Fragment Sequence Alignment Using K-mer Hashing and Gap Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Next-generation sequencing technologies face challenges in efficiently mapping large numbers of short DNA/RNA sequence reads against a human-sized genome within a short timeframe, requiring rapid and accurate alignment methods to identify insertions and deletions.

Innovation Solution

A processor-based method for aligning fragment sequences to a reference sequence using both ungapped local and gapped alignment techniques, allowing for the identification of insertions and deletions, with a scoring function to determine the best alignment quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If billions of short reads are mapped against a human-sized genome using traditional alignment methods, then mapping completeness is improved, but mapping time becomes excessively long

Engineering Contradiction:
Improvemapping completenessVSAvoidmapping time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent divides the sequence read into multiple k-mers (subsequences of length k) and processes each k-mer independently through hash table lookups. This segmentation allows parallel processing of billions of reads by breaking them into smaller, manageable units that can be quickly matched against the genome, dramatically reducing mapping time while maintaining completeness through systematic reconstruction of full alignments from k-mer matches.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary indexing of the reference genome by pre-computing and storing all possible k-mer sequences and their genomic positions in hash tables before actual read mapping begins. This preliminary action creates an optimized lookup structure that enables rapid retrieval during mapping, eliminating the need for computationally intensive real-time sequence searching and reducing overall mapping time.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If rapid alignment methods are used to map billions of reads quickly, then mapping speed is improved, but accuracy in identifying insertions and deletions deteriorates

Engineering Contradiction:
Improvemapping speedVSAvoidinsertion and deletion detection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent applies different processing qualities to different regions of the alignment. For well-matched regions, rapid k-mer based matching is sufficient, while for regions showing potential insertions or deletions (indicated by mismatches or gaps in k-mer continuity), the system performs more rigorous local realignment and quality assessment. This localized quality adjustment maintains high speed overall while ensuring accurate detection of structural variations where needed.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system incorporates feedback mechanisms where initial rapid k-mer matching results are evaluated, and regions showing anomalies (such as unexpected mismatches or discontinuities) trigger secondary verification steps. The feedback loop allows the system to automatically allocate additional computational resources to problematic regions, ensuring accurate insertion and deletion detection without requiring exhaustive processing of all reads at high computational cost.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20210210164A1Systems and methods for mapping sequence reads
Publication Date: 2021.07.08 LIFE TECHNOLOGIES CORP
  • US20210210164A1 patent drawing
  • US20210210164A1 patent drawing
  • US20210210164A1 patent drawing

AI summary

Systems, methods, and computer program products for aligning a fragment sequence to a target sequencing. The alignment is allowed at most one gap, such as an insertion or a deletion. In some embodiments, both a gapped alignment and an ungapped alignment can be produced. A selection can be made between the gapped alignment and the ungapped alignment based on a quality value for each alignment.