Genome Sequence Identification via Coverage Threshold Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current state-of-the-art techniques are inefficient in distributing relative abundance values across related strains within the same species in metagenomic sequencing, particularly at higher resolutions of species and strain levels due to high inter-strain similarity.
Innovation Solution
A method for identifying candidate genome sequences by estimating coverage, which involves indexing reference genome sequences, mapping genomic read sequences, computing relative abundances, and refining estimates using coverage and cardinality thresholds to select candidate genome sequences through iterative refinement techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Maximum Likelihood Estimation (MLE) approach is used for species level abundance estimation, then species level abundance estimation accuracy is improved, but strain level abundance estimation remains challenging due to high inter-strain similarity
Solution Approach 1:
The patent segments the abundance estimation problem by separately addressing species-level and strain-level estimation. It uses MLE for species-level estimation while introducing a coverage-based refinement step specifically for strain-level estimation, dividing the complex problem into manageable parts with different methodologies optimized for each resolution level
Solution Approach 2:
The patent performs preliminary coverage estimation for each strain before final abundance calculation. By computing coverage values and sorting strains based on coverage thresholds in advance, the system prepares candidate strain sets that simplify the subsequent iterative refinement process and improve strain-level estimation accuracy
2Measurement precision
If coverage estimation and iterative refinement are performed for all reference genome sequences, then strain level abundance estimation accuracy is improved, but computational time and processing complexity increase
Solution Approach 1:
The patent applies partial action by first computing coverage for all strains, then using coverage thresholds to select only the top candidate strains for iterative refinement. This partial application of the refinement process to only promising candidates significantly reduces computational time while maintaining high estimation accuracy for the most abundant strains
Solution Approach 2:
The patent implements feedback through an iterative refinement process where coverage estimates are used to select candidate strains, which then undergo iterative abundance estimation. The process monitors convergence and can terminate when improvements become negligible, using feedback from each iteration to guide subsequent processing and optimize computational efficiency
Data Source
AI summary
This disclosure relates generally to identifying candidate genome sequences. Next generation sequencing (NGS) is a massively parallel sequencing technique for identifying candidate genome sequences. The current state-of-the-art techniques for identifying candidate genome sequences does not efficiently address the problem of distributing abundance values across several related strains that are present in the reference under the same species. The disclosed technique proposes a technique for identifying candidate genome sequences by estimating coverage. The disclosed technique includes a local search-based optimization to compute maximum likelihood-based estimates using constrains on coverage/cardinality thresholds for identifying candidate genome sequences.