Method and system for disease detection using marker combinations

• US 20040126767A1
• Filed: 12/27/2002
• Published: 07/01/2004
• Est. Priority Date: 12/27/2002
• Status: Abandoned Application

• Current Assignee: BIOSITE INCORPORATED (Abbott Laboratories Incorporated)
• Original Assignee: BIOSITE INCORPORATED (Abbott Laboratories Incorporated)
• Inventors: Buechler, Kenneth F., Dahlen, Jeffrey R., McPherson, Paul H., Kirchick, Howard J., Anderberg, Joseph Michael
• Application Number:
  US10/331,127
• Publication Number:
  US 20040126767A1
• Time in Patent Office: Days
• US Class Current: 435/6
• 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 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

Abstract

The present invention relates to methods and system for the diagnosis diseases or conditions. In a particular aspect, a disclosed method for determining a panel includes calculating a panel response for each patient in a set of diseased patients and in a set of non-diseased patients. The panel response is a function of the value of each of a plurality of markers in a panel of markers. The method also includes calculating a value for an objective function. The objective function is indicative of the effectiveness of the panel. The steps of calculating a panel response for each patient and calculating a value for an objective function are iterated by varying at least one of the parameters relating to the panel response function and a sense of each marker to facilitate optimization of the objective function. The objective function may be a measure of an overlap of panel responses of diseased patients and panel responses of non-diseased patients. The contribution of each marker to the objective function may be determined, and the panel size may be reduced by removing the poorest markers. Thus, an optimum panel of markers and an optimal panel response function for the diagnosis of a disease or condition may be determined.

Claims

1. A method of identifying a panel of markers for diagnosis of a disease or a condition, comprising:

a) calculating a panel response for each patient in a set of diseased patients and in a set of non-diseased patients, said panel response being a function of value of each of a plurality of markers in a panel of markers;

b) calculating a value for an objective function, said objective function being indicative of an effectiveness of the panel; and

c) iterating steps a) and b) by varying at least one of parameters relating to said panel response function and a sense of each marker to facilitate optimization of said objective function.

2. The method according to claim 1, wherein said objective function is a measure of an overlap of panel responses of diseased patients and panel responses of non-diseased patients.

3. The method according to claim 1, wherein said panel response is a function of value of an indicator for each of a plurality of markers in a panel of markers and a weighting coefficient for each marker, said indicator being a mapping, for each of said plurality of markers, of marker levels, said mapping being according to an indicator function;

and

wherein said iterating includes varying at least one of said weighting coefficients, parameters relating to said indicator function, and a sense of each marker to facilitate optimization of said objective function.

4. The method according to claim 3, wherein each indicator has a first value for marker levels below a cutoff region and a second value for marker values above a cutoff region, said cutoff region being defined by a location and a length.

5. The method according to claim 4, wherein said parameters include said location of said cutoff region and said length of said cutoff region.

6. The method according to claim 4, wherein said length of said cutoff region is zero.

7. The method according to claim 4, wherein said length of said cutoff region is greater than zero.

8. The method according to claim 7, wherein said indicators have values between said first value and said second value for marker levels within said cutoff region.

9. The method according to claim 8, wherein said indicators have values varying linearly from said first value to said second value across said cutoff region.

10. The method according to claim 8, wherein said indicators have values varying non-linearly from said first value to said second value across said cutoff region.

11. The method according to claim 10, wherein said non-linear variation is indicative of an error function of a distribution of marker values of diseased patients and an error function of a distribution of marker values of non-diseased patients within said cutoff region.

12. The method according to claim 3, wherein said calculating a panel response includes calculating, for each patient, Σ

w_iI_i, where w is a weighting coefficient for a marker i, I is the indicator value for the marker i, and Σ

is a summation over all of said plurality of markers.

13. The method according to claim 1, wherein said calculating a value for an objective function includes generating a receiver operating characteristic (ROC) curve for said panel response, said ROC curve being indicative of a sensitivity of said panel response as a function of one minus a specificity of said panel response.

14. The method according to claim 13, wherein said objective function is associated with an area under said ROC curve.

15. The method according to claim 13, wherein said objective function is associated with a knee of said ROC curve.

16. The method according to claim 13, wherein said objective function is associated with a sensitivity at a selected specificity level.

17. The method according to claim 13, wherein said objective function is associated with a specificity at a selected sensitivity level.

18. The method according to claim 13, wherein said objective function is associated with two or more of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level.

