Nucleic Acid Sequencing Error Correction via Consensus Tagging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current nucleic acid sequencing technologies suffer from high error rates, leading to false-positive mutation calls, particularly for mutations with minor allele frequencies, which limits the reliable detection of single-nucleotide polymorphisms (SNPs) below 1%.
Innovation Solution
A method that aligns genetic reads to a reference genome, groups them based on shared positions, creates consensus sequences by setting the most abundant nucleotide at each position if it meets a threshold ratio, and uses a tag 'N' for positions below this threshold, then identifies true variations by comparing consensus sequences and calculating a ratio between 'N' and variant sequences to determine authenticity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If standard next-generation sequencing is used to detect genetic variations, then high throughput sequencing is achieved, but false-positive mutation calls occur due to high error rates
Solution Approach 1:
The patent segments the sequencing data processing into multiple stages: initial read alignment, grouping reads into families based on shared positions, creating consensus sequences for each family, and then identifying variations. This segmentation allows error correction at the consensus sequence level while maintaining high throughput processing of individual reads.
Solution Approach 2:
The patent changes the parameter of error tolerance by requiring a minimum threshold (at least 76%) of reads within a family to agree on a nucleotide position before calling it as the consensus. This parameter change filters out random sequencing errors while preserving true biological variations.
2Speed
If standard sequencing error rates of 0.5-1% are accepted, then sequencing speed is maintained, but mutations with minor allele frequencies below 1% cannot be reliably detected
Solution Approach 1:
The patent creates multiple copies (consensus sequences) of the true genetic signal by grouping similar reads into families. Each consensus sequence represents a amplified copy of the true biological variation, making it distinguishable from random sequencing errors even at low allele frequencies.
Solution Approach 2:
The patent uses disposable read-level error correction by creating consensus sequences from groups of reads. Each consensus sequence is a temporary computational construct used to filter errors, then discarded after variation calling, allowing continuous processing without carrying forward error corrections.
3Reliability
If strict consensus thresholds are applied to reduce false positives, then reliability improves, but sensitivity to detect true low-frequency variants decreases
Solution Approach 1:
The patent adds a new dimension of analysis by evaluating not just the frequency of variant alleles but also the frequency of 'N' calls (uncertain positions) within each consensus sequence. True variants show low N-frequency while false positives show high N-frequency, providing an orthogonal criterion for discrimination.
Solution Approach 2:
The patent implements feedback by using the ratio of N-calls to variant-calls as a quality metric. Consensus sequences with high N-ratios are flagged as potentially erroneous and excluded from variation calls, while those with low N-ratios are accepted as true variants, creating a self-correcting system.
Data Source
Figure 1(A)~1(B)
Figure 2(A)~2(C)
Figure 3(A)~3(B)
AI summary
The present invention relates to methods and systems for nucleic acid sequencing. In particular, the present invention relates to methods and systems for reducing the number of false-positives in nucleic acid sequencing. The method comprises: aligning a plurality of genetic reads to a reference genetic sequence; grouping the genetic reads into a plurality of groups; creating a consensus sequence for each group of the plurality of groups by setting a representation of the most abundant nucleotide man_p or a tag N based on a ratio r; and identifying a variation as a true variation if a ratio r* between the number of consensus sequences comprising the tag N at a specific position p and the number of the consensus sequences comprising the variation at the specific position p is below a threshold t*.