Reduction of bias in genomic coverage measurements

1. A method of characterizing a sample, the method comprising:

labeling a plurality of sample molecules with a first label, wherein the sample molecules comprise a genome or genomic fragment;

translocating the plurality of labeled sample molecules though a fluidic channel;

detecting counts of signals from the labeled sample molecules so as to ascertain a pattern or plurality of patterns characteristic of the genome or genomic fragment;

correlating signals from the labeled sample molecules to a reference to ascertain a coverage depth of signals of one or more regions of the genome or genomic fragment by the sample molecules;

scaling the coverage depths of signals to a subset of coverage depths of signals corresponding to regions of the genome or genomic fragment that does not comprise a sex chromosome or a fragment thereof, thereby providing scaled coverage depths; and

normalizing the scaled coverage depths by one, two, or three of:

characteristic molecular length of the plurality of labeled sample molecules; or

characteristic number of labels per interval for a plurality of intervals of the reference, wherein the reference comprises a plurality of intervals; or

characteristic number of labels per molecule or characteristic number of labels within a segment of predetermined length per molecule,

thereby generating a copy number profile of the sample molecules in which bias due to label density and bias due to factors other than label density in the copy number profile are minimized or eliminated;

wherein normalizing the scaled coverage depths comprises obtaining normalized label coverage depths as provided by the formula: n=Q/[E+GC(1/lambda−1/lambda 0 ), wherein n represents normalized label coverage depths, Q represents scaled label coverage depths, G and E respectively represent gradient and zero-order coefficient of a linear regression of scaled label coverage depths versus abscissa for a plurality of samples in a training set, lambda represents characteristic sample-specific molecule length, and lambda 0 represents median characteristic molecule length for the plurality of samples of the training set.

2. The method of claim 1 , wherein the plurality of intervals of the reference are of equal size.

3. The method of claim 1 , further comprising automatically determining a presence or absence of possible structural variation in of the genome or genomic fragment,

wherein automatically determining a presence or absence of possible regional structural variation comprises identifying possible breakpoints in the copy number profile, wherein an interval in the copy number profile with a significantly different copy number than a neighboring interval comprises a possible breakpoint.

4. The method of claim 3 , wherein automatically determining a presence or absence of possible regional structural variation comprises determining GROM copy number breakpoints.

5. The method of claim 4 , further comprising, for each of the copy number breakpoints:

identifying a first region of a reference sequence on a first side of the breakpoint, and masking a second region of the reference sequence on a second side of the breakpoint, wherein the second side is opposite the first side; and

scoring only single molecule alignments that align with the reference in the first region.

6. The method of claim 1 , wherein the copy number profile is generated in less than five minutes after the signals are detected.

7. A method of characterizing a sample, the method comprising:

labeling a plurality of sample molecules with a first label, wherein the sample molecules comprise a genome or genomic fragment;

translocating the plurality of labeled sample molecules though a fluidic channel;

detecting counts of signals from the labeled sample molecules so as to ascertain a pattern or plurality of patterns characteristic of the genome or genomic fragment;

correlating signals from the labeled sample molecules to a reference to ascertain a coverage depth of signals of one or more regions of the genome or genomic fragment by the sample molecules;

scaling the coverage depths of signals to a subset of coverage depths of signals corresponding to regions of the genome or genomic fragment that does not comprise a sex chromosome or a fragment thereof, thereby providing scaled coverage depths; and

normalizing the scaled coverage depths by one, two, or three of:

characteristic molecular length of the plurality of labeled sample molecules; or

characteristic number of labels per interval for a plurality of intervals of the reference, wherein the reference comprises a plurality of intervals; or

characteristic number of labels per molecule or characteristic number of labels within a segment of predetermined length per molecule,

thereby generating a copy number profile of the sample molecules in which bias due to label density and bias due to factors other than label density in the copy number profile are minimized or eliminated;

wherein normalizing the scaled coverage depths comprises Sex Chromosome Normalization, comprising:

scaling the scaled label coverage depths for a training sample based on the number of sex chromosomes in the training sample; and

normalizing scaled label coverage depths with respect to characteristic molecular length of the plurality of labeled sample molecules, and further comprises dividing normalized label coverage depths by a median of normalized coverage depths for a plurality of sex chromosomes of a training set.

