Methods for removing artifact from biological profiles

• US 6,950,752 B1
• Filed: 12/23/1998
• Issued: 09/27/2005
• Est. Priority Date: 10/27/1998
• Status: Expired due to Fees

• Current Assignee: Rosetta Inpharmatics LLC (Merck & Co., Inc.), Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Rosetta Inpharmatics LLC (Merck & Co., Inc.)
• Inventors: Stoughton, Roland, He, Yudong, Friend, Stephen H.
• Primary Examiner(s): Marschel, Ardin H.
• Application Number:
  US09/220,275
• Time in Patent Office: 2,470 Days
• Field of Search: 702/19, 702/21, 435/6, 435/69.1, 703/11, 436/2
• US Class Current: 702/19
• CPC Class Codes: C12Q 1/6809 Methods for determination or identification of nucleic acids involving differential detection
• CPC Class Codes: C12Q 1/6834 Enzymatic or biochemical coupling of nucleic acids to a solid phase
• CPC Class Codes: G01N 33/6803 General methods of protein analysis not limited to specific proteins or families of proteins
• 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
• 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 40/20 Supervised data analysis
• CPC Class Codes: G16B 40/30 Unsupervised data analysis

Abstract

The invention provides methods for removing unwanted response components (i.e., “artifacts”) from a measured biological profile comprising measurements of a plurality of cellular constituents of a cell or organism in response to a perturbation. The methods involve subtracting from the measured biological profile one or more artifact patterns, each of which comprises measurements of changes in cellular constituents as a result of deviation of one or more experimental variables, such as cell culture density and temperature, hybridization temperature, as well as concentrations of total RNA and/or hybridization reagents, from desired values.

Claims

1. A method for removing one or more artifacts from a measured biological response profile, said measured biological response profile comprising measurements of a plurality of cellular constituents of a living cell or organism in response to a perturbation to said living cell or organism, said method comprising subtracting one or more artifact patterns from the measured biological response profile, wherein each said artifact pattern comprises measurements of said cellular constituents and represents contribution of one of said one or more artifacts in said measured biological response profile, wherein each of said one or more artifacts results from deviation of one or more experimental variables from conditions for measuring said measured biological response profile free of said artifact, and wherein each said artifact pattern is scaled by an independent scaling coefficient, thereby removing said one or more artifacts from said measured biological response profile.

2. The method of claim 1, wherein each of the one or more artifact patterns comprises measurements of amplitudes of said cellular constituents associated with the artifact to which the artifact pattern corresponds.

3. The method of claim 1, wherein each of the one or more artifact patterns is determined by experiments with perturbations of said one or more experimental variables to which the artifact pattern corresponds.

4. The method of claim 1, wherein each of the one or more artifact patterns is determined by a cluster analysis of control biological response profiles, the control biological response profiles comprising measurements of a plurality of cellular constituents in experiments wherein said one or more experimental variables are perturbed to produce said artifact pattern.

5. The method of claim 1, wherein the scaling coefficients are determined by a method comprising determining the values of said scaling coefficients such that the value of a function of the difference between the measured biological response profile and the sum of the one or more scaled artifact patterns is minimized.

6. The method of claim 5, wherein the function is minimized by a least squares minimization.

7. The method of claim 5 or 6, wherein the difference term associated with each cellular constituent in said function is weighed by a weighing factor, wherein said weighing factor is selected according to the significance of the measured value of said cellular constituent.

8. A method for removing one or more artifacts from a measured biological response profile, said measured biological response profile comprising measurements of a plurality of cellular constituents of a living cell or organism in response to a perturbation to said living cell or organism, each of said one or more artifacts (i) resulting from deviation of one or more experimental variables from conditions for measuring said measured biological response profile free of said artifact, and (ii) comprising measurements of said cellular constituents, said method comprising subtracting an artifact template from the measured biological response profile, wherein said artifact template comprises an artifact signature having greatest similarity to said biological response profile and is obtained by a method comprising comparing said measured biological response profile to a library of artifact signatures, each of said artifact signatures comprising measurements of amplitudes of said cellular constituents corresponding to different levels of said one or more experimental variables, said comparing comprising pattern matching of said measured biological response profile against said library, thereby removing said one or more artifacts from said measured biological response profile.

10. The method of claim 8 or 9, wherein said comparing comprises minimizing a function of the difference between said measured biological response profile and said library of artifact signatures.

11. The method of claim 10, wherein said function is minimized by a least squares minimization.

12. The method of claim 10, wherein the difference term associated with each cellular constituent in said function is weighted by a weighing factor, wherein said weighing factor is selected according to the significance of the measured value of said cellular constituent.

13. The method of claim 11, wherein the difference term associated with each cellular constituent in said function is weighted by a weighing factor, wherein said weighing factor is selected according to the significance of the measured value of said cellular constituent.

9. A method for removing one or more artifacts from a measured biological response profile, said measured biological response profile comprising measurements of a plurality of cellular constituents of a living cell or organism in response to a perturbation to said living cell or organism, each of said one or more artifacts (i) resulting from deviation of one or more experimental variables from conditions for measuring said measured biological response profile free of said artifact, (ii) and comprising measurements of said cellular constituents, said method comprising:

(a) comparing said measured biological response profile to a library of artifact signatures to generate an artifact template, each of said artifact signatures comprising measurements of amplitudes of said cellular constituents corresponding to different levels of said one or more experimental variables, said artifact template comprising an artifact signature having greatest similarity to said biological response profile, said comparing comprising pattern matching of said measured biological response profile against said library; and

(b) subtracting said artifact template from said measured biological response profile, thereby removing said one or more artifacts from said measured biological response profile.