19. The method according to claim 13, wherein said iterating constrains at least one of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level to be above about 0.9.

20. The method according to claim 1, further comprising:

d) removing at least one of said markers from said panel;

e) calculating a value of said objective function; and

f) determining a contribution of said at least one of said markers to said objective function based on a result of step e).

21. The method according to claim 20, further comprising:

g) repeating steps d) through f) by removing a different at least one of said markers from said panel; and

h) eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function.

22. The method according to claim 1, further comprising:

d) removing at least one of said markers from said panel;

e) iterating steps a) and b) by varying parameters relating to said panel response function to facilitate optimization of said objective function; and

f) determining a contribution of said at least one of said markers to said objective function based on a result of step c).

23. The method according to claim 22, further comprising:

g) repeating steps d) through f) by removing a different at least one of said markers from said panel; and

h) eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function.

115. The method according to claim 3, wherein said indicator function is monotonic with marker value.

116. The method according to claim 115, wherein said indicator function is one of the group consisting of:

a ramp function, a step function, and a sigmoid function.

117. The method according to claim 3, wherein said indicator function is adapted to localize a marker value.

118. The method according to claim 117, wherein said indicator function is one of the group consisting of:

a triangle, a square, and Gaussian.

119. The method according to claim 1, wherein at least one of said plurality of markers is a derived marker.

120. The method according to claim 119, wherein said derived marker is the ratio of two other markers.

121. The method according to claim 1, wherein said iterating includes using a downhill simplex method.

122. The method according to claim 121, wherein said iterating further includes simulated annealing.

123. The method according to claim 122, wherein said simulated annealing includes performing a statistically sufficient number of optimizations to evaluate a most common solution.

124. The method according to claim 1, wherein said optimization is adapted to provide a stable solution.

125. The method according to claim 124, wherein said adaptation includes varying the marker values by a random percentage.

126. The method according to claim 124, wherein said adaptation includes varying one or more parameters of said panel function.

127. The method according to claim 124, wherein said adaptation includes generating a seed simplex about a minimum.

128. The method according to claim 124, wherein said adaptation includes increasing an annealing temperature until an achieved solution is not recovered.

140. The method according to claim 20, wherein said removing includes setting a weighting coefficient of said at least one of said markers to approximately zero.

161. The method according to claim 119, wherein said derived marker is indicative of the change in another marker over time.

162. The method according to claim 119, wherein said derived marker is indicative of the change in said panel response over time.

165. The method according to claim 1, wherein said optimization is adapted to simultaneously at least one of optimize and constrain a plurality of objective functions calculated from a plurality of groups of data.

24. A system for identifying a panel of markers for diagnosis of a disease or a condition, comprising:

means for calculating a panel response for each patient in a set of diseased patients and in a set of non-diseased patients, said panel response being a function of value of each of a plurality of markers in a panel of markers;

means for calculating a value for an objective function, said objective function being indicative of an effectiveness of said panel; and

means for iteratively activating said means for calculating a panel response and said means for calculating a value for an objective function, by varying at least one of parameters relating to said panel response function and a sense of each marker to facilitate optimization of said objective function.

25. The system according to claim 24, wherein said objective function is a measure of an overlap of panel responses of diseased patients and panel responses of non-diseased patients.

26. The system according to claim 24, wherein said panel response is a function of value of an indicator for each of a plurality of markers in a panel of markers and a weighting coefficient for each marker, said indicator being a mapping, for each of said plurality of markers, of marker levels, said mapping being according to an indicator function;

and

wherein said means for iteratively activating is adapted to vary at least one of said weighting coefficients, parameters relating to said indicator function, and a sense of each marker to facilitate optimization of said objective function.

27. The system according to claim 26, wherein each indicator has a first value for marker levels below a cutoff region and a second value for marker values above a cutoff region, said cutoff region being defined by a location and a length.

28. The method according to claim 27, wherein said parameters include said location of said cutoff region and said length of said cutoff region.

29. The system according to claim 27, wherein said length of said cutoff region is zero.

30. The system according to claim 27, wherein said length of said cutoff region is greater than zero.

31. The system according to claim 30, wherein said indicators have values between said first value and said second value for marker levels within said cutoff region.

32. The system according to claim 31, wherein said indicators have values varying linearly from said first value to said second value across said cutoff region.

33. The system according to claim 32, wherein said indicators have values varying non-linearly from said first value to said second value across said cutoff region.

34. The system according to claim 33, wherein said non-linear variation is indicative of an error function of a distribution of marker values of diseased patients and an error function of a distribution of marker values of non-diseased patients within said cutoff region.

