Genomic Reference Reconstruction in Compressed Sequence Reads

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for compressing aligned genomic sequence reads face inefficiencies, particularly in reference-less compression scenarios where the reference genome is not available, leading to poor compression ratios and inability to reconstruct data without the reference sequence.

Innovation Solution

A method that builds contigs by merging aligned reads and uses specific entropy coders to compress genomic descriptors, allowing for lossless compression and reconstruction of genomic sequence reads and reference sequences without requiring the reference genome, by encoding contig length and mismatches between contigs and reference sequences.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If reference-based compression is used to compress aligned genomic sequences, then compression ratio is improved, but the reference sequence must be available at both encoding and decoding sides which reduces adaptability

Engineering Contradiction:
Improvecompression ratioVSAvoidavailability of reference sequence
Core Design Contradiction:
Quantity of substanceVSAdaptability or versatility

Solution Approach 1:

The patent extracts the reference sequence information from the external environment and embeds it within the compressed data structure. Specifically, the reference sequence is encoded using differential coding (storing only differences from a known reference) and included in the compressed file, allowing the decoder to reconstruct the reference sequence locally without requiring external availability. This resolves the contradiction by making the system self-contained while maintaining high compression ratios.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary encoding of the reference sequence during the compression phase. The reference sequence is pre-processed and stored in a compressed differential format within the output file before the actual genomic data compression occurs. This preliminary action ensures that when decomposition is needed, the reference sequence is already prepared and available within the data structure itself, eliminating the need for external reference availability.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If general purpose compressors are used for reference-less compression, then adaptability is improved, but compression ratio deteriorates to poor levels

Engineering Contradiction:
Improveindependence from reference sequenceVSAvoidcompression ratio
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The patent introduces an intermediary approach that combines elements of both reference-based and reference-less methods. Instead of using general-purpose compressors directly on raw genomic data, the system first performs differential coding against a reference sequence (creating a simplified intermediate representation), then applies entropy coding to this intermediate form. This intermediary step captures the redundancy in genomic data while maintaining independence from external reference availability, achieving both good compression ratios and adaptability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the parameter representation of genomic data by encoding positions and differences rather than raw sequences. By transforming the data into a differential format (storing offset positions and difference values rather than complete sequences), the system enables efficient compression without requiring external reference sequences, thus improving compression ratio while maintaining adaptability.

Inventive Principle:
Principle #35Parameter changes

3Reliability

If reference genome is included in compressed representation for storage, then reconstruction capability is improved, but compression efficiency deteriorates

Engineering Contradiction:
Improvereconstruction capabilityVSAvoidcompression efficiency
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent changes how the reference sequence is represented by storing only differential information (positions and differences) rather than the complete reference sequence. This parameter transformation reduces the storage requirement from O(N) where N is the reference sequence length to O(M) where M is the number of differences, which is typically much smaller. This enables reliable reconstruction while maintaining compression efficiency.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent extracts only the essential differential information needed for reconstruction rather than storing the entire reference sequence. By taking out only the necessary difference data (positions and offset values) and embedding this minimal required information in the compressed file, the system achieves reconstruction capability without the storage overhead of including the full reference genome, thus maintaining compression efficiency.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3583249B1Method and systems for the reconstruction of genomic reference sequences from compressed genomic sequence reads
Publication Date: 2023.07.19 GENOMSYS SA
  • EP3583249B1 patent drawingFigure 1
  • EP3583249B1 patent drawingFigure 2
  • EP3583249B1 patent drawingFigure 3

AI summary

The method and apparatus described in this disclosure include representing a reference genome in terms of syntax elements describing the differences between said reference genome and previously aligned genomic sequences. Each of the aligned genomic sequence is described by means of a subset of syntax elements. Syntax elements describing all the genomic sequences are partitioned in blocks according to their statistical properties. Each block of syntax elements is entropy coded. The entropy coded blocks are then concatenated to form a compressed bitstream. The differences between the reference genome and the aligned sequences are expressed in terms of syntax elements, which are embedded in the bitstream of coded blocks of syntax elements describing aligned reads. The disclosed method enables the reconstruction of the reference genome used for alignment when decoding the compressed genomic sequences while preserving different options of random access on the compressed data and enabling efficient compression.