Microbial Genomic Sequence Detection with Local Depth Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Next-generation sequencing technologies face challenges in accurately identifying specific genomic sequences due to variable sequencing depth and high error rates, leading to false positives and negatives, particularly in the context of SARS-CoV-2 variant identification.
Innovation Solution
A method that computes local sequencing depth by counting k-mers in a neighborhood of target genomic sequences, using a machine learning model with adaptive thresholding based on local sequencing depth estimation, to enhance accuracy and reduce false positives and negatives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If average sequencing depth is used for detection, then the detection process is simple, but false positives and false negatives increase due to highly variable sequencing depth
Solution Approach 1:
The patent applies local quality by computing sequencing depth locally around each target genomic sequence position rather than using a global average. For each position, the method counts reads overlapping a window around that position, creating position-specific depth values that reflect local variations in sequencing coverage. This resolves the contradiction by maintaining operational simplicity while eliminating false positives/negatives caused by using inappropriate global averages in regions with highly variable depth.
2Reliability
If complex bioinformatics pipelines are used to process reads, then detection accuracy may improve, but workflow complexity increases
Solution Approach 1:
The patent extracts and isolates the critical detection function from complex bioinformatics pipelines. Instead of requiring full genome assembly, alignment, and annotation pipelines, the method directly counts reads overlapping target genomic sequence positions and compares to locally computed depth thresholds. This extraction maintains high detection accuracy for specific genomic sequences while dramatically simplifying the workflow to essential operations only.
Solution Approach 2:
The method employs self-service by using the sequencing data itself to compute local depth thresholds without requiring external reference datasets or complex statistical models. The local sequencing depth is calculated directly from the read counts in the dataset being analyzed, allowing the system to self-calibrate detection thresholds based on its own data characteristics, eliminating dependency on complex external pipelines.
3Measurement precision
If enrichment steps are performed before sequencing, then sensitivity for specific genomic regions is improved, but sequencing depth becomes highly variable leading to more false results
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the detection threshold based on local sequencing depth at each position. Rather than using a fixed threshold, the method computes the local depth (average reads per position in a window) and sets the threshold as a fraction of this local value. This adapts the detection criterion to the local enrichment level, maintaining reliability despite variable depth caused by enrichment steps.
Data Source
AI summary
A method for detecting target genomic sequences at predetermined positions in a sequenced genome of a microbial organism in the form of reads. The method comprises computing a local sequencing depth of the genome in a neighborhood of the position of a target genomic sequence, the computing including a) detecting in the reads a set of digital genomic sequences from at least one reference genome of the microbial organism, b) generating of third set of digital genomic sequences comprising reads and digital genomic sequences of the second set belonging to the neighborhood c) counting the number of copies of each digital genomic sequences of the third set of digital genomic sequences, and d) computing the local sequencing depth as being equal to the maximum of the counted numbers of copies.