Methods for estimating probe cell locations in high-density synthetic DNA microarrays

• US 6,993,173 B2
• Filed: 09/30/2002
• Issued: 01/31/2006
• Est. Priority Date: 10/12/2001
• Status: Expired due to Fees

• Current Assignee: Duke University
• Original Assignee: Duke University
• Inventors: Zuzan, Harry, Johnson, Valen E.
• Primary Examiner(s): Patel, Kanjibhai
• Assistant Examiner(s): Tabatabai, Abolfazl
• Application Number:
  US10/262,488
• Publication Number:
  US 20030073896A1
• Time in Patent Office: 1,219 Days
• Field of Search: 382/128, 382/129, 382/130, 382/131, 382/132, 382/133, 382/134, 600/352, 600/407, 600/427, 600/562, 424/1.11, 424/1.17, 424/1.41, 424/1.45, 435/5, 435/6, 435/69.1, 435/91.4, 435/333, 435/252.3, 435/477, 436/94, 436/63, 436/508, 436/519.22, 702/20, 702/27, 702/47
• US Class Current: 382/129
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06T 2207/30072 Microarray; Biochip, DNA array; Well plate
• CPC Class Codes: G06T 7/74 involving reference images or patches

Abstract

Methods, systems, and computer program products for estimating the location of a probe cell in an image of a high-density microarray DNA chip interrogate a plurality of different closely spaced estimated locations to identify the most likely estimated location of the probe cell in the image.

Claims

1. A method for estimating the probe cell locations in a hybridized microarray, comprising:

obtaining an image of the intensity of a microarray having a plurality of individual probe cells;

estimating the center of individual probe cell locations in the image to provide the first estimated locations of the centers of estimated probe cell locations; and

for respective probe cell locations undergoing analysis;

generating a first fitting region that has a center that coincides with the location of the first estimated center and that has a boundary that can include partial pixel volumes;

obtaining pixel intensity data of the first fitting region;

generating a second fitting region that has a center that is spaced apart from the center of the first fitting region by less than about one pixel, the second fitting region having a boundary that can include partial pixel volumes;

obtaining pixel intensity data of the second fitting region;

comparing the variance of pixel intensity from the first and second fitting regions; and

determining the likely estimated probe cell location in the image based on the step of comparing the pixel intensity data in the first and second fitting regions.

2. A method according to claim 1, further comprising:

generating a third fitting region that has a center which is spaced apart from the centers of the first and second fitting regions by less than about one pixel, and

obtaining pixel intensity data of the third fitting region, and wherein the step of comparing considers the variance of the pixel intensity data within each of the first, second, and third fitting regions.

3. A method according to claim 2, wherein said first, second, and third fitting regions share a majority of the same pixels in the image.

4. A method according to claim 1, wherein the first and second fitting regions have respective centers which are spaced apart by about 0.5 pixels or less.

5. A method according to claim 1, wherein said steps of obtaining pixel intensity data from the first and second fitting regions comprises calculating partial pixel volumes of pixels dissected by the boundary of the fitting regions, and calculating pixel variance for the pixel data from the respective fitting regions.

6. A method according to claim 1, further comprising generating an array defining spaced apart potential prospective estimated center locations of a respective probe cell location undergoing analysis, wherein a plurality of fitting regions are generated, the number defined by the quantity of points in the array, a respective one fitting region corresponding to one of the estimated center locations defined by the array.

7. A method according to claim 6, wherein the array includes the first estimated center location as one of its estimated center locations, and wherein the first estimated center location can be an initial estimated center location or a subsequent current estimated location identified by said step of determining.

8. A method according to claim 7, wherein the array is a symmetrical array of evenly spaced estimated center locations.

9. A method according to claim 8, wherein the array is configured so that the first estimated center location is in the center.

10. A method according to claim 8, wherein the array is a 3×

3 array.

11. A method according to claim 7, wherein the array is circular with an array center, and a plurality of prospective estimated center locations disposed circumferentially evenly spaced about the perimeter thereof.

12. A method according to claim 6, wherein the array comprises at least six discrete estimated center locations points.

13. A method according to claim 6, wherein said step of determining the estimated probe cell location, comprises interrogating each of the fitting regions associated with the prospective estimated potential probe cell center locations defined by the array, and

assigning a first penalty value for proposed estimated probe cell locations centered at each of the proposed center locations in the array for centers being misaligned with the estimated centers of neighboring probe cell locations, the penalty being greater for greater degrees of misalignment;

assigning a second penalty value for intensity variance within the proposed estimated probe cell locations centered at each of the proposed center locations in the array, the penalty being greater for greater degrees of variance; and

determining which of the proposed estimated probe cell locations corresponding to the proposed centers in the array has the lowest penalty to thereby identify the estimated center location and the estimated probe cell location.

14. A method according to claim 13, wherein the pixel intensity variation within each estimated probe cell location defined by its corresponding fitting region having one of the proposed estimated centers of the array as its center is calculated by using log transformed intensity data averaged over the pixel area associated with the respective fitting region.

15. A method according to claim 1, wherein the first estimated center location is an initial estimated center location defined by the centroid of the centers of the neighboring probe cell locations.

16. A method according to claim 1, wherein the first and second fitting regions are the same size, share a major number of the same pixels in the image, and cover a subset of the estimated probe cell location in the image.

17. A method according to claim 1, wherein said determining step comprises considering the extent of misalignment of the estimated center location relative to the estimated center locations of its neighboring probe cell locations in the image.

18. A method according to claim 1, wherein said method further comprises logarithmically transforming the intensity data associated with the obtained pixel intensities of the first and second fitting regions.