35. The system according to claim 26, wherein said means for calculating a panel response is adapted to calculate, for each patient, Σ

w_iI_i, where w is a weighting coefficient for a marker i, I is the indicator value for the marker 1, and Σ

is a summation over all of said plurality of markers.

36. The system according to claim 24, wherein said means for calculating a value for an objective function is adapted to generate a receiver operating characteristic (ROC) curve for said panel response, said ROC curve being indicative of a sensitivity of said panel response as a function of one minus a specificity of said panel response.

37. The system according to claim 36, wherein said objective function is associated with an area under said ROC curve.

38. The system according to claim 36, wherein said objective function is associated with a knee of said ROC curve.

39. The system according to claim 38, wherein said objective function is associated with a sensitivity at a selected specificity level.

40. The system according to claim 36, wherein said objective function is associated with a specificity at a selected sensitivity level.

41. The system according to claim 36, wherein said objective function is associated with two or more of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level.

42. The system according to claim 36, wherein said means for iteratively activating is adapted to constrain at least one of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level to be above about 0.9.

43. The system according to claim 24, further comprising:

means for determining a contribution of said at least one of said markers to said objective function, said means for determining being adapted to remove at least one of said markers from said panel and to activate said means for calculating a value for an objective function.

44. The system according to claim 43, further comprising:

means for eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function, said means for eliminating being adapted to activate said means for determining a contribution by removing a different at least one of said markers from said panel.

45. The system according to claim 24, further comprising:

means for determining a contribution of said at least one of said markers to said objective function, said means for determining being adapted to remove at least one of said markers from said panel and to iteratively activate said means for calculating a panel response and said means for calculating a value for an objective function, by varying parameters relating to said panel response function to facilitate optimization of said objective function.

46. The system according to claim 45, further comprising:

means for eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function, said means for eliminating being adapted to activate said means for determining a contribution by removing a different at least one of said markers from said panel.

141. The system according to claim 26, wherein said indicator function is monotonic with marker value.

142. The system according to claim 141, wherein said indicator function is one of the group consisting of:

a ramp function, a step function, and a sigmoid function.

143. The system according to claim 26, wherein said indicator function is adapted to localize a marker value.

144. The system according to claim 143, wherein said indicator function is one of the group consisting of:

a triangle, a square, and Gaussian.

145. The system according to claim 43, wherein said removing includes setting a weighting coefficient of said at least one of said markers to approximately zero.

166. The system according to claim 24, wherein said means for iteratively activating is adapted to simultaneously at least one of optimize and constrain a plurality of objective functions calculated from a plurality of groups of data.

47. A program product, comprising machine readable program code for causing a machine to perform following method steps:

a) calculating a panel response for each patient in a set of diseased patients and in a set of non-diseased patients, said panel response being a function of value of each of a plurality of markers in a panel of markers;

b) calculating a value for an objective function, said objective function being indicative of an effectiveness of said panel; and

c) iterating steps a) and b) by varying at least one of parameters relating to said panel response function and a sense of each marker to facilitate optimization of said objective function.

48. The program product according to claim 47, wherein said objective function is a measure of an overlap of panel responses of diseased patients and panel responses of non-diseased patients.

49. The program product according to claim 47, wherein said panel response is a function of value of an indicator for each of a plurality of markers in a panel of markers and a weighting coefficient for each marker, said indicator being a mapping, for each of said plurality of markers, of marker levels, said mapping being according to an indicator function;

and

wherein said iterating includes varying at least one of said weighting coefficients, parameters relating to said indicator function, and a sense of each marker to facilitate optimization of said objective function.

50. The program product according to claim 49, wherein each indicator has a first value for marker levels below a cutoff region and a second value for marker values above a cutoff region, said cutoff region being defined by a location and a length.

51. The program product according to claim 50, wherein said parameters include said location of said cutoff region and said length of said cutoff region.

52. The program product according to claim 50, wherein said length of said cutoff region is zero.

53. The program product according to claim 50, wherein said length of said cutoff region is greater than zero.

54. The program product according to claim 53, wherein said indicators have values between said first value and said second value for marker levels within said cutoff region.

55. The program product according to claim 54, wherein said indicators have values varying linearly from said first value to said second value across said cutoff region.

56. The program product according to claim 54, wherein said indicators have values varying non-linearly from said first value to said second value across said cutoff region.

