Methods and computer software products for multiple probe gene expression analysis
AssigneeAffymetrix, Inc.
InventorMing-Hsiu Ho
Life of this patent
Estimating the term…
Abstract
Methods and computer software products are provided for analyzing gene expression data. In one embodiment, the expression of a gene is determined by multiple probes in several experiments. A principal component analysis is performed to obtain the relative expression of the gene in these experiments.
Claims32
A computer implemented method for determining the relative level of a biological molecule in a plurality of experiments comprising:
- a) providing a plurality of signals where each of said signals relects said level of said biological molecule in one of said experiments;
- b) determining said relative level of said biological molecule by a principal component.
A method for selecting nucleic acid probes from a pool of candidate nucleic acid probes for a target nucleic acid comprising:
- a) measuring hybridization intensities between each of said candidate probes with said target nucleic acid in a plurality of experiments;
- and
- b) selecting said nucleic acid probes based upon the inner product Of normalized eigenvector associated with the largest eigenvalue and normalized experimental hybridization intensity for each of said candidate probes.
A computer software product comprising:
- a) Computer program code that inputs a plurality of signals where each of said signals reflects the level of a biological molecule in one of a plurality of experiments;
- b) Computer program code that determines said relative level of said biological molecule by calculating a principal component;
- and
- c) A computer readable media storing said computer codes.
A method for determining a canonical vector for analyzing multiple probe nucleic acid hybridization comprising:
- a) providing a matrix S, wherein:
- 14 S = [ S 11 · S 1 j · S 1 n · · · · · · · · · · ;
- · · · · S m1 · S mi · S mn ]
- wherein Sij is the hybridization intensity of a jth probe in ith experiment;
- and
- b) determining said canonical vector by calculating the eigenvector of a matrix P;
- wherein said eigenvector is associated with the largest eigenvalue and said matrix P={tilde over (S)}·S.
A computer implemented method for determining the level of a nucleic acid comprising:
- providing a plurality of hybridization intensities (Si . . . Sj . . . Sn);
- wherein Sj reflects the hybridization between jth probe and said nucleic acid and n is the total number of probes and n is greater than 2;
- and
- calculating said level as
- 15 C · [ S 1 · S j · S n ] = [ c 1 · c j · c n ] · [ S 1 · S j · S n ] ;
- ;
- wherein said C is a canonical vector determined using principal component analysis.
Dependencies are inferred from the claim text.