19. A method according to claim 1, further comprising generating a relational data set that provides the average of log transformed pixel intensity data within the first and second fitting regions for a corresponding probe cell location in the image.

20. A method for estimating the probe cell locations in an image of a hybridized high-density microarray having a plurality of individual hybridized probe cells thereon, comprising:

providing a first estimated center location of a respective probe cell location in the image;

generating an array of a plurality of potential estimated center locations such that the array includes the first estimated center as one of the estimated center locations;

interrogating a plurality of estimated probe cell locations, a respective one probe cell location having a center defined by one of the plurality of potential estimated center locations in the array to obtain pixel intensity data of pixels within the estimated probe cell locations; and

comparing the intensity data for each of the interrogated estimated probe cell locations to determine a likely estimated or revised estimated probe cell location.

21. A method according to claim 20, wherein said interrogating step comprises generating fitting regions having associated centers, one of which coincides with a respective one of the estimated centers of the array, the fitting regions having a size selected to sample a subset of the probe cell location in the image, the fitting regions being configured to be able to include partial pixels at their boundaries.

22. A method according to claim 21, wherein said comparing step comprises logarithmically transforming the intensity data.

23. A method according to claim 21, wherein the fitting regions are sized the same and share a major portion of the same pixels in the image.

24. A method according to claim 21, further comprising determining partial pixel volumes of the pixels dissected by the boundaries of the fitting regions and evaluating which of the proposed center estimate locations associated with the array provides the best new estimate of the center of the estimated probe cell location.

25. A method according to claim 20, wherein, the interrogated estimated probe cell locations have centers which are spaced apart by less than about 1 pixel.

26. A method according to claim 20, wherein the array is defined by a symmetrically spaced apart plurality of points.

27. A method according to claim 20, wherein said comparing step comprises analyzing the variation in the intensity of pixels associated with estimated probe cell locations.

28. A method according to claim 27, further comprising:

assigning a first penalty value for those proposed estimated probe cell locations interrogated having centers which are misaligned with neighboring probe cell locations, the penalty being greater for greater degrees of misalignment;

assigning a second penalty value for intensity variance within the proposed estimated probe cell locations, the penalty being greater for greater degrees of variance; and

determining which proposed center in the array has the lowest penalty to thereby identify the new center location for the estimated probe cell location.

29. A computer program product stored in computer readable medium for estimating the probe cell locations of probe cells in an image of an expressed hybridized DNA microarray chip, the computer program product comprising:

a computer readable storage medium having computer readable program code embodied in said medium, said computer-readable program code comprising;

computer readable program code for obtaining an image of the intensities of a microarray having a plurality of individual probe cells;

computer readable program code for providing a first estimated center location of a respective probe cell location in the image;

computer readable program code for generating an array of a plurality of prospective estimated center locations for a respective probe cell such that the array includes the first estimated center as one of the estimated center locations;

computer readable program code for generating a fitting region for each of the plurality of plurality of prospective estimated center locations;

computer readable program code for interrogating the plurality of fitting regions to obtain pixel intensity data of pixels within estimated probe cell locations corresponding to the fitting regions; and

computer readable program code for comparing the intensity data for each of the interrogated estimated probe cell locations to determine a likely estimated or revised estimated probe cell location.

30. A computer program product according to claim 29, wherein said computer program product further comprises computer readable program code for carrying out the interrogation by generating fitting regions having associated centers, one of which coincides with a respective one of the estimated centers of the array, the fitting regions having a size selected to sample a subset of the probe cell location in the image, the fitting regions being configured to be able to include partial pixels at their boundaries.

31. A computer program product according to claim 30, wherein said computer program product further comprises computer readable program code for determining partial pixel volumes of the pixels dissected by the boundaries of the fitting regions and evaluating which of the proposed center estimate locations associated with the array provides the best new estimate of the center of the estimated probe cell location.

32. A computer program product according to claim 29, wherein said computer program product further comprises computer readable program code for logarithmically transforming the intensity data of individual probe cells undergoing analysis.

33. A computer program product according to claim 32, wherein the potential estimated centers so as to be spatially offset from the another estimated center location in the array by less than one pixel.

34. A computer program product according to claim 29, wherein the perimeter of the fitting regions may dissect pixels, and wherein the computer program code for interrogating the plurality of fitting regions to obtain pixel intensity data can consider partial pixel volumes.

35. A computer program product according to claim 29, wherein said computer program product further comprises computer readable program code for calculating and analyzing the variation in the intensity of pixels associated with estimated probe cell locations.

36. A computer program product according to claim 29, wherein said computer program product further comprises computer readable program code for assigning a first penalty value for those proposed estimated probe cell locations interrogated having centers which are misaligned with neighboring probe cell locations, the penalty being greater for greater degrees of misalignment,

computer readable program code for assigning a second penalty value for intensity variance within the proposed estimated probe cell locations, the penalty being greater for greater degrees of variance;

and

computer readable program code for determining which proposed center in the array has the lowest penalty to thereby identify the new center location for the estimated probe cell location.

37. A system for estimating probe cell location in an image of a hybridized DNA array, comprising:

a processor; and

computer program code for estimating probe cell locations in an image of a hybridized microarray comprising, for a respective probe cell location in an image, interrogating a plurality of estimated probe cell locations by generating first and second fitting regions having centers which are spaced apart by less than about one pixel, the first and second fitting regions being able to include partial pixels at their boundaries, analyzing the pixel intensity variance within each of the first and second fitting regions, and selecting the potential estimated location for that probe cell based thereon.