57. The program product according to claim 56, wherein said non-linear variation is indicative of an error function of a distribution of marker values of diseased patients and an error function of a distribution of marker values of non-diseased patients within said cutoff region.

58. The program product according to claim 49, wherein said calculating a panel response includes calculating, for each patient, Σ

w_iI_i, where w is a weighting coefficient for a marker i, I is the indicator value for the marker i, and Σ

is a summation over all of said plurality of markers.

59. The program product according to claim 47, wherein said calculating a value for an objective function includes generating a receiver operating characteristic (ROC) curve for said panel response, said ROC curve being indicative of a sensitivity of said panel response as a function of one minus a specificity of said panel response.

60. The program product according to claim 59, wherein said objective function is associated with an area under said ROC curve.

61. The program product according to claim 59, wherein said objective function is associated with a knee of said ROC curve.

62. The program product according to claim 59, wherein said objective function is associated with a sensitivity at a selected specificity level.

63. The program product according to claim 59, wherein said objective function is associated with a specificity at a selected sensitivity level.

64. The program product according to claim 59, wherein said objective function is associated with two or more of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level.

65. The program product according to claim 59, wherein said iterating constrains at least one of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level to be above about 0.9.

66. The program product according to claim 47, further comprising machine readable program code for causing a machine to perform following method steps:

d) removing at least one of said markers from said panel;

e) calculating a value of said objective function; and

f) determining a contribution of said at least one of said markers to said objective function based on a result of step e).

67. The program product according to claim 66, further comprising machine readable program code for causing a machine to perform following method steps:

g) repeating steps d) through f) by removing a different at least one of said markers from said panel; and

h) eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function.

68. The program product according to claim 47, further comprising machine readable program code for causing a machine to perform following method steps:

d) removing at least one of said markers from said panel;

e) iterating steps a) and b) by varying parameters relating to said panel response function to facilitate optimization of said objective function; and

f) determining a contribution of said at least one of said markers to said objective function based on a result of step e).

69. The program product according to claim 68, further comprising machine readable program code for causing a machine to perform following method steps:

g) repeating steps d) through f) by removing a different at least one of said markers from said panel; and

h) eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function.

70. The program product according to claim 47, wherein said machine readable code is embedded in a portable meter.

71. The program product according to claim 70, wherein said portable meter is a fluorometer.

72. The program product according to claim 70, wherein said portable meter is a reflectometer.

73. The program product according to claim 47, wherein said machine readable code is embedded in a computer.

74. The program product according to claim 73, wherein said computer is a portable computer.

75. The program product according to claim 73, wherein said computer is adapted to be accessed through a network.

76. The program product according to claim 75, wherein said network is the Internet.

77. The program product according to claim 73, wherein said computer is adapted to be coupled to an analyzer.

78. The program product according to claim 77, wherein said analyzer is an immunoassay analyzer.

79. The program product according to claim 77, wherein said analyzer is a single nucleotide polymorphism detector.

80. The program product according to claim 77, wherein said analyzer is adapted to sort and count similar and different particles and cells.

146. The program product according to claim 49, wherein said indicator function is monotonic with marker value.

147. The program product according to claim 49, wherein said indicator function is one of the group consisting of:

a ramp function, a step function, and a sigmoid function.

148. The program product according to claim 49, wherein said indicator function is adapted to localize a marker value.

149. The program product according to claim 49, wherein said indicator function is one of the group consisting of:

a triangle, a square, and Gaussian.

164. The program product according to claim 66, wherein said removing includes setting a weighting coefficient of said at least one of said markers to approximately zero.

167. The program product according to claim 47, wherein said optimization is adapted to simultaneously at least one of optimize and constrain a plurality of objective functions calculated from a plurality of groups of data.

81. A method of identifying a panel of markers for diagnosis of a disease or a condition, comprising:

a) selecting a panel of markers, said panel including a plurality of markers measured in a set of diseased patients and a set of non-diseased patients;

b) defining a cutoff region of marker levels for each of said plurality of markers, said cutoff region having a location and a length;

c) selecting a weighting coefficient for each of said plurality of markers;

d) mapping, for each of said plurality of markers, marker levels to an indicator, each of said indicators having a first value for marker levels below said cutoff region and a second value for marker levels above said cutoff region;

e) calculating a panel response for each patient in said set of diseased patients and in said set of non-diseased patients, said panel response being a function of value of said indicator for each marker and said weighting coefficient for each marker;

f) calculating a value for an objective function, said objective function being indicative of an effectiveness of said panel; and

g) iterating steps e) and f) by varying at least one of said location of said cutoff region, said length of said cutoff region, said weighting coefficients, and a sense of each marker to facilitate optimization of said objective function.

