PBWT Indexing for Scalable IBD Segment Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying long Identity by Descent (IBD) segments between a query and a panel of haplotypes are inefficient and not scalable for large databases, as they rely on pairwise comparisons with computational complexity O(NM2), where N is the genome length and M is the number of haplotypes.
Innovation Solution
The system employs a dynamic pool of Positional Burrows-Wheeler Transform (PBWT) genetic indexes, a haplotype ingestion engine, and a haplotype query engine to efficiently identify all long IBD segments between a query and a panel of haplotypes, independent of the number of haplotypes, by using a long match query algorithm and maintaining LEAP arrays for efficient searching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pairwise comparison methods are used to identify IBD segments, then all pairs of individuals can be compared, but the computational complexity increases to O(NM2)
Solution Approach 1:
The patent introduces a PBWT index as an intermediary data structure that mediates between the query haplotype and the panel haplotypes. Instead of directly comparing all pairs, the system first builds a PBWT index from the panel haplotypes, then uses this index to efficiently find matches. This intermediary structure reduces the computational complexity from O(NM2) to O(N+c) while maintaining the ability to identify all long IBD segments
Solution Approach 2:
The system performs preliminary action by pre-processing the panel haplotypes to build the PBWT index before actual querying. The index construction organizes the haplotype data in a way that enables efficient searching, allowing subsequent queries to be answered quickly without performing exhaustive pairwise comparisons each time
2Productivity
If traditional indexing methods are used, then database storage is simplified, but genealogical search speed decreases
Solution Approach 1:
The patent implements a dynamic PBWT indexing system that can efficiently adapt to queries of varying lengths. The index structure maintains flexibility to handle different search parameters while providing fast access. The dynamic nature of the index allows it to optimize search operations based on the specific query requirements, achieving O(N+c) search complexity regardless of database size
3Measurement precision
If exhaustive searching is performed to ensure accuracy, then all long matches are identified, but the time required increases significantly
Solution Approach 1:
The patent replaces the mechanical exhaustive search process with an algorithmic approach based on PBWT. Instead of systematically comparing every pair of haplotypes, the system uses the properties of the PBWT index to directly identify matching segments. This substitution of the search mechanism maintains accuracy in identifying all long matches while dramatically reducing the time required from O(NM2) to O(N+c)
Data Source
AI summary
A system for indexing, updating, and search haplotypes for genetic genealogical discovery in genotype databases. The system includes a pool of genetic indexes, a haplotype ingestion engine, and a haplotype query engine. The haplotypes of a number of individuals in a genotype database are indexed by a pool of multiple panels, and each panel pool can be dynamically updated by the insertion or deletion of individual haplotypes. A genetic genealogical search of a query haplotype against the database is achieved by first projecting the query onto a subset of panels in the pool, then conducting long match queries over each panel, and finally aggregating the identified long matches into Identical-by-Descent segments, i.e., DNA matches, between the query and the haplotypes in the database.