Artificial intelligence and global normalization methods for genotyping

• US 20060224529A1
• Filed: 11/08/2005
• Published: 10/05/2006
• Est. Priority Date: 03/24/2004
• Status: Active Grant

• Current Assignee: Illumina Incorporated
• Original Assignee: Illumina Incorporated
• Inventors: Kermani, Bahram Ghaffarzadeh
• Granted Patent: US 7,467,117 B2
• Time in Patent Office: Days
• US Class Current: 706/13
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• CPC Class Codes: G16B 20/20 Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
• CPC Class Codes: G16B 20/40 Population genetics; Linkage disequilibrium
• 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: G16B 40/20 Supervised data analysis
• CPC Class Codes: G16B 40/30 Unsupervised data analysis

Abstract

Described herein are systems and methods for normalizing data without the use of external controls. Also described herein are systems and methods for analyzing cluster data, such as genotyping data, using an artificial neural network.

Claims

1-29. -29. (canceled)

30. A method of clustering genetic data for n loci, wherein n is an integer greater than one, comprising

(a) obtaining genetic data comprising n sets of first and second signal values related in a coordinate system, wherein said first and second signal values are indicative of the levels of a first and second allele, respectively, at n loci;

(b) comparing fit of said genetic data to each of a plurality of cluster models using an artificial neural network, thereby determining a best fit cluster model; and

(c) assigning said signal values to at least one cluster according to said best fit cluster model.

31. The method of claim 30, wherein if said best fit cluster model contains at least one actual cluster and at least one missing cluster, then using a second artificial neural network to propose a location for said at least one missing cluster.

32. The method of claim 30, wherein said plurality of cluster models comprises at least seven cluster models.

33. The method of claim 30, further comprising training said first or second artificial neural network with an algorithm selected from the group consisting of a genetic algorithm, back-propagation algorithm, Levenberg-Marquardt algorithm and Bayesian algorithm.

34. The method of claim 30, wherein said first or second artificial neural network comprises a 3 layer feed-forward artificial neural network or a two layer artificial neural network.

35. The method of claim 30, wherein said coordinate system comprises two dimensions.

36. The method of claim 35, wherein step (c) comprises assigning said signal values to at least one cluster according to said best fit cluster model, wherein if said best fit cluster model contains at least one actual cluster and fewer than three actual clusters, then using a second artificial neural network to propose a location for at least one missing cluster, wherein the sum of actual and missing clusters is three.

37. The method of claim 36, wherein if said best fit cluster model contains one actual cluster, then using said second artificial neural network to propose a location for two missing clusters.

38. The method of claim 37, further comprising separately training said artificial neural network for proposing locations for said two missing clusters.

39. The method of claim 36, wherein if said best fit cluster model contains two actual clusters, then using said second artificial neural network to propose a location for one missing clusters.

40. The method of claim 30, wherein said genetic data comprises n sets of first, second and third signal values related in a coordinate system, wherein said first, second and third signal values are indicative of the levels of a first, second and third allele, respectively, at n loci.

41. The method of claim 40, wherein said coordinate system comprises three dimensions.

42. The method of claim 41, wherein step (c) comprises assigning said signal values to at least one cluster according to said best fit cluster model, wherein if said best fit cluster model contains at least one actual cluster and fewer than six actual clusters, then using a second artificial neural network to propose a location for at least one missing cluster, wherein the sum of actual and missing clusters is six.

43. The method of claim 30, wherein said genetic data is represented in a graphical format.

44. The method of claim 43, wherein said graphical format comprises Cartesian coordinates.

45. The method of claim 30, wherein said genetic data is provided in a tabular format.

46. The method of claim 30, wherein n is at least 2.

47-48. -48. (canceled)