Methods for classifying samples and ascertaining previously unknown classes

• US 7,239,986 B2
• Filed: 02/12/2002
• Issued: 07/03/2007
• Est. Priority Date: 04/09/1999
• Status: Expired due to Fees

• Current Assignee: Dana-Farber Cancer Institute, Whitehead Institute for Biomedical Research
• Original Assignee: Dana-Farber Cancer Institute, Whitehead Institute for Biomedical Research
• Inventors: Golub, Todd R., Lander, Eric S., Tamayo, Pablo, Slonim, Donna, Mesirov, Jill
• Primary Examiner(s): Moran; Marjorie A.
• Application Number:
  US10/074,789
• Publication Number:
  US 20030017481A1
• Time in Patent Office: 1,967 Days
• Field of Search: 706/20, 702/19, 514/908, 536/23.5, 703/2
• US Class Current: 703/2
• CPC Class Codes: G16B 25/00 ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression
• CPC Class Codes: G16B 25/10 Gene or protein expression profiling; Expression-ratio estimation or normalisation

Abstract

Methods and apparatus for classifying or predicting the classes for samples based on gene expression are described. Also described are methods and apparatus for ascertaining or discovering new, previously unknown classes based on gene expression.

Claims

1. A method of assigning a sample to a known class, comprising the steps of:

a. determining a weighted vote for one or more informative genes for the known class in said sample in accordance with a model built with a weighted voting scheme, wherein the magnitude of each vote depends on the expression level of the one or more informative genes in said sample and on the degree of correlation of the one or more informative genes'"'"' expression with class distinction;

b. summing the votes to determine a winning class and a prediction strength;

c. assigning said sample to the winning class, which winning class is a known class, if the prediction strength is greater than a prediction strength threshold; and

d. providing an output indicating assignment of said sample to the winning class, wherein the known class is a cancer disease class selected from any of Acute Lymphoblastic Leukemia (ALL), Acute Myeloid Leukemia (AML), glioblastoma, medulloblastoma, folicular lymphoma, or diffuse large B cell lymphoma.

2. The method of claim 1, wherein the cancer disease class is Acute Lymphoblastic Leukemia (ALL) or Acute Myeloid Leukemia (AML).

3. The method of claim 1, wherein the prediction strength is determined by:

(V_win−

V_lose)/(V_win+V_lose),

wherein V_win and V_lose are the vote totals for one or more winning and losing classes, respectively.

4. The method of claim 3, wherein the cancer disease class is Acute Lymphoblastic Leukemia (ALL) or Acute Myeloid Leukemia (AML).

5. The method of claim 3, wherein the number of informative genes used in the weighted voting scheme is at least 50.

6. The method of claim 5, wherein the cancer disease class is glioblastoma or medulloblastoma.

7. The method of claim 5, wherein the cancer disease class is folicular lymphoma or diffuse large B cell lymphoma.

8. The method of claim 5, wherein the cancer disease class is Acute Lymphoblastic Leukemia (ALL) or Acute Myeloid Leukemia (AML).

9. The method of claim 8, wherein the informative genes encode proteins selected from the group consisting of:

c-myb, MB-1, cyclinD3, Rb Ap48, SNF2, E2A, topoisomerase II β

, TFIIEβ

, MCM3, Op 18, zyxin, HoxA9, CD33, and CD11c.

10. The method of claim 8, wherein the informative genes are markers of hematopoietic lineage.

11. The method of claim 1, wherein the output comprises an output assembly.

12. The method of claim 11, wherein the output assembly comprises a graphical representation.