82. The method according to claim 81, wherein said objective function is a measure of an overlap of panel responses of diseased patients and panel responses of non-diseased patients.

83. The method according to claim 81, wherein said length of said cutoff region is zero.

84. The method according to claim 81, wherein said length of said cutoff region is greater than zero.

85. The method according to claim 84, wherein said indicators have values between said first value and said second value for marker levels within said cutoff region.

86. The method according to claim 85, wherein said indicators have values varying linearly from said first value to said second value across said cutoff region.

87. The method according to claim 85, wherein said indicators have values varying non-linearly from said first value to said second value across said cutoff region.

88. The method according to claim 87, wherein said non-linear variation is indicative of an error function of a distribution of marker values of diseased patients and an error function of a distribution of marker values of non-diseased patients within said cutoff region.

89. The method according to claim 81, wherein said calculating a panel response includes calculating, for each patient, Σ

w_iI_i, where w is a weighting coefficient for a marker i, I is the indicator value for the marker I, and Σ

is a summation over all of said plurality of markers.

90. The method according to claim 81, wherein said calculating a value for an objective function includes generating a receiver operating characteristic (ROC) curve for said panel response, said ROC curve being indicative of a sensitivity of said panel response as a function of one minus a specificity of said panel response.

91. The method according to claim 90, wherein said objective function is associated with an area under said ROC curve.

92. The method according to claim 90, wherein said objective function is associated with a knee of said ROC curve.

93. The method according to claim 90, wherein said objective function is associated with a sensitivity at a selected specificity level.

94. The method according to claim 90, wherein said objective function is associated with a specificity at a selected sensitivity level.

95. The method according to claim 90, wherein said objective function is associated with two or more of an area under said ROC curve, a knee of said ROC curve, a sensitivity at a selected specificity level, and a specificity at a selected sensitivity level.

96. The method according to claim 81, further comprising:

h) setting said weighting coefficient of at least one of said markers to approximately zero;

i) calculating a value of said objective function with remaining weighting coefficients; and

j) determining a contribution of said at least one of said markers to said objective function.

97. The method according to claim 96, further comprising:

k) repeating steps h) through j) by setting said weighting coefficient of a different at least one of said markers to approximately zero; and

l) eliminating a marker from said panel of markers in accordance with said contribution of said marker to said objective function.

98. A method of diagnosing a subject for a disease or condition, comprising:

a) measuring a value of each of a plurality of markers in said subject;

b) calculating a panel response for said subject, said panel response being a function of values of a plurality of markers; and

c) diagnosing said disease or condition in said patient when said panel response is greater than a predetermined cutoff.

129. The method according to claim 98, wherein said calculating a panel response includes using a panel response function and parameters relating to said panel response function, said panel response function and parameters being determined by:

a) calculating a panel response for each patient in a set of diseased patients and in a set of non-diseased patients, said panel response being a function of value of each of a plurality of markers in a panel of markers;

b) calculating a value for an objective function, said objective function being indicative of an effectiveness of the panel; and

c) iterating steps a) and b) by varying at least one of parameters relating to said panel response function and a sense of each marker to facilitate optimization of said objective function.

130. The method according to claim 129, wherein said panel of markers is adapted to diagnose two or more diseases or conditions.

131. The method according to claim 130, wherein different panel response functions and parameters are used for each of said diseases or conditions.

132. The method according to claim 98, wherein said at least one of said plurality of markers is a derived marker.

133. The method according to claim 132, wherein said derived marker is a ratio of two other markers.

134. The method according to claim 132, wherein said derived marker is indicative of a change over time in a marker.

135. The method according to claim 98, wherein said calculating a panel response includes calculating, for each patient, Σ

w_iI_i, where w is a weighting coefficient for a marker i, I is an indicator value for the marker i, and Σ

is a summation over all of said plurality of markers, said indicator value being a mapping, for each of said plurality of markers, of marker levels, said mapping being according to an indicator function.

136. The method according to claim 135, wherein said indicator function is monotonic with marker value.

137. The method according to claim 136, wherein said indicator function is one of the group consisting of:

a ramp function, a step function, and a sigmoid function.

138. The method according to claim 135, wherein said indicator function is adapted to localize a marker value.

139. The method according to claim 138, wherein said indicator function is one of the group consisting of:

a triangle, a square, and Gaussian.