8. The method of claim 7 , wherein normalized label coverage depths for an X chromosome are divided by the median normalized coverage depths for a plurality of female samples of a training set.

9. The method of claim 7 , wherein normalized label coverage depths for a Y chromosome are divided by the median normalized coverage depths for a plurality of male samples of a training set and are further divided by two.

10. The method of claim 7 , wherein the plurality of intervals of the reference are of equal size.

11. The method of claim 7 , wherein the copy number profile is generated in less than five minutes after the signals are detected.

12. The method of claim 7 , further comprising automatically determining a presence or absence of possible structural variation in of the genome or genomic fragment,

wherein automatically determining a presence or absence of possible regional structural variation comprises identifying possible breakpoints in the copy number profile, wherein an interval in the copy number profile with a significantly different copy number than a neighboring interval comprises a possible breakpoint.

13. The method of claim 12 , wherein automatically determining a presence or absence of possible regional structural variation comprises determining GROM copy number breakpoints.

14. The method of claim 13 , further comprising, for each of the copy number breakpoints:

identifying a first region of a reference sequence on a first side of the breakpoint, and masking a second region of the reference sequence on a second side of the breakpoint, wherein the second side is opposite the first side; and

scoring only single molecule alignments that align with the reference in the first region.

15. A method of characterizing a sample, the method comprising:

labeling a plurality of sample molecules with a first label, wherein the sample molecules comprise a genome or genomic fragment;

translocating the plurality of labeled sample molecules though a fluidic channel;

detecting counts of signals from the labeled sample molecules so as to ascertain a pattern or plurality of patterns characteristic of the genome or genomic fragment;

correlating signals from the labeled sample molecules to a reference to ascertain a coverage depth of signals of one or more regions of the genome or genomic fragment by the sample molecules;

scaling the coverage depths of signals to a subset of coverage depths of signals corresponding to regions of the genome or genomic fragment that does not comprise a sex chromosome or a fragment thereof, thereby providing scaled coverage depths; and

normalizing the scaled coverage depths by one, two, or three of:

characteristic molecular length of the plurality of labeled sample molecules; or

characteristic number of labels per interval for a plurality of intervals of the reference, wherein the reference comprises a plurality of intervals; or

characteristic number of labels per molecule or characteristic number of labels within a segment of predetermined length per molecule,

thereby generating a copy number profile of the sample molecules in which bias due to label density and bias due to factors other than label density in the copy number profile are minimized or eliminated;

wherein normalizing the scaled coverage depths per interval comprises obtaining normalized label coverage depths as provided by the formula n=(c−GL)/E, wherein n represents normalized label coverage depths, c represents scaled coverage depth, L represents gradient of a linear regression of scaled coverage depth vs. the number of labels per interval for a plurality of samples in a training set, and G and E respectively represent gradient and zero-order coefficient of a linear regression of scaled coverage depth versus abscissa for a plurality of samples in a training set.

16. The method of claim 15 , wherein the plurality of intervals of the reference are of equal size.

17. The method of claim 15 , wherein the copy number profile is generated in less than five minutes after the signals are detected.

18. The method of claim 15 , further comprising automatically determining a presence or absence of possible structural variation in of the genome or genomic fragment,

wherein automatically determining a presence or absence of possible regional structural variation comprises identifying possible breakpoints in the copy number profile, wherein an interval in the copy number profile with a significantly different copy number than a neighboring interval comprises a possible breakpoint.

19. The method of claim 18 , wherein automatically determining a presence or absence of possible regional structural variation comprises determining GROM copy number breakpoints.

20. The method of claim 19 , further comprising, for each of the copy number breakpoints:

identifying a first region of a reference sequence on a first side of the breakpoint, and masking a second region of the reference sequence on a second side of the breakpoint, wherein the second side is opposite the first side; and

scoring only single molecule alignments that align with the reference in the first region.