Detecting and classifying copy number variation
The invention provides a method for determining copy number variations (CNV) of a sequence of interest in a test sample that comprises a mixture of nucleic acids that are known or are suspected to differ in the amount of one or more sequence of interest. The method comprises a statistical approach that accounts for accrued variability stemming from process-related, interchromosomal and inter-sequencing variability. The method is applicable to determining CNV of any fetal aneuploidy, and CNVs known or suspected to be associated with a variety of medical conditions. CNV that can be determined according to the method include trisomies and monosomies of any one or more of chromosomes 1-22, X and Y, other chromosomal polysomies, and deletions and/or duplications of segments of any one or more of the chromosomes, which can be detected by sequencing only once the nucleic acids of a test sample.
1. A method, implemented using one or more computers comprising one or more processors and system memory, for classifying a copy number variation in a fetal genome, the method comprising:
(a) sequencing cell free DNA from a maternal test sample obtained from a pregnant woman carrying a fetus to provide at least about one million sequence reads in an electronic format;
(b) aligning the at least about one million sequence reads to a reference genome thereby providing sequence tags corresponding to the sequence reads;
(c) identifying a number of those sequence tags that are from one or more bins and determining that a first bin of interest in the fetus harbors a copy number variation, wherein the first bin of interest is in a first chromosome of interest, and wherein the first chromosome of interest is selected from the group consisting of chromosomes 1-22, X, and Y;
(d) determining a first fetal fraction value by a first method that uses information from the sequence tags from the first bin of interest, wherein the first method includes a ploidy assumption that the first bin of interest harbors a complete aneuploidy;
(e) determining a second fetal fraction value by a second method that does not use information from the sequence tags from the first bin of interest; and
(f) repeating (c)-(e) for a plurality of additional bins, and for each bin comparing the first fetal fraction value to the second fetal fraction value and using the comparison to classify the copy number variation of the first chromosome of interest in the genome of the fetus, wherein operation (f) classifies the copy number variation into a classification selected from the group consisting of: a partial chromosomal aneuploidy and a mosaic, wherein the first bin and the plurality of additional bins comprise substantially all bins in the first chromosome of interest, and wherein each bin has a bin size of about 1-100 k base pairs (kbp),
wherein (a)-(f) are performed using the one or more computers.
2. The method of claim 1 , wherein the comparing the first fetal fraction value to the second fetal fraction value comprises determining whether the two fetal fraction values have a difference that is less than or greater than a defined threshold.
3. The method of claim 2 , wherein the two fetal fraction values are determined to have a difference less than the defined threshold, and the method further comprises determining that the ploidy assumption that the first bin of interest harbors a complete aneuploidy is true.
4. The method of claim 2 , wherein the two fetal fraction values are determined to have a difference larger than the defined threshold, and the method further comprises analyzing sequence tag information for the first bin of interest to determine whether (i) the first bin of interest harbors a partial aneuploidy, or (ii) the fetus is a mosaic.
5. The method of claim 4 , wherein analyzing the sequence tag information for the first bin of interest comprises:
dividing the sequence for the first bin of interest into a plurality of sub-bins;
determining that any of said sub-bins contains significantly more a significantly less nucleic acid than one or more other sub-bins as determined by a defined threshold difference; and
determining that the first bin of interest harbors a partial aneuploidy.
6. The method of claim 5 , further comprises determining a locus of the partial aneuploidy on the first bin of interest.
7. The method of claim 6 , wherein determining the locus of the partial aneuploidy on the first bin of interest comprises categorizing sequence tags aligned to the first bin of interest as belonging to sub-bins in the first bin of interest; and counting sequence tags in each sub-bin, and wherein the sub-bins are sub-chromosomal regions.
8. The method of claim 4 , wherein analyzing the sequence tag information for the first bin of interest comprises:
dividing the sequence for the first bin of interest into a plurality of sub-bins;
determining none of said sub-bins contains significantly more or significantly less nucleic acid than one or more other sub-bins as determined by a defined threshold difference; and
determining that the fetus is a mosaic.
9. The method of claim 1 , wherein the sequencing comprises next generation sequencing (NGS), massively parallel sequencing using sequencing-by-synthesis, sequencing-by-synthesis with reversible dye terminators, sequencing-by-ligation, or single molecule sequencing.
10. The method of claim 1 , wherein calculating the first fetal fraction value comprises (i) determining a number of sequence tags from the first bin of interest and normalizing the number of sequence tags from the first bin to account for inter bin bias.
11. The method of claim 10 , wherein the normalizing comprises accounting for GC content of the first bin of interest.
12. The method of claim 11 , wherein the normalizing comprises calculating a bin ratio value according to the equation
BRV
j
=
Tags
j
∑
Tags
km
wherein BRV j is a bin ratio value of the first bin of interest, Tags j is the number of sequence tags that align to the first bin of interest and ΣTags km is the sum of the numbers of sequence tags that align to a defined number of other bins in the maternal test sample, wherein the other bins are selected from one or more bins from the reference sequence and have GC content that are closest to the GC content of the first bin of interest.
