K-Mer Pseudoalignment for Fast Reference-Based Variant Calling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current variant calling methods are inefficient in identifying genetic variants with high speed, accuracy, and technology versatility, particularly in single-cell sequencing data, and do not effectively utilize existing databases like COSMIC to improve sensitivity and interpretability.
Innovation Solution
A reference-based variant calling tool, varseek, uses k-mer pseudoalignment and kallisto to create a variant-containing reference sequence index, enabling fast and accurate identification of variants in DNA, bulk RNA, and single-cell RNA sequencing data by matching sequencing reads to a reference database of variants.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional variant calling methods are used, then variant identification can be performed, but the process is slow and computationally inefficient
Solution Approach 1:
The patent pre-generates variant-containing reference sequences (VCRSs) from a reference genome and known variant database before actual variant calling. This preliminary action creates a ready-to-use index that eliminates the need for real-time complex computations during sequencing data analysis, thereby significantly improving both speed and computational efficiency.
Solution Approach 2:
The patent creates copies of reference sequences with known variants embedded (VCRSs) and builds an index from these copies. During variant calling, sequencing reads are matched against this pre-built index rather than performing de novo alignment, which dramatically accelerates the variant identification process while maintaining accuracy.
2Measurement precision
If comprehensive variant databases are utilized, then sensitivity improves, but the complexity of the analysis increases
Solution Approach 1:
The patent segments the comprehensive variant database into pre-processed VCRS entries that are incorporated into the reference index. This segmentation allows the system to leverage comprehensive variant information without requiring complex real-time database queries, as the variant data is already organized and integrated in the pre-built index.
Solution Approach 2:
The VCRS index acts as an intermediary between the comprehensive variant database and the sequencing reads. Instead of directly querying the complex variant database during analysis, the system uses the simplified VCRS index as a mediator that has already processed and organized the variant information, reducing analysis complexity while maintaining sensitivity.
3Measurement precision
If traditional alignment methods are used, then accurate variant calling can be achieved, but runtime is excessive
Solution Approach 1:
The patent creates copied and modified reference sequences (VCRSs) that contain known variants embedded within them. These copies are indexed and used for rapid matching against sequencing reads, replacing traditional time-consuming alignment methods while preserving accuracy through the embedded variant information.
Solution Approach 2:
The patent performs preliminary alignment and variant embedding by pre-generating VCRSs from the reference genome and variant database. This preliminary action moves the computationally intensive alignment work to before the actual variant calling process, allowing rapid accurate variant detection during runtime without sacrificing precision.
Data Source
AI summary
Disclosed herein include systems, devices, methods, and computer readable media for reference-based mutation or variant detection. The detection can be based on an index of variant-containing reference sequences and k-mer-based psuedoalignment.