MOLECULAR LABEL COUNTING ADJUSTMENT METHODS

• US 20170344866A1
• Filed: 05/25/2017
• Published: 11/30/2017
• Est. Priority Date: 05/26/2016
• Status: Active Grant

• Current Assignee: Becton, Dickinson & Co
• Original Assignee: Cellular Research, Inc. (Becton, Dickinson & Co)
• Inventors: Fan, Jue, Tsai, Jennifer, Shum, Eleen, Deng, Lisha, Fu, Glenn K.
• Granted Patent: US 11,397,882 B2
• Time in Patent Office: Days
• CPC Class Codes: G06K 19/06028 using bar codes
• CPC Class Codes: G16B 30/00 ICT specially adapted for sequence analysis involving nucleotides or amino acids
• 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 40/10 Signal processing, e.g. from mass spectrometry [MS] or from PCR

Abstract

Disclosed herein are methods and systems for determining the numbers of targets. In some embodiments, the method comprise: stochastically barcoding targets using stochastic barcodes; obtaining sequencing data; for one or more of the targets: counting the number of molecular labels with distinct sequences associated with the target in the sequencing data; identifying clusters of molecular labels of the target using directional adjacency; collapsing the sequencing data using the clusters of molecular labels of the target identified; and estimating the number of the target.

Claims

1. A method for determining the numbers of targets, comprising:

(a) stochastically barcoding a plurality of targets using a plurality of stochastic barcodes to create a plurality of stochastically barcoded targets, wherein each of the plurality of stochastic barcodes comprises a molecular label;

(b) obtaining sequencing data of the stochastically barcoded targets; and

(c) for one or more of the plurality of targets;

(i) counting the number of molecular labels with distinct sequences associated with the target in the sequencing data;

(ii) identifying clusters of molecular labels of the target using directional adjacency;

(iii) collapsing the sequencing data obtained in (b) using the clusters of molecular labels of the target identified in (ii); and

(iv) estimating the number of the target, wherein the number of the target estimated correlates with the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) after collapsing the sequencing data in (ii).

3. The method of claim 1, wherein molecular labels of the target within a cluster are within a predetermined directional adjacency threshold of one another.

5. The method of claim 1, wherein the molecular labels of the target within the cluster comprise one or more parent molecular labels and children molecular labels of the one or more parent molecular labels, and wherein the occurrence of the parent molecular label is greater than or equal to a predetermined directional adjacency occurrence threshold.

6. The method of claim 5, wherein the predetermined directional adjacency occurrence threshold is twice the occurrence of a child molecular label less one.

7. The method of claim 1, wherein collapsing the sequencing data obtained in (b) using the clusters of molecular labels of the target identified in (ii) comprises:

attributing the occurrence of the child molecular label to the parent molecular label.

8. The method of claim 1, further comprising:

determining a sequencing depth of the target.

9. The method of claim 8, wherein estimating the number of the target, if the sequencing depth of the target is above a predetermined sequencing depth threshold, comprises adjusting the sequencing data counted in (i).

11. The method of claim 9, wherein adjusting the sequencing data counted in (i) comprises:

thresholding molecular labels of the target to determine true molecular labels and false molecular labels associated with the target in the sequencing data obtained in (b).

12. The method of claim 11, wherein thresholding the molecular labels of the target comprises performing a statistical analysis on the molecular labels of the target.

13. The method of claim 12, wherein performing the statistical analysis comprises:

fitting the distribution of the molecular labels of the target and their occurrences to two negative binomial distributions;

determining the number of true molecular labels n using the two negative binomial distributions; and

removing the false molecular labels from the sequencing data obtained in (b), wherein the false molecular labels comprise molecular labels with occurrences lower than the occurrence of the nth most abundant molecular label, and wherein the true molecular labels comprise molecular labels with occurrences greater than or equal to the occurrence of the nth most abundant molecular label.

14. The method of claim 13, wherein the two negative binomial distributions comprise a first negative binomial distribution corresponding for the true molecular labels and a second negative binomial distribution for the false molecular labels.

2. (canceled)

4. (canceled)

10. (canceled)

15. A method for determining the numbers of targets, comprising:

(a) stochastically barcoding a plurality of targets using a plurality of stochastic barcodes to create a plurality of stochastically barcoded targets, wherein each of the plurality of stochastic barcodes comprises a molecular label;

(b) obtaining sequencing data of the stochastically barcoded targets; and

(c) for one or more of the plurality of targets;

(i) counting the number of molecular labels with distinct sequences associated with the target in the sequencing data;

(ii) determining a number of noise molecular labels with distinct sequences associated with the target in the sequencing data; and

(iii) estimating the number of the target, wherein the number of the target estimated correlates with the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) adjusted according to the number of noise molecular labels determined in (ii).

16. The method of claim 15, further comprising determining a sequencing status of the target in the sequencing data.

17. The method of claim 16, wherein the sequencing status of the target in the sequencing data is saturated sequencing, under sequencing, or over sequencing.

18. The method of claim 17, wherein the saturated sequencing status is determined by the target having a number of molecular labels with distinct sequences greater than a predetermined saturation threshold.

21. The method of claim 17, wherein, if the sequencing status of the target in the sequencing data is the saturated sequencing status, the number of noise molecular labels determined in (ii) is zero.

22. The method of claim 17, wherein the under sequencing status is determined by the target having a depth smaller than a predetermined under sequencing threshold, wherein the depth of the target comprises an average, a minimum, or a maximum depth of the molecular labels with distinct sequences associated with the target in the sequencing data.