13. The method of claim 10 , wherein the normalizing comprises accounting for process-related variability.
14. The method of claim 1 , wherein calculating the second fetal fraction value comprises using information from one or more polymorphisms exhibiting an allelic imbalance in the fetal and maternal nucleic acids of the maternal test sample.
15. The method of claim 1 , wherein calculating the second fetal fraction value comprises determining the number of sequence tags aligning to the X or Y chromosome of male fetal and maternal nucleic acids of the maternal test sample.
16. The method of claim 1 wherein calculating the first fetal fraction value comprises evaluating the following expression:
ff =abs(2 z i CV i )
where ff is the first fetal fraction value, CV i is a coefficient of variation for the first bin of interest across multiple samples, and z i is a z-score for the first bin of interest.
17. The method of claim 1 , further comprising calculating a normalized chromosome value (NCV) from a chromosome dose, wherein the NCV relates to the chromosome dose to the mean or median of a corresponding chromosome dose in a set of qualified samples as:
NCV
iA
=
R
iA
-
R
iU
_
σ
iU
where R iU and σ iU are the estimated mean and standard deviation, respectively, for the i-th chromosome dose in the set of qualified samples, and R iA is the chromosome dose calculated for a chromosome of interest.
18. The method of claim 1 , wherein operations (b) and (c) are performed for at least 20 chromosomes of interest selected from chromosomes 1-22, X, and Y, and wherein the presence or absence of complete fetal chromosomal aneuploidies of the at least 20 chromosomes of interest is determined.
19. The method of claim 1 , wherein the one or more bins from the reference sequence each have a size of about 1-50 kbp.
20. A computer program product comprising a non-transitory computer readable medium on which is provided program instructions for classifying a copy number variation in a fetal genome, the instructions comprising:
(a) code for receiving at least about one million sequence reads obtained by sequencing cell free DNA from a maternal test sample obtained from a pregnant woman carrying a fetus;
(b) code for aligning the at least about one million sequence reads to a reference genome thereby providing sequence tags corresponding to the sequence reads;
(c) code for identifying a number of those sequence tags that are from one or more bins and determining that a first bin of interest in the fetus harbors a copy number variation, wherein the first bin of interest is in a first chromosome of interest, and wherein the first chromosome of interest is selected from the group consisting of chromosomes 1-22, X, and Y;
(d) code for determining a first fetal fraction value by a first method that uses information from the sequence tags from the first bin of interest, wherein the first method includes a ploidy assumption that the first bin of interest harbors a complete aneuploidy;
(e) code for determining a second fetal fraction value by a second method that does not use information from the sequence tags from the first bin of interest; and
(f) code for repeating (c)-(e) for a plurality of additional bins, and for each bin comparing the first fetal fraction value to the second fetal fraction value and using the comparison to classify the copy number variation of the first chromosome of interest in the genome of the fetus, wherein operation (f) classifies the copy number variation into a classification selected from the group consisting of: a partial chromosomal aneuploidy and a mosaic, wherein the first bin and the plurality of additional bins comprise substantially all bins in the first chromosome of interest, and wherein each bin has a bin size of about 1-100 kbp.
21. A system for classifying a copy number variation in a fetal genome, the system comprising: (1) an interface for receiving at least about 10,000 sequence reads from fetal and maternal nucleic acids in a maternal test sample, wherein the sequence reads are provided in an electronic format; (2) memory for storing, at least temporarily, a plurality of said sequence reads; and (3) a processor configured to:
(a) receive at least about one million sequence reads obtained by sequencing cell free DNA from a maternal test sample obtained from a pregnant woman carrying a fetus;
(b) align the at least about one million sequence reads to a reference genome thereby providing sequence tags corresponding to the sequence reads;
(c) identify a number of those sequence tags that are from one or more bins and determining that a first bin of interest in the fetus harbors a copy number variation, wherein the first bin of interest is in a first chromosome of interest, and wherein the first chromosome of interest is selected from the group consisting of chromosomes 1-22, X, and Y;
(d) determine a first fetal fraction value by a first method that uses information from the sequence tags from the first bin of interest, wherein the first method includes a ploidy assumption that the first bin of interest harbors a complete aneuploidy;
(e) determine a second fetal fraction value by a second method that does not use information from the sequence tags from the first bin of interest; and
(f) repeat (c)-(e) for a plurality of additional bins, and for each bin compare the first fetal fraction value to the second fetal fraction value and using the comparison to classify the copy number variation of the chromosome of interest in the genome of the fetus, wherein operation (f) classifies the copy number variation into a classification selected from the group consisting of: complete chromosomal duplications, complete chromosomal multiplications, complete chromosomal deletions, partial chromosomal duplications, partial chromosomal multiplications, partial chromosomal deletions, and mosaics, wherein the first bin and the plurality of additional bins comprise substantially all bins in the first chromosome of interest, and wherein each bin has a bin size of about 1-100 kbp.
22. The method of claim 19 , wherein the one or more bins from the reference sequence each have a size of about 1-10 kbp.