Methods for estimating genome-wide copy number variations

• US 8,725,422 B2
• Filed: 10/11/2011
• Issued: 05/13/2014
• Est. Priority Date: 10/13/2010
• Status: Active Grant

• Current Assignee: Complete Genomics Incorporated (BGI Genomics Co., Ltd.)
• Original Assignee: Complete Genomics Incorporated (BGI Genomics Co., Ltd.)
• Inventors: Halpern, Aaron, Pant, Krishna
• Primary Examiner(s): Skibinsky, Anna
• Application Number:
  US13/270,989
• Publication Number:
  US 20120095697A1
• Time in Patent Office: 945 Days
• US Class Current: 702/19
• CPC Class Codes: C12Q 1/6809 Methods for determination or identification of nucleic acids involving differential detection
• CPC Class Codes: C12Q 1/6827 for detection of mutation or polymorphism
• CPC Class Codes: C12Q 2537/16 Assays for determining copy number or wherein the copy number is of special importance
• CPC Class Codes: C12Q 2537/165 Mathematical modelling, e.g. logarithm, ratio
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• CPC Class Codes: G16B 20/10 Ploidy or copy number detection
• CPC Class Codes: G16B 20/20 Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
• CPC Class Codes: G16B 25/00 ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression
• CPC Class Codes: G16B 25/20 Polymerase chain reaction [PCR]; Primer or probe design; Probe optimisation
• CPC Class Codes: G16B 30/00 ICT specially adapted for sequence analysis involving nucleotides or amino acids
• CPC Class Codes: G16B 30/10 Sequence alignment; Homology search
• CPC Class Codes: G16B 30/20 Sequence assembly
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
• CPC Class Codes: G16B 40/30 Unsupervised data analysis

Abstract

Methods for determining the copy number of a genomic region at a detection position of a target sequence in a sample are disclosed. Genomic regions of a target sequence in a sample are sequenced and measurement data for sequence coverage is obtained. Sequence coverage bias is corrected and may be normalized against a baseline sample. Hidden Markov Model (HMM) segmentation, scoring, and output are performed, and in some embodiments population-based no-calling and identification of low-confidence regions may also be performed. A total copy number value and region-specific copy number value for a plurality of regions are then estimated.

Claims

1. A method for determining copy number variation of a genomic region at a detection position of a target polynucleotide sequence in a sample, said method comprising:

obtaining, using a computer, coverage values for each given position in a baseline or reference sample for the sequence coverage of said target polynucleotide using data generated from mate-pair mappings;

correcting, using the computer, the coverage values for each given position for sequence coverage bias, wherein correcting the coverage values for each given position in a baseline or reference sample comprises performing ploidy-aware baseline correction; and

estimating, using the computer, a total copy number value and region-specific copy number value for each of a plurality of genomic regions based at least on the corrected coverage values for each given position in a baseline or reference sample.

2. The method of claim 1, wherein the method further comprises performing Hidden Markov Model (HMM) segmentation, scoring, and output based on the coverage values for each given position in a baseline or reference sample.

3. The method of claim 1, wherein the method further comprises performing population-based no-calling and identification of low-confidence regions.

4. The method of claim 1 wherein said method further comprises normalizing the coverage values for each given position in a baseline or reference sample for the sequence coverage by comparison to sequence data obtained from a baseline sample.

5. The method of claim 1 wherein obtaining the coverage values for each given position in a baseline or reference sample for the sequence coverage comprises measuring sequence coverage depth at every position of the genome.

6. The method of claim 1 wherein correcting the coverage values for each given position in a baseline or reference sample for the sequence coverage bias comprises calculating window-averaged coverage.

7. The method of claim 1 wherein correcting the coverage values for each given position in a baseline or reference sample for the sequence coverage bias comprises performing adjustments to account for GC bias in the nucleic acid templates generated from the sample during the library construction and/or sequencing process.

8. The method of claim 1 wherein correcting the coverage values for each given position in a baseline or reference sample for the sequence coverage bias comprises performing adjustments based on additional weighting factor associated with individual mappings to compensate for bias.

9. The method of claim 1 wherein the sequence coverage, c_i, is determined by

10. The method of claim 1 wherein the sample-comprises at least one of:

a genome;

sequences of a plurality of approximately random fragments of a genome;

a reference genome; and

an assembled sequence whereby coverage values for each given position in a baseline or reference sample can be obtained as mapping data and/or coverage data.

11. The method of claim 2 wherein performing the HMM segmentation further comprises generating an initial model that estimates the number of states and their means based on the overall coverage distribution.

12. The method of claim 11 wherein performing the HMM segmentation comprises optimizing the initial model by performing one or more of modifying the number of states in the model and optimizing the parameters of each state.

13. The method of claim 11 wherein the corrected coverage at position i is:

14. The method of claim 4 wherein normalizing the measurement data comprises determining normalized corrected coverage by using the equation:

15. The method of claim 1 further comprising using sequence coverage estimation to generate mappings of a sequenced fragment to more than one location on the genome, and using confidence measurements on each of the mappings to fractionally attribute said each mapping to each detection location.

16. The method of claim 1 further comprising performing HMM calculations to determine a ploidy number at each detection position.

17. The method of claim 1 further comprising performing HMM calculations to determine a ploidy score at each detection position, said ploidy score representing a confidence that the determined ploidy number at said detection position is correct.

18. The method of claim 1 further comprising performing HMM calculations to determine a CNV-Type-Score at each detection position, the CNV-Type-Score representing a confidence that said determined ploidy number at said detection position correctly indicates decreased ploidy, expected ploidy, or increased ploidy at said detection position.

