The patent divides the compressed genomic dataset into multiple independently decodable blocks, each representing a specific genomic region. This segmentation allows random access to individual blocks or subsets of blocks, enabling efficient data manipulation and processing of specific genomic regions without requiring decomposition of the entire compressed file.
Solution Approach 2:
The patent creates a dynamic data structure where compressed blocks can be selectively accessed, combined, or processed based on query requirements. The system adapts to different access patterns by allowing flexible combination of blocks, providing both storage efficiency and manipulation ease through dynamic block-level operations.
Data Source
AI summary
Method and apparatus for the coding and selective access of compressed genomic sequence data produced by genomic sequencing machines. The coding process is based on aligning sequence reads with respect to pre-existing or constructed reference sequences, on classifying and coding the sequence reads by means of sets of descriptors, and further partitioning the descriptor sets into access units of different types. Efficient selective access to specific genomic regions with the guarantee of retrieving all sequence reads mapped to those regions, is provided by: signaling the type of data mapping configuration used to store or transmit the descriptor sets, determining the minimum number of access units that need to be retrieved and decoded to access a genomic region, providing a master index table that contain all information for optimizing the data access process.