163. The method according to claim 132, wherein said derived marker is indicative of a change over time in said panel response.

168. The method according to claim 134, wherein said calculating a panel response is adapted to calculate a panel response without said derived marker when said derived marker is not available and to calculate a panel response using said derived marker when said derived marker is available.

169. The method according to claim 168, wherein said panel response without said derived marker and said panel response using said derived marker are calculated according to different panel response functions and parameters relating to said panel response functions.

170. The method according to claim 163, wherein said calculating a panel response is adapted to calculate a panel response without said derived marker when said derived marker is not available and to calculate a panel response using said derived marker when said derived marker is available.

171. The method according to claim 170, wherein said panel response without said derived marker and said panel response using said derived marker are calculated according to different panel response functions and parameters relating to said panel response functions.

99. A program product, comprising machine readable program code for causing a machine to perform following method steps:

a) calculating a value of each of a plurality of markers in a subject;

b) calculating a panel response for said subject, said panel response being a function of values of a plurality of markers; and

c) diagnosing a disease or condition in said patient when said panel response is greater than a predetermined cutoff.

100. The program product according to claim 99, wherein said machine readable program code is embedded in a portable meter.

101. The program product according to claim 100, wherein said portable meter is a fluorometer.

102. The program product according to claim 100, wherein said portable meter is a reflectometer.

103. The program product according to claim 99, wherein said machine readable program code is embedded in a computer.

104. The program product according to claim 103, wherein said computer is a portable computer.

105. The program product according to claim 103, wherein said computer is adapted to be accessed through a network.

106. The program product according to claim 105, wherein said network is the Internet.

107. The program product according to claim 103, wherein said computer is adapted to be coupled to an analyzer.

108. The program product according to claim 107, wherein said analyzer is an immunoassay analyzer.

109. The program product according to claim 107, wherein said analyzer is a single nucleotide polymorphism detector.

110. The program product according to claim 107, wherein said analyzer is adapted to sort and count similar and different particles and cells.

150. The program product according to claim 99, wherein said calculating a panel response includes using a panel response function and parameters relating to said panel response function, said panel response function and parameters being determined by:

a) calculating a panel response for each patient in a set of diseased patients and in a set of non-diseased patients, said panel response being a function of value of each of a plurality of markers in a panel of markers;

b) calculating a value for an objective function, said objective function being indicative of an effectiveness of the panel; and

c) iterating steps a) and b) by varying at least one of parameters relating to said panel response function and a sense of each marker to facilitate optimization of said objective function.

151. The program product according to claim 150, wherein said panel of markers is adapted to diagnose two or more diseases or conditions.

152. The program product according to claim 151, wherein different panel response functions and parameters are used for each of said diseases or conditions.

153. The program product according to claim 99, wherein said at least one of said plurality of markers is a derived marker.

154. The program product according to claim 153, wherein said derived marker is a ratio of two other markers.

155. The program product according to claim 153, wherein said derived marker is indicative of a change over time in a marker.

156. The program product according to claim 99, wherein said calculating a panel response includes calculating, for each patient, Σ

w_iI_i, where w is a weighting coefficient for a marker i, I is an indicator value for the marker i, and Σ

is a summation over all of said plurality of markers, said indicator value being a mapping, for each of said plurality of markers, of marker levels, said mapping being according to an indicator function.

157. The program product according to claim 156, wherein said indicator function is monotonic with marker value.

158. The program product according to claim 157, wherein said indicator function is one of the group consisting of:

a ramp function, a step function, and a sigmoid function.

159. The program product according to claim 156, wherein said indicator function is adapted to localize a marker value.

160. The program product according to claim 159, wherein said indicator function is one of the group consisting of:

a triangle, a square, and Gaussian.

111. A method of identifying a panel of markers for diagnosis of a disease or a condition, comprising:

a) identifying a cutoff region for each of a plurality of markers, said cutoff region being substantially centered about an overlap region of marker values for a set of diseased patients and a set of non-diseased patients, said cutoff region having a location and a length;

b) determining an effectiveness value of each of said plurality of markers in distinguishing said set of diseased patients from said set of non-diseased patients; and

c) defining a panel response as a function of said effectiveness value of each marker and a measured level of each marker.

112. The method according to claim 111, wherein said cutoff region has a length of zero.

113. The method according to claim 111, wherein said cutoff region has a non-zero length.

114. The method according to claim 111, wherein said effectiveness value of each marker is represented by an area under a ROC curve.