Genomic Reference Graphs for Precise NAHR Event Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing genomics and bioinformatics techniques struggle to accurately detect non-allelic homologous recombination (NAHR) events, particularly in whole genome sequencing, due to flawed alignment and subsequent variant calling, which often miss or misinterpret NAHR-related deletions and other genomic rearrangements.
Innovation Solution
A system and method utilizing a genomic reference graph, such as a directed acyclic graph (DAG), to represent potential NAHR events by linking homologous segments, allowing rapid alignment of sequence reads and identifying recombination events through adjacency lists, thereby improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional flat-file reference genomes are used for alignment, then the system is simple to implement, but the detection precision of NAHR events is insufficient
Solution Approach 1:
The patent transitions from a traditional linear flat-file reference genome to a multi-dimensional graph-based genomic reference structure. This graph structure incorporates multiple paths representing different genomic arrangements, including alternative loci and potential NAHR events, allowing sequence reads to be aligned against multiple possible configurations simultaneously, thereby significantly improving NAHR detection precision.
Solution Approach 2:
The genomic graph structure implements a nested organization where nodes represent genomic segments and edges represent relationships between them. The structure nests multiple levels of genomic information including primary reference sequences, alternative loci, and predicted NAHR events within a hierarchical framework, enabling efficient traversal and alignment while maintaining organizational simplicity.
2Reliability
If whole genome sequencing is performed to detect NAHR events, then comprehensive genomic coverage is achieved, but the alignment accuracy deteriorates due to flawed variant calling
Solution Approach 1:
The system performs preliminary actions by pre-processing the genomic reference to identify and mark potential NAHR events, creating a structured graph with predicted recombination sites before alignment occurs. This pre-marking allows the alignment algorithm to efficiently focus on regions of interest without requiring exhaustive analysis of the entire genome, thereby improving both accuracy and computational efficiency.
Solution Approach 2:
The patent introduces an intermediary graphical structure that mediates between the raw sequence reads and the final variant calling process. This genomic graph acts as an intermediate representation layer that captures complex genomic relationships, allowing alignment algorithms to operate more efficiently and accurately by working with the structured graph rather than directly with flat reference files.
3Speed
If the genomic reference is stored as traditional flat files, then storage and access are simple, but the system cannot rapidly identify alternative paths representing NAHR events
Solution Approach 1:
The genomic reference is segmented into discrete nodes and edges within the graph structure, where each node represents a specific genomic segment and each edge represents a relationship or potential recombination event. This segmentation allows the system to efficiently traverse only relevant portions of the genome when searching for NAHR events, dramatically improving identification speed compared to scanning entire flat-file references.
Solution Approach 2:
The graph-based genomic reference implements a dynamic structure that can efficiently adapt during alignment operations. The system can dynamically traverse different paths through the graph based on the sequence reads being analyzed, rapidly identifying alternative paths that represent potential NAHR events without requiring predetermined knowledge of all possible variants.
Data Source
AI summary
A method for screening for disease in a genomic sample is includes receiving a representation of a reference genome comprising a sequence of symbols. The presence of a predicted mutational event is identified in a location of the reference genome. An alternate path is created in the reference genome representing the predicted mutational event. A plurality of sequence reads are obtained from a genomic sample, wherein at least one sequence read comprises at least a portion of the predicted mutational event. The at least one sequence read is then mapped to the reference genome and a location is determined corresponding to the predicted mutational event. The predicted mutational event is then identified as present in the genomic sample. The method may be used to detect evidence of non-allelic homologous recombination (NAHR) occurring in genomic samples.