24. The method of claim 22, wherein the under sequencing threshold is independent of the number of molecular labels with distinct sequences.

25. The method of claim 17, wherein, if the sequencing status of the target in the sequencing data is the under sequencing status, the number of noise molecular labels determined in (ii) is zero.

26. The method of claim 17, wherein the over sequencing status is determined by the target having a depth greater than a predetermined over sequencing threshold, wherein the depth of the target comprises an average, a minimum, or a maximum depth of the molecular labels with distinct sequences associated with the target in the sequencing data.

28. The method of claim 26, further comprising, if the sequencing status of the target in the sequencing data is the over sequencing status:

subsampling the number of molecular labels with distinct sequences associated with the target in the sequencing data to about the predetermined over sequencing threshold.

29. The method of claim 17, wherein determining the number of noise molecular labels with distinct sequences associated with the target in the sequencing data comprises:

if a negative binomial distribution fitting condition is satisfied,

(iv) fitting a signal negative binomial distribution to the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i), wherein the signal negative binomial distribution corresponds to a number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) being signal molecular labels;

(v) fitting a noise negative binomial distribution to the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i), wherein the noise negative binomial distribution corresponds to a number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) being noise molecular labels; and

(vi) determining the number of noise molecular labels using the signal negative binomial distribution fitted in (v) and the noise negative binomial distribution fitted in (vi).

30. The method of claim 29, wherein the negative binomial distribution fitting condition comprises:

the sequencing status of the target in the sequencing data is not the under sequencing status or the over sequencing status.

31. The method of claim 29, determining the number of noise molecular labels using the signal negative binomial distribution fitted in (v) and the noise negative binomial distribution fitted in (vi) comprises:

for each of the distinct sequences associated with the target in the sequencing data;

determining a signal probability of the distinct sequence to be in the signal negative binomial distribution;

determining a noise probability of the distinct sequence to be in the noise negative binomial distribution; and

determining the distinct sequence to be a noise molecular label if the signal probability is smaller than the noise probability.

32. The method of claim 17, wherein determining the number of noise molecular labels with distinct sequences associated with the target in the sequencing data comprises:

adding pseudopoints to the number of molecular labels with distinct sequences associated with the target in the sequencing data prior to determining the number of noise molecular labels with distinct sequences associated with the target in the sequencing data in (ii), if the sequencing status of the target in the sequencing data is not the under sequencing status or the over sequencing status and the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) is less than a pseudopoints threshold.

34. The method of claim 17, wherein determining the number of noise molecular labels with distinct sequences associated with the target in the sequencing data comprises:

removing non-unique molecular labels when determining the number of noise molecular labels with distinct sequences associated with the target in the sequencing data in (ii), if the sequencing status of the target in the sequencing data is not the under sequencing status or the over sequencing status and the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) is not less than a pseudopoints threshold.

35. The method of claim 34, wherein removing the non-unique molecular labels comprises removing the non-unique molecular labels when determining the number of noise molecular labels with distinct sequences associated with the target in the sequencing data in (ii) if the number of molecular labels with distinct sequences associated with the target in the sequencing data is greater than a predetermined recycled molecular label threshold.

37. The method of claim 34, wherein removing the non-unique molecular labels comprises:

determining a theoretical number of non-unique molecular labels for the number of molecular labels with distinct sequences associated with the target in the sequencing data; and

removing a molecular label with an occurrence greater than the nth most abundant molecular label of the molecular labels with distinct sequences associated with the target in the sequencing data, wherein n is the theoretical number of non-unique molecular labels.

19-20. -20. (canceled)

23. (canceled)

27. (canceled)

33. (canceled)

36. (canceled)

38-39. -39. (canceled)

40. A method for determining the numbers of targets, comprising:

(a) stochastically barcoding a plurality of targets using a plurality of stochastic barcodes to create a plurality of stochastically barcoded targets, wherein each of the plurality of stochastic barcodes comprises a molecular label;

(b) obtaining sequencing data of the stochastically barcoded targets; and

(c) for one or more of the plurality of targets;

(i) counting the number of molecular labels with distinct sequences associated with the target in the sequencing data;

(ii) determining a quality status of the target in the sequencing data obtained in (b);

(iii) determining one or more sequencing data errors in the sequencing data obtained in (b), wherein determining the one or more sequencing data errors in the sequencing data comprises determining one or more of;

the number of the molecular labels with distinct sequences associated with the target in the sequencing data, the quality status of the target in the sequencing data, and the number of molecular labels with distinct sequences in the plurality of stochastic barcodes; and

(iv) estimating the number of the target, wherein the number of the target estimated correlates with the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) adjusted according to the one or more sequencing data errors determined in (iii).

41-110. -110. (canceled)

111. A method for determining the numbers of targets, comprising:

(a) stochastically barcoding a plurality of targets using a plurality of stochastic barcodes to create a plurality of stochastically barcoded targets, wherein each of the plurality of stochastic barcodes comprises a molecular label;

(b) obtaining sequencing data of the stochastically barcoded targets; and

(c) for one or more of the plurality of targets;

(i) counting the number of molecular labels with distinct sequences associated with the target in the sequencing data;

(ii) identifying clusters of molecular labels of the target using directional adjacency;

(iii) collapsing the sequencing data obtained in (b) using the clusters of molecular labels of the target identified in (ii); and

(iv) estimating the number of the target, wherein the number of the target estimated correlates with the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (i) after collapsing the sequencing data in (ii).

112-126. -126. (canceled)