Mathematical analysis for the estimation of changes in the level of gene expression

• US 6,502,039 B1
• Filed: 05/24/2000
• Issued: 12/31/2002
• Est. Priority Date: 05/24/2000
• Status: Expired due to Fees

• Current Assignee: Aventisub II, Inc. (Sanofi )
• Original Assignee: Aventis Pharmaceuticals Incorporated (Sanofi-Aventis U.S., LLC)
• Inventors: Fuchs, Rainer, Theilhaber, Joachim, Bushnell, Steven
• Primary Examiner(s): Marschel, Ardin H.
• Assistant Examiner(s): SHEINBERG, MONIKA B
• Application Number:
  US09/577,634
• Time in Patent Office: 951 Days
• Field of Search: 435/6, 702/19
• US Class Current: 702/19
• 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

The present invention relates to a highly accurate and reproducible mathematically-based methodology for quantifying the levels of differential gene expression from microassay protocols. Specifically, the present invention provides a simple deductive approach, grounded in a Bayesian framework, to circumvent the heuristic-based limitation of previous methodologies. Rather than seeking a point-estimate of the fold-change of the level of gene expression, the present invention utilizes the derivation of mathematical formula to determine the a posteriori distribution of all the fold-changes of differential gene expression which may be inferred from the given measurements. From this a posteriori distribution the following information may be obtained: (i) an estimator for the fold-change of the level of gene expression; (ii) confidence limits for the fold-change, at any given confidence level; and (iii) a P-value for assessing the statistical significance of change. The present invention also possesses the advantage that fold-change estimates and confidence limits may even be assigned to signal pairs where both signals are zero or negative, without resorting to heuristic thresholds.

Claims

1. A method for calculating indicia of differences in gene-expression level in a plurality of array hybridizations, the method comprising:

(a) determining an indication of experimental noise associated with intensity of hybridization signal for a gene on each array hybridization;

(b) using the determined experimental noise indication to determine a first analytical probability distribution function describing distribution values of intensity for the gene on each array hybridization, assuming the noise is Gaussian and by applying Bayes Theorem;

(c) using the first analytic probability distribution functions to derive a second analytical probability distribution function describing an indication of a gene-expression difference of a differentially-expressed gene; and

(d) applying the second probability distribution function of the differentially-expressed gene using experimentally-derived intensities and noise values from the array hybridizations to determine values related to the gene-expression difference.

2. The method of claim 1 wherein the determined values include estimated fold changes in the concentration of gene transcripts.

3. The method of claim 2 wherein the determined values include a quality metric associated with at least one of the estimated fold changes.

4. The method of claim 3 wherein the quality metric represents at least one of a probability that a fold change could be less than 1 when an estimated fold change is greater than 1, and a probability that a fold change is greater than 1 when the estimated fold change is less than 1.

5. The method of claim 1 wherein the determined values include confidence limits on the fold change given specific confidence intervals.

11. The method of claim 1 wherein an indication of experimental noise is determined, the first probability distribution function is determined, and the second probability distribution function is derived and applied for each gene on each array hybridization.

6. A computer program product, for calculating indicia of differences gene-expression level in a plurality of array hybridizations, residing on a computer-readable medium and comprising instructions for causing a computer to:

(a) determine an indication of experimental noise associated with intensity of hybridization signal for a gene on each array hybridization;

(b) use the determined experimental noise indication to determine a first analytical probability distribution function describing distribution values of intensity for the gene on each array hybridization, assuming the noise is Gaussian and by applying Bayes Theorem;

(c) use the first analytic probability distribution functions to derive a second analytical probability distribution function describing an indication of a gene-expression difference of a differentially-expressed gene;

(d) apply the second probability distribution function of the differentially-expressed gene using experimentally-derived intensities and noise values from the array hybridizations to determine values related to the gene-expression difference.

7. The computer program product of claim 6 wherein the determined values include estimated fold changes in the concentration of gene transcripts.

8. The computer program product of claim 7 wherein the determined values include a quality metric associated with at least one of the estimated fold changes.

9. The computer program product of claim 8 wherein the quality metric represents at least one of a probability that a fold change could be less than 1 when an estimated fold change is greater than 1, and a probability that a fold change is greater than 1 when the estimated fold change is less than 1.

10. The computer program product of claim 6 wherein the determined values include confidence limits on the fold change given specific confidence intervals.

12. The computer program product of claim 6 wherein the instructions for causing a computer to determine an indication of experimental noise, determine the first probability distribution function, derive the second probability distribution function, and apply the second probability distribution function, cause the computer to do so for each gene on each array hybridization.