19. The method of claim 2, wherein a plurality of states of the HMM correspond to respective copy numbers, and wherein if the sample is a normal sample, performing the HMM segmentation, scoring, and output includes:

initializing a mean of an emission distribution of the HMM for each state with copy number N greater than zero to N/2 multiplied by the median of the coverage in a portion of the sample expected to be diploid; and

initializing the mean of the emission distribution for the state with copy number 0 to a positive value smaller than that used for the state with copy number 1.

20. The method of claim 2, wherein plural states of the HMM correspond to respective copy numbers, and wherein if the sample is a tumor sample, performing the HMM segmentation, scoring, and output includes:

estimating the number of states and a mean of each state based on a distribution of the coverage to generate an initial model for the HMM;

optimizing the initial model by modifying the number of states in the model as well as optimizing the parameters of each state; and

modifying the number of states in the model by sequentially adding states to the model and then sequentially removing states, or a combination thereof.

21. The method of claim 20, wherein modifying the initial model comprises:

a) adding a new state between a pair of states if adding said new state improves a likelihood associated with the HMM beyond a first predetermined threshold;

b) repeating (a) recursively between each pair of states until no more additions are possible;

c) removing a state from the HMM if removal of said state does not decrease the likelihood beyond a second predetermined threshold; and

d) repeating (c) iteratively for all the states.

22. The method of claim 2, wherein plural states of the HMM correspond to respective copy numbers, and wherein performing the HMM segmentation, scoring, and

output includes initializing a variance of an emission distribution of the HMM for each state with copy number N to a constant multiplied by a mean of the emission distribution for said state.

23. The method of claim 10 wherein obtaining the coverage values for the sequence coverage comprises:

a) determining reads that represent the sequences of a plurality of approximately random fragments of the genome of the sample, wherein said plurality provides a sampling of the genome whereby on average a base position of the genome is sampled one or more times;

b) obtaining mapping data by mapping said reads to the reference genome, or by mapping said reads to an assembled sequence; and

c) obtaining coverage data by measuring an intensity of sampled sequences along the reference genome or along the assembled sequence;

wherein the coverage values comprises the mapping data and the coverage data.

24. The method of claim 23 wherein determining the reads further comprises steps of:

a) providing a plurality of amplicons, wherein;

i) each amplicon comprises multiple copies of a fragment of the target polynucleotide sequence,

ii) each amplicon comprises a plurality of interspersed adaptors at predetermined sites within the fragment, each adaptor comprising at least one anchor probe hybridization site,

iii) said plurality of amplicons comprise fragments that substantially cover target nucleic acids in the sample;

b) providing a random array of said amplicons fixed to a surface at a density such that at least a majority of said amplicons are optically resolvable;

c) hybridizing one or more anchor probes to said random array;

d) hybridizing one or more sequencing probes to said random array to form perfectly matched duplexes between said one or more sequencing probes and fragments of target nucleic acid;

e) ligating the anchor probes to the sequencing probes;

f) identifying at least one nucleotide adjacent to at least one interspersed adaptor; and

g) repeating steps (c) through (f) until a nucleotide sequence of said target nucleic acid is identified;

wherein steps (a)-(g) are performed by a sequencing machine.

25. A non-transitory computer-readable medium comprising instructions tangibly embodied thereon, the instructions when executed by a computer processor causing the processor to perform the operations of:

obtaining, using the processor, coverage values for each given position in a baseline or reference sample for the sequence coverage of said target polynucleotide using data generated from mate-pair mappings;

correcting, using the processor, the coverage values for each given position for sequence coverage bias, wherein correcting the coverage values for each given position in a baseline or reference sample comprises performing ploidy-aware baseline correction; and

estimating, using the processor, a total copy number value and region-specific copy number value for each of a plurality of genomic regions based at least on the corrected coverage values for each given position in a baseline or reference sample.

26. A non-transitory computer-readable medium comprising instructions tangibly embodied thereon, the instructions when executed by a computer processor causing the processor to perform the operations of:

obtaining, using the processor, coverage values for each given position in a baseline or reference sample for the sequence coverage of said target polynucleotide using data generated from mate-pair mappings;

correcting, using the processor, the coverage values for each given position for sequence coverage bias, wherein correcting the coverage values for each given position in a baseline or reference sample comprises performing ploidy-aware baseline correction; and

performing Hidden Markov Model (HMM) segmentation, scoring, and output based on the corrected coverage values for each given position in a baseline or reference sample;

based on the HMM scoring and output, performing population-based no-calling and identification of low-confidence regions; and

based on the HMM scoring and output, estimating a total copy number value and region-specific copy number value for a plurality of regions.

27. A system of determining copy number variation of a genomic region at a detection position of a target polynucleotide sequence, comprising:

a computer processor; and

a computer-readable storage medium coupled to said processor, the storage medium having instructions tangibly embodied thereon, the instructions when executed by said processor causing said processor to perform the operations of;

obtaining, using the processor, coverage values for each given position in a baseline or reference sample for the sequence coverage of said target polynucleotide using data generated from mate-pair mappings;

correcting, using the processor, the coverage values for each given position for sequence coverage bias, wherein correcting the coverage values for each given position in a baseline or reference sample comprises performing ploidy-aware baseline correction; and

estimating, using a computer, a total copy number value and region-specific copy number value for each of a plurality of genomic regions based at least on the corrected coverage values for each given position in a baseline or reference sample.