Reference Genome Cell Arrays for Scalable Sample Read Location
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional genome sequencing methods, both de novo and reference-aligned, face challenges with high computational cost, memory resource requirements, and limited scalability due to the need to compare large volumes of sample reads against reference genomes, which are constrained by the number of compute threads and memory channels.
Innovation Solution
A reference-guided device using arrays and circuitry to probabilistically locate sample reads within a reference genome by storing overlapping reference sequences, allowing for efficient identification of matching groups of cells, thereby reducing memory and computational demands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional de novo sequencing stores large group of sample reads in shared memory, then sample reads can be compared to determine locations, but the number of compute threads is limited due to shared memory constraints
Solution Approach 1:
The patent divides the reference genome into multiple smaller reference panels, each stored in separate memory. This segmentation allows multiple compute threads to access different reference panels simultaneously without contending for the same shared memory resources, thereby increasing the number of parallel compute threads while reducing the memory burden on each thread.
Solution Approach 2:
The patent transitions from a single shared memory architecture to a distributed memory architecture where reference panels are distributed across multiple memory locations. This dimensional change in memory organization enables parallel access by multiple compute threads, effectively increasing productivity without proportionally increasing total memory requirements.
2Measurement precision
If conventional reference-aligned sequencing stores full reference genome in dedicated memory, then sample reads can be located, but memory resources and computational cost are significant
Solution Approach 1:
The reference genome is segmented into multiple reference panels that are distributed across different memory locations. Each compute thread accesses only the relevant reference panel needed for its specific sample read location task, rather than loading the entire reference genome into memory. This reduces memory resources while maintaining location accuracy through targeted access to appropriate reference segments.
Solution Approach 2:
Different compute threads are assigned to access different local reference panels based on their specific needs for locating particular sample reads. This local quality approach ensures that each compute thread has access to the precise reference data it needs without wasting memory resources on unrelated reference genome portions, optimizing both memory usage and location precision.
3Productivity
If conventional methods partition sample reads into groups for processing, then computational cost is reduced, but scalability is limited due to memory constraints
Solution Approach 1:
The patent segments both the reference genome into distributed reference panels and the sample reads into processing groups, where each group is assigned to compute threads that access specific reference panels. This dual segmentation enables scalable processing by allowing the system to add more compute threads and reference panel memory locations without proportionally increasing the complexity of the memory architecture, as each component remains independently addressable.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A device for locating a sample read with respect to a reference genome includes a plurality of groups of cells. Each group of cells stores a reference sequence representing reference bases from the reference genome corresponding to an order of cells in the respective group of cells. Each group of cells further stores a current substring sequence representing sample bases from the sample read corresponding to the order of the cells in the respective group of cells. Each group of cells stores the same current substring sequence and a reference sequence representing a portion of the reference genome that partially overlaps at least one other portion of the reference genome represented by one or more other reference sequences stored in one or more other groups of cells. Groups of cells are identified among the plurality of groups of cells where the stored reference sequence matches the current substring sequence.