# US 2023/0053844 A1 · Improved Quality Value Compression Framework in Aligned Sequencing Data Based on Novel Contexts

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0053844 A1
- Document ID: US20230053844A1
- Status: Published application
- Kind: A1
- Application number: 17/797,146
- Filed: 2021-01-27
- Published: 2023-02-23
- Earliest claimed: 2020-02-07
- Assignees: Koninklijke Philips N.V.
- Applicants: Koninklijke Philips N.V.
- Inventors: Shubham Chandak; Yee Him Cheung
- Web page: https://patentgrep.com/patent/US20230053844A1

## CPC

- G16B 50/50: Compression of genetic data
- G06F 16/215: Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
- G16B 30/10: Sequence alignment; Homology search
- G16B 40/20: Supervised data analysis
- G16B: BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method for compressing information includes accessing a read of genomic sequencing data, aligning the read to a reference, generating alignment data based on alignment of the read, obtaining a set of contexts based on the alignment data, and compressing quality values corresponding to the alignment data based on the set of contexts. The alignment data may provide an indication of errors in the genomic sequencing data, and each of the quality values may provide an indication of a probability of error at one or more bases in the genomic sequencing data.

## Claims

### Claim 1

A method for compressing genomic information, comprising:
(a) accessing a read of genomic sequencing data;
(b) aligning the read to a reference;
(c) generating alignment data, which comprises an alignment position on the reference and a characterization of matches and mismatches of bases in the read compared to bases in the reference at alignment position, such as a CIGAR representation, after alignment of the read;
(d) arithmetically compressing quality values of the read, wherein each of the quality values provides an indication of a Probability of error of a base in the genomic sequencing data, wherein the compressing comprises:
(e) selecting a set of one or more arithmetic compression contexts, characterized in that at least one of the arithmetic compression contexts is based on the alignment data, wherein such at least one of the arithmetic compression contexts based on the alignment data comprises selecting from:
whether the base in the read at the position of the quality value matches the base in the reference;
the type of mismatch at the position of the quality value, being an insertion, deletion or substitution,
the type of bases in the reference surrounding the position on the reference corresponding to the position of the quality value in the read, and
a mean quality value of bases surrounding the position on the the reference corresponding to the position of the quality value in the read.

### Claim 2 (depends on 1)

The method of claim 1, wherein the aligned genomic sequencing data is compressed in (e) based on count-based adaptive arithmetic coding.

### Claim 3 (depends on 1)

The method of claim 1, wherein the aligned genomic sequencing data is compressed in (e) based on neural-network prediction-based arithmetic coding.

### Claim 4 (depends on 1)

The method of claim 1, wherein the aligned genomic sequencing data is compressed in (e) based on arithmetic coding, with the arithmetic coding mode and training procedure chosen based on one or more criteria, wherein the one or more criteria include data size, prediction power, processing efficiency, availability of training data, or compatibility with other systems or uses.

### Claim 5 (depends on 1)

The method of claim 1, wherein the set of contexts includes a match between the read and a reference base.

### Claim 6 (depends on 1)

The method of claim 1, wherein the set of contexts includes at least one of:
existence of mismatch, and type of mismatch.

### Claim 7 (depends on 1)

The method of claim 1, wherein the set of contexts includes a number of bases in a reference sequence surrounding the position in the reference sequence corresponding to a quality value to be compressed.

### Claim 8 (depends on 1)

The method of claim 1, wherein the set of contexts includes mean quality value across a plurality of bases at current or nearby genomic coordinates.

### Claim 9 (depends on 1)

The method of claim 1, wherein the set of contexts includes errors at current and nearby bases measured using a pileup of reads mapping to the same genomic coordinate.

### Claim 10 (depends on 1)

The method of claim 1, wherein step (e) includes:
selecting the set of contexts based on one or more criteria,
wherein the one or more criteria include dataset type, dataset size, context size, predictive power of context, or amount of data to be compressed.

### Claim 11

A system for compressing information, comprising:
a memory configured to store instructions; and
a processor configured to execute the instructions to:
(a) access a read of genomic sequencing data;
(b) align the read to a reference;
(c) generate alignment data based on alignment of the read, which comprises an alignment position on the reference and a characterization of matches and mismatches of bases in the read compared to bases in the reference at alignment position, such as a CIGAR representation;
(d) obtain a sequence of quality values for the read, wherein each of the quality values provides an indication of a probability of error of a base in the genomic sequencing data,
(e) select a set of one or more arithmetic encoding contexts, characterized in that these one or more arithmetic encoding contexts are based on the alignment data, wherein such at least one of the arithmetic compression contexts based on the alignment data comprises is selected from:
whether the base in the read at the position of the quality value matches the base in the reference;
the type of mismatch at the position of the quality value, being an insertion, deletion or substitution,
the type of bases in the reference surrounding the position on the reference corresponding to the position of the quality value in the read, and
a mean quality value of bases surrounding the position on the the reference corresponding to the position of the quality value in the read; and;
(f) arithmetically compress the quality values using the set of contexts.

### Claim 12 (depends on 11)

The system of claim 11, wherein the processor is to compress the aligned genomic sequencing data based on count-based adaptive arithmetic coding.

### Claim 13 (depends on 11)

The system of claim 11, wherein the processor is to compress the aligned genomic sequencing data based on neural-network, prediction-based arithmetic coding.

### Claim 14 (depends on 11)

The system of claim 11, wherein the processor is to compress the aligned genomic sequencing data based on arithmetic coding, with the arithmetic coding mode and training procedure chosen based on one or more criteria, wherein the one or more criteria include data size, prediction power, processing efficiency, availability of training data, or compatibility with other systems or uses.

### Claim 15 (depends on 11)

The system of claim 11, wherein the set of contexts includes a match between the read and a reference base.

### Claim 16 (depends on 11)

The system of claim 11, wherein the set of contexts includes at least one of:
existence of a mismatch, and
type of mismatch.

### Claim 17 (depends on 11)

The system of claim 11, wherein the set of contexts includes a number of bases in a reference sequence surrounding a position in the reference corresponding to a quality value to be compressed.

### Claim 18 (depends on 11)

The system of claim 11, wherein the set of contexts includes mean quality value across a plurality of bases at current or nearby genomic coordinates.

### Claim 19 (depends on 11)

The system of claim 11, wherein the set of contexts includes errors at current and nearby bases measured using a pileup of reads mapping to the same genomic coordinate.

### Claim 20 (depends on 11)

The system of claim 11, wherein (de) includes:
selecting the set of contexts based on one or more criteria,
wherein the one or more criteria include dataset type, dataset size, context size, predictive power of context, or amount of data to be compressed.

### Claim 21 (depends on 1)

A method of decompressing genomic information which was compressed according to the method of claim 1.

### Claim 22 (depends on 11)

A system of decompressing genomic information which was compressed using a system as claimed in claim 11.
