Method and system for detecting discriminatory data patterns in multiple sets of data

• US 20050022168A1
• Filed: 06/14/2004
• Published: 01/27/2005
• Est. Priority Date: 06/11/2003
• Status: Active Grant

• Current Assignee: The Research Foundation for The State University of New York (State University of New York)
• Original Assignee: The Research Foundation for The State University of New York (State University of New York)
• Inventors: Zhu, Wei, Kovach, John S., Wang, Xuena
• Granted Patent: US 8,478,534 B2
• Time in Patent Office: Days
• US Class Current: 717/124
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• 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
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: H01J 49/0036 Step by step routines describing the handling of the data generated during a measurement
• CPC Class Codes: Y02A 90/10 Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

A comprehensive analysis procedure for analyzing and comparing multiple sets of data to detect hidden discriminatory data patterns. The inventive procedure identifies a best subset of markers for optimal discrimination between two or more sets of data. A point-wise test on two or more sets of data is performed to calculate test statistic values and to generate a statgram, a two- or higher- dimensional map of the test statistic values along the range of data. A threshold is then determined for isolating critical regions of the statgram at each significance level to provide candidate markers. A subset of markers from the candidate markers is then selected to discriminate among the sets of data. The two or more sets of data are classified using the subset of markers.

Claims

1. A method for detecting discriminatory data patterns in multiple sets of data, each unit of data being described uniquely by one or more coordinates, comprising:

performing a point-wise test on a training set of data to calculate a plurality of test statistic values for corresponding units of data from two or more groups of data representing known conditions;

determining a threshold test statistic value based on a selected significance level;

selecting those units of data having a test statistic value with an absolute value that exceeds the threshold, the selected units of data comprising candidate marker elements;

selecting a subset of marker elements from the candidate marker elements to discriminate among the two or more groups; and

classifying a testing set of data into the two or more groups using the subset of marker elements.

2. The method of claim 1, wherein classifying includes determining a quantitative difference in the subset of marker elements between the multiple sets of data.

3. The method of claim 1, comprising standardizing the multiple sets of data.

4. The method of claim 1, comprising randomly selecting the training set of data from the multiple sets of data.

5. The method of claim 1, comprising randomly selecting the testing set of data from the multiple sets of data.

6. The method of claim 1, comprising presmoothing the multiple sets of data with a Gaussian kernel.

7. The method of claim 1, wherein the sets of data comprise spectral data.

8. The method of claim 1, wherein determining a threshold includes performing multiple-test correction of the multiple sets of data.

9. The method of claim 8, wherein performing multiple-test correction includes performing at least one of a random field theory, a Gaussian random field, and a t random field multiple-test correction.

10. The method of claim 1, wherein selecting a subset of marker elements includes eliminating any duplicative marker elements.

11. The method of claim 1, comprising checking the candidate marker elements for variance stability.

12. The method of claim 1, comprising checking the subset of marker elements for variance stability.

31. The method of claim 1, wherein performing a point-wise test includes performing at least one of a z-test and a t-test.

32. The method of claim 1, wherein performing a point-wise test includes performing at least one of an ANOVA and an ANCOVA F-test.

33. The method of claim 8, wherein performing multiple-test correction includes performing an F random field theory multiple-test correction.

34. The method of claim 1, wherein classifying includes determining a probability that a subject belongs to a category.

35. The method of claim 34, wherein classifying includes performing a likelihood ratio test.

36. The method of claim 34, wherein classifying includes determining a Mahalanobis distance.

37. The method of claim 1, comprising clustering data.

38. The method of claim 1, comprising clustering marker elements.

13. A method of detecting a disease from data of biological samples, comprising:

randomly sampling data that has been standardized and smoothed to divide the data into a training set and a testing set, each set comprising random samples from subjects affected and unaffected by the disease;

performing a point-wise test on the training set of data to determine test statistic values indicative of the difference between corresponding data values of the samples of the affected and the unaffected subjects;

determining a threshold test statistic value based on a selected significance level;

selecting those data values having a test statistic value with an absolute value that exceeds the threshold, the selected units of data comprising candidate marker elements;

selecting a subset of marker elements from the candidate marker elements to discriminate between the affected and the unaffected samples of the training set; and

classifying the testing set of data as representing affected or unaffected samples using the subset of marker elements.

