Tandem Duplication Detection via Soft-Clipped Read Analysis
Solution Overview
Problem
Current methods for detecting tandem duplications, particularly in the FLT3 gene, using next-generation sequencing technologies face challenges in accurately identifying long inserts and soft-clipped regions, leading to incomplete detection of tandem duplications.
Innovation Solution
A method involving the analysis of partially mapped reads with soft-clipped portions, where the algorithm checks for anchors and determines insert sizes by comparing soft-clipped and mapped portions, categorizes reads based on insert sizes, and merges features to accurately determine tandem duplication locations and sizes, utilizing a processor and machine-readable media to execute these steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard next-generation sequencing pipelines are used for tandem duplication detection, then processing speed and throughput are maintained, but detection accuracy for long inserts and soft-clipped regions deteriorates
Solution Approach 1:
The patent segments the read analysis into distinct components: mapped portions that align to the reference sequence and soft-clipped portions that represent unmapped regions. This segmentation enables specialized processing of each component, allowing accurate detection of tandem duplications while maintaining overall pipeline efficiency.
Solution Approach 2:
The patent introduces an intermediary analysis step that processes soft-clipped portions separately from standard alignment results. This intermediary layer captures duplication information that would otherwise be lost, bridging the gap between standard sequencing output and accurate duplication detection.
2Measurement precision
If standard alignment algorithms are used, then mapping speed is maintained, but ability to detect soft-clipped regions and long inserts deteriorates
Solution Approach 1:
The patent performs preliminary classification of reads into mapped and soft-clipped portions during the alignment process. This preliminary action identifies which reads require special duplication analysis, allowing efficient processing by focusing computational resources only on relevant reads rather than reprocessing all reads through complex analysis.
Solution Approach 2:
The patent applies partial action by analyzing only the soft-clipped portions of reads for duplication detection, rather than reanalyzing the entire read sequence. This selective approach maintains processing efficiency while achieving accurate insert size determination for the critical regions.
3Measurement precision
If reads are classified by insert size categories, then duplication detection accuracy improves, but computational steps increase
Solution Approach 1:
The patent changes the parameter of read classification from binary (mapped/unmapped) to multi-category based on insert size. By establishing size categories (e.g., small, medium, large inserts), the system enables more precise duplication detection while managing computational complexity through structured classification rather than exhaustive analysis.
Data Source
AI summary
A method for detecting a tandem duplication in an FLT3 gene of a sample, includes mapping reads corresponding to targeted regions of exons of the FLT3 gene to a reference sequence. A partially mapped read includes a mapped portion, a soft-clipped portion and a breakpoint. Analyzing the partially mapped reads intersecting a column of the pileup includes detecting a duplication in the soft-clipped portion by comparing the soft-clipped portion to the mapped portion adjacent to the breakpoint; determining an insert size of the duplication in the soft-clipped portion; and assigning the partially mapped read to a category based on the insert size. Categories correspond to insert sizes. The categories are filtered and converted into features corresponding to the column. The features corresponding to one or more columns representing a same insert are merged to determine a location and size of a tandem duplication.