Systems and methods for genotyping with graph reference

• US 10,262,102 B2
• Filed: 02/24/2016
• Issued: 04/16/2019
• Est. Priority Date: 02/24/2016
• Status: Active Grant

• Current Assignee: Seven Bridges Genomics, Inc.
• Original Assignee: Seven Bridges Genomics, Inc.
• Inventors: Brown, Richard
• Primary Examiner(s): Negin, Russell S
• Application Number:
  US15/052,354
• Publication Number:
  US 20170242958A1
• Time in Patent Office: 1,147 Days
• Field of Search: None
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• CPC Class Codes: G16B 20/20 Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
• CPC Class Codes: G16B 45/00 ICT specially adapted for bioinformatics-related data visualisation, e.g. displaying of maps or networks

Abstract

Genomic references are structured as a reference graph that represents diploid genotypes in organisms. A path through a series of connected nodes and edges represents a genetic sequence. Genetic variation within a diploid organism is represented by multiple paths through the reference graph. The graph may be transformed into a traversal graph in which a path represents a diploid genotype. Genetic analysis using the traversal graph allows an organism'"'"'s diploid genotype to be elucidated, e.g., by mapping sequence reads to the reference graph and scoring paths in the traversal graph based on the mapping to determine the path through the traversal graph that best fits the sequence reads.

Claims

1. A method of determining a genotype, the method comprising:

providing a reference graph in computer memory, the reference graph representing a plurality of genomic sequences and comprising nodes connected by edges into a plurality of paths, wherein variation across the plurality of genomic sequences is incorporated as divergent paths converging to nodes representing conserved sequence, such that each of the plurality of genomic sequences is represented as a path in the reference graph;

wherein the plurality of paths further comprises a first alternate path representing a first haplotype A at a position and a second alternate path representing a second haplotype B at the position, wherein each of the first and second alternate paths converge to a node representing a conserved sequence;

identifying one or more genotypes corresponding to either or both of the first haplotype A and the second haplotype B, the one or more genotypes comprising a first genotype AA, a second genotype AB, and a third genotype BB;

mapping a plurality of sequence reads from an organism to the reference graph, wherein the mapping comprises looking backwards to predecessor paths to identify an optimal location for a sequence read across multiple alternate paths;

assigning, based on properties of the mapped plurality of sequence reads to the reference graph, scores to each of the identified genotypes, wherein the assigning considers whether a sequence read mapping to the first haplotype A changes the probability of the genotype including the second haplotype B; and

identifying one of the first genotype AA, second genotype AB, or third genotype BB as having a highest score in said assigning step, thereby providing a genotype.

2. The method of claim 1, wherein the properties of the mapped reads are selected from the group consisting of position, count, and quality.

3. The method of claim 1, further comprising enumerating a plurality of genotypes based on a plurality of haplotypes in the reference graph, assigning scores to each genotype, and identifying a most likely sequence of genotypes for the organism using dynamic programming.

4. The method of claim 3, wherein the dynamic programming comprises using Viterbi decoding.

5. The method of claim 1, wherein the plurality of sequence reads represent at least a majority of a genome of an organism.

6. The method of claim 1, further comprising:

representing the first genotype AA, the second genotype AB, and the third genotype BB using a traversal graph comprising nodes connected by edges into a plurality of paths, the plurality of paths including a first traversal path representing the first genotype AA, a second traversal path representing the second genotype AB, and a third traversal path representing the third genotype BB; and

determining weights for the first traversal path, the second traversal path, and the third traversal path based on properties of the sequence reads that map to the reference graph.

7. The method of claim 6, further comprising enumerating a plurality of genotypes from the reference graph, representing the plurality of genotypes as paths in the traversal graph, and identifying the most likely sequence of genotypes by determining a most probable path through the traversal graph.

8. The method of claim 7, wherein identifying the most likely sequence of genotypes by determining a most probable path through the traversal graph comprises using dynamic programming.

9. The method of claim 8, wherein using dynamic programming comprises using a Dijkstra operation to find a least-cost path.

10. The method of claim 1, wherein A represents an allele of a gene known to be dominant and B represents an allele of the gene known to be recessive.

11. The method of claim 10, wherein the organism is a patient, the method further comprising providing a report for the patient that reports the identified one of the first genotype AA, the second genotype AB, and the third genotype BB.

12. The method of claim 1, wherein providing a reference graph further comprises obtaining the plurality of genomic sequences and transforming the plurality of genomic sequences into a reference graph, wherein matching homologous segments of the known sequences are each represented by a single object in the reference graph.

13. The method of claim 12, wherein matching homologous segments of the plurality of genomic sequences are each represented by a single node.

14. The method of claim 12, wherein assigning scores to each of the identified genotypes comprises determining the likelihood of sequence reads mapping to each edge of a reference graph given each possible state.

15. A system for determining a genotype, the system comprising a processor coupled to a tangible memory subsystem having stored therein:

a reference graph representing a plurality of genomic sequences and comprising nodes connected by edges into a plurality of paths, wherein variation across the plurality of genomic sequences is incorporated as divergent paths converging to nodes representing conserved sequence, such that each of the plurality of genomic sequences is represented as a path in the reference graph;

wherein the plurality of paths further comprises a first alternate path representing a first haplotype A at a position and a second alternate path representing a second haplotype B at the position, wherein each of the first and second alternate paths converge to a node representing a conserved sequence; and

instructions that when executed by the processor cause the system to;

identify one or more genotypes corresponding to either or both of the first haplotype A and the second haplotype B, the one or more genotypes comprising a first genotype AA, a second genotype AB, and a third genotype BB;

map a plurality of sequence reads from an organism to the reference graph, wherein the mapping comprises looking backwards to predecessor paths to identify an optimal location for a sequence read across multiple alternate paths;

assign—

based on properties of the mapped plurality of sequence reads to the reference graph—

scores to each of the identified genotypes, wherein the assigning considers whether a sequence read mapping to the first haplotype A changes the probability of the genotype including the second haplotype B; and

identify one of the first genotype AA, second genotype AB, or third genotype BB as having a highest score in said assigning step, thereby providing a genotype.

16. The system of claim 15, wherein the instructions are further operable to:

represent the first genotype AA, the second genotype AB, and the third genotype BB using a traversal graph comprising nodes connected by edges into a plurality of paths, the plurality of paths including a first traversal path representing the first genotype AA, a second traversal path representing the second genotype AB, and a third traversal path representing the third genotype BB; and

determine weights for the first traversal path, the second traversal path, and the third traversal path based on numbers of the sequence reads that map to the traversal graph.

17. The system of claim 16, wherein the instructions are further operable to:

enumerate a plurality of genotypes from the reference graph;

represent the plurality of genotypes as paths in the traversal graph; and

identify the most likely sequence of genotypes by determining a most probable path through the traversal graph.

18. The system of claim 17, wherein identifying the most likely sequence of genotypes by determining a most probable path through the traversal graph comprises using dynamic programming.

19. The system of claim 18, wherein using dynamic programming comprises using a Dijkstra operation to find a least-cost path.

20. The system of claim 15, wherein the organism is a patient and the system is operable to provide a report for the patient that gives the identified one of the first genotype AA, the second genotype AB, and the third genotype BB.