14. The method of claim 13, comprising classifying data of a biological sample as being affected or unaffected using the subset of markers.

15. The method of claim 13, wherein the data comprises spectral intensity values in a spectrum for each of the biological samples, the method comprising pre-processing the data to standardize and smooth the spectral intensity values across the spectrum for each of the biological samples.

16. The method of claim 15, wherein pre-processing includes presmoothing the data with a Gaussian kernel.

17. The method of claim 13, wherein the test statistic value t(x) for the corresponding data values y₁(x) and y₂(x) corresponding to x is:

$t<br><br>\left(x\right)=\frac{\stackrel{\_}{y_1<br><br>\left(x\right)}-\stackrel{\_}{y_2<br><br>\left(x\right)}}{\sqrt{s_1^2<br><br>\left(x\right)/n_1+s_2^2<br><br>\left(x\right)/n_2}}$

where {overscore (y)}₁ (x), {overscore (y)}₂ (x), s₁²(x), and s₂²(x) are the means and variances, respectively, of the affected and unaffected samples of the training set and n₁ and n₂ are the numbers of affected and unaffected samples, respectively, in the training set.

18. The method of claim 13, wherein determining a threshold test statistic value includes performing multiple-test correction of the data of the training set.

19. The method of claim 13, wherein selecting a subset of marker elements includes eliminating any duplicative marker elements.

20. The method of claim 13, comprising checking the candidate marker elements for variance stability.

21. The method of claim 13, comprising checking the subset of marker elements for variance stability.

22. The method of claim 13, wherein classifying the testing set includes using normal kernel density estimates with unequal bandwidth of the subset of marker elements.

23. The method of claim 14, wherein classifying data of the biological sample includes determining a change in a state of the disease using the subset of marker elements.

24. A system for detecting discriminatory data patterns in multiple sets of data, each unit of data being described uniquely by one or more coordinates, comprising:

a point-wise test module for performing a point-wise test on a training set of data to calculate a plurality of test statistic values for corresponding units of data from two or more groups of data representing known conditions;

a threshold module for determining a threshold test statistic value based on a selected significance level and selecting those units of data having a test statistic value with an absolute value that exceeds the threshold, the selected units of data comprising candidate marker elements;

a marker selection module for selecting a subset of marker elements from the candidate marker elements to discriminate among the two or more groups; and

a classification module for classifying a testing set of data into the two or more groups using the subset of marker elements.

25. The system of claim 24, comprising a variance stability module for checking at least one of the candidate marker elements and the subset of marker elements for variance stability.

26. The system of claim 24, wherein the threshold module performs a multiple-test correction of the data of the training set.

27. The system of claim 24, comprising a pre-processing module for presmoothing the multiple sets of data with a Gaussian kernel.

28. A system for detecting a disease from data of biological samples, comprising:

a sampling module for randomly sampling data that has been standardized and smoothed to divide the data into a training set and a testing set, each set comprising random samples from subjects affected and unaffected by the disease;

a point-wise test module for performing a point-wise test on the training set of data to determine test statistic values indicative of the difference between corresponding data values of the samples of the affected and the unaffected subjects;

a threshold module for determining a threshold test statistic value based on a selected significance level and for selecting those data values having a test statistic value with an absolute value that exceeds the threshold, the selected units of data comprising candidate marker elements;

a marker selection module for selecting a subset of marker elements from the candidate marker elements to discriminate between the affected and the unaffected samples of the training set; and

a classification module for classifying the testing set of data as representing affected or unaffected samples using the subset of marker elements.

29. The system of claim 28, wherein the data comprises spectral intensity values in a spectrum for each of the biological samples, the system comprising a pre-processing module for standardizing and smoothing the spectral intensity values across the spectrum for each of the biological samples.

30. A computer readable medium comprising code for detecting discriminatory data patterns in multiple sets of data, each unit of data being described uniquely by one or more coordinates, the code comprising instructions for:

performing a point-wise test on a training set of data to calculate a plurality of test statistic values for corresponding units of data from two or more groups of data representing known conditions;

determining a threshold test statistic value based on a selected significance level and selecting those units of data having a test statistic value with an absolute value that exceeds the threshold, the selected units of data comprising candidate marker elements;

selecting a subset of marker elements from the candidate marker elements to discriminate among the two or more groups; and

classifying a testing set of data into the two or more groups using the subset of marker elements.