Genomic Quality Score Adjustment for Higher Compression Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for compressing genomic sequencing data are limited by the large volume and high entropy of quality score values, leading to inefficient data handling and storage, with lossy compression schemes risking valuable information loss.
Innovation Solution
A system and method that utilize Coding Theory principles, specifically Bayesian estimation and Hamming/Edit distances, to adjust and boost quality scores, improving their accuracy and enabling better compression without modifying the underlying data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If lossy compression schemes are used to compress genomic sequencing data, then data size is reduced, but valuable information is lost
Solution Approach 1:
The patent transforms quality score values from their original Phred scale representation into a new parameter space using logistic transformation and quantization. This parameter transformation enables more efficient compression by changing the distribution characteristics of the data, allowing lossless compression to achieve better ratios without information loss.
Solution Approach 2:
The patent creates compressed representations of quality score data through systematic encoding schemes that preserve all original information. By using reversible transformation and encoding methods, the system produces a compact copy of the data that can be perfectly reconstructed, avoiding both information loss and the need for lossy compression.
2Quantity of substance
If conventional compression techniques are used on quality score values, then some compression is achieved, but compression ratios are limited due to high entropy
Solution Approach 1:
The patent applies parameter transformation to quality score data, converting Phred scores through logistic functions and quantization schemes. This changes the entropy distribution of the data, creating a more favorable probability distribution for compression algorithms and achieving significantly higher compression ratios.
Solution Approach 2:
The patent segments quality score data into discrete quantization bins and processes different segments with optimized encoding schemes. By dividing the quality score range into distinct levels and applying targeted compression to each segment, the system overcomes the limitations of uniform compression approaches.
3Quantity of substance
If quality score values are stored in compressed form, then storage efficiency improves, but the alphabet size and entropy contribute to larger data quantities
Solution Approach 1:
The patent transforms quality score parameters from a large-alphabet representation (256 possible values) into a smaller quantized representation (e.g., 6-10 levels). This parameter reduction directly decreases the alphabet size effect on data quantity while maintaining storage efficiency through the transformed parameter space.
Data Source
Figure 1
Figure 2
Figure 3a~3b
AI summary
There is provided a system (900) for generating output genomics data (970) from biological genetic material (930). The system (900) includes: (i) a genetic reading apparatus (910) which is operable to read bases included in the biological genetic material (930) to generate a genome sequence read, wherein the genome sequence read comprises at least one sequence of bases and associated quality scores; and (ii) a data processing apparatus (950) for processing the genome sequence read to generate the output genomics data (970), characterized in that the data processing apparatus (950) is operable: (a) to determine similarity criteria (300) for performing searching in a reference genome corpus (310); (b) to perform a search of the at least one sequence of bases in the reference genome corpus (310), based upon the similarity criteria (300); (c) to calculate (330) an adjustment for one or more of the associated quality scores, based upon results of the search (320); and (d) to adjust one or more of the associated quality scores according to the calculated adjustment from (c).