Method of analysing a representation of a separation pattern

• US 7,747,049 B2
• Filed: 08/26/2005
• Issued: 06/29/2010
• Est. Priority Date: 07/15/2005
• Status: Active Grant

• Current Assignee: Biosignatures Limited, WesternGeco LLC
• Original Assignee: Biosignatures Limited
• Inventors: Bramwell, David, Morns, Ian, Kapferer, Anna
• Primary Examiner(s): Carter; Aaron W
• Application Number:
  US11/212,478
• Publication Number:
  US 20070014450A1
• Time in Patent Office: 1,768 Days
• Field of Search: 382/128, 382129-134, 702 19- 21
• US Class Current: 382/128
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• CPC Class Codes: G16B 45/00 ICT specially adapted for bioinformatics-related data visualisation, e.g. displaying of maps or networks

Abstract

The invention relates principally to the statistical analysis of protein separation patterns. The invention solves the problems associated with producing models which are predictive of classification using unreduced data. The invention provides a method of analysing a representation of a separation pattern, the representation including a neighborhood representing a region of the separation pattern, the neighborhood including a plurality of data points, the method comprising augmenting data by representing the entire region using each data point of the neighborhood; and building a classification model using some or all of the data points.

Claims

1. A method of performing operations, in a computing device, on protein samples for the analysis of a representation of a separation pattern, the representation including a neighborhood representing a region of the separation pattern, the neighborhood including a plurality of data points relating to the protein samples, the method comprising using a computer to perform:

augmenting data by representing the entire region using each data point of the neighborhood; and

building a classification model using some or all of the data points.

2. The method of claim 1 wherein augmenting data includes, in a computing device, separating the data points of the neighborhood into respective recombined representations, the location of a data point in a recombined representation corresponding to the location of its neighborhood in the representation.

3. The method of claim 2, comprising associating, in a computing device, each recombined representation with a particular location of a data point in a neighborhood.

4. The method of claim 3, including varying the association of recombined representations with locations between neighborhoods.

5. The method of claim 4, including varying a position of a particular neighborhood in relation to the representation.

6. The method of claim 1 including comparing, in a computing device, a first distribution built using neighborhoods of a first size to a second distribution built using neighborhoods of a second size and selecting criteria based on the first and second sizes if a statistical change in the distributions is detected.

7. The method of claim 6 wherein F-ratio or student t tests are used to determine a probability threshold based on the probability that the distributions significantly differ in means.

8. The method of claim 6 wherein non-parametric tests are used.

9. The method of claim 1 including building a distribution in respect of each data point, the distribution including values of neighboring data points.

10. The method of claim 9 including determining, in a computing device, a number of data points the values of which are unlikely given the distributions associated with those data points.

11. The method of claim 9 including determining, in a computing device, an absolute difference between the actual value of each data point and a value predicted using the distribution associated with that data point.

12. A method of analyzing, in a computing device, a representation of a separation pattern, the representation including a neighborhood representing a region of the separation pattern, the neighborhood including a plurality of data points, the method comprising using a computer to perform:

augmenting data by representing the entire region using each data point of the neighborhood; and

building a classification model using some or all of the data points.

13. The method of claim 12 wherein augmenting data includes separating, in a computing device, the data points of the neighborhood into respective recombined representations, the location of a data point in a recombined representation corresponding to the location of its neighborhood in the representation.

14. The method of claim 13, comprising associating each recombined representation with a particular location of a data point in a neighborhood.

15. The method of claim 14, including varying the association of recombined representations with locations between neighborhoods.

16. The method of claim 13, including varying the position of a particular neighborhood in relation to the representation.

17. The method of claim 12 including comparing, in a computing device, a first distribution built using neighborhoods of a first size to a second distribution built using neighborhoods of a second size and selecting criteria based on the first and second sizes if a statistical change in the distributions is detected.

18. The method of claim 17, wherein F-ratio or student t tests are used to determine a probability threshold based on the probability that the distributions significantly differ in means.

19. The method of claim 12 wherein non-parametric tests are used.

20. The method of claim 12 including building, in a computing device, a distribution in respect of each data point, the distribution including values of neighboring data points.

21. The method of claim 20 including determining, in a computing device, a number of data points the values of which are unlikely given the distributions associated with those data points.

22. The method of claim 21 including determining, in a computing device, an absolute difference between the actual value of each data point and a value predicted using the distribution associated with that data point.

23. Apparatus for performing operations on protein samples for the analysis of a representation of a separation pattern, the representation including a neighborhood representing a region of the separation pattern, the neighborhood including a plurality of data points, the apparatus comprising means for augmenting data by representing the entire region using each data point of the neighborhood;

and building a classification model using some or all of the data points.

24. Apparatus for analyzing a representation of a separation pattern, the representation including a neighborhood representing a region of the separation pattern, the neighborhood including a plurality of data points, the apparatus comprising means for augmenting data by representing the entire region using each data point of the neighborhood;

and building a classification model using some or all of the data points.

25. A non-transitory computer-readable medium having computer executable instructions for performing a method of performing operations on protein samples for the analysis of a representation of a separation pattern, the representation including a neighborhood representing a region of the separation pattern, the neighborhood including a plurality of data points, the method comprising augmenting data by representing the entire region using each data point of the neighborhood;

and building a classification model using some or all of the data points.

26. The computer-readable medium of claim 25, wherein augmenting data includes separating the data points of the neighborhood into respective recombined representations, the location of a data point in a recombined representation corresponding to the location of its neighborhood in the representation.

27. The computer-readable medium of claim 26, comprising a computer executable instruction to associate each recombined representation with a particular location of a data point in a neighborhood.

28. The computer-readable medium of claim 27, comprising a computer executable instruction to vary at least one of a group consisting of the association of recombined representations with locations between neighborhoods and the position of a particular neighborhood in relation to the representation.

29. The computer-readable medium of claim 25 comprising a computer executable instructions to compare, in a computing device, a first distribution built using neighborhoods of a first size to a second distribution built using neighborhoods of a second size and to select criteria based on the first and second sizes if a statistical change in the distributions is detected.