Methods for predicting chemosensitivity or chemoresistance

A method for determining whether a compound is effective for the treatment of cancer, comprising:

• a) assessing the sample to obtain gene expression values for one or more informative genes;
• b) determining a weighted vote for a chemosensitive class or a chemoresistant class for the informative genes in said sample in accordance with a classifier built with a weighted voting scheme from samples which have been determined to be either sensitive or resistant to a compound by growth inhibition, wherein the magnitude of each vote depends on the expression level of the gene in said sample and on the degree of correlation of the gene's expression with being either chemosensitive or chemoresistant;
• and
• c) summing the votes to determine the winning class.

A method for determining a mechanism of action in the proliferation or non-proliferation of cells in a sample which is determined to be either sensitive or resistant to a compound by growth inhibition, wherein more than one compound is assessed, comprising the steps of:

• a) sorting genes by the degree to which their expression in said sample which is determined to be either sensitive or resistant to a compound by growth inhibition correlates with a chemosensitive class or a chemoresistant class;
• b) determining whether said correlation is stronger than expected by chance;
• and
• c) repeating steps a) and b) for each compound being assessed;
• wherein a gene whose expression correlates with the chemoresistant class more strongly than expected by chance is a gene involved in the proliferation of cells;
• and wherein a gene whose expression correlates with the chemosensitive class more strongly than expected by chance is a gene involved in the non-proliferation of cells, thereby determining a mechanism of action.

A method of determining a treatment plan for an individual undergoing chemotherapy, comprising:

• a) obtaining a sample from the individual;
• b) assessing the sample for the level of gene expression for one or more genes;
• and
• c) using a model built with a weighted voting scheme from gene expression data from samples which are determined to be either chemosensitive or chemoresistant to more than one compound by growth inhibition, classifying the sample including comparing the gene expression values of the sample to that of the model, to thereby determine a treatment plan.

In a computer system, method for classifying a sample obtained from an individual in a chemosensitive class or a chemoresistant class, wherein gene expression values are determined for the sample to be tested, comprising:

• a) receiving the gene expression values for two or more genes from the sample to be tested;
• b) using a model built with a weighted voting scheme from gene expression data from samples which are determined to be either sensitive or resistant to more than one compound by growth inhibition, classifying the sample in a chemosensitive class or a chemoresistant class including comparing the gene expression values of the sample to that of the model, to thereby produce a classification of the sample;
• and
• c) providing an output indication of the classification.

In a computer system, method for classifying a sample obtained from an individual in a chemosensitive class or a chemoresistant class, wherein gene expression values are determined for the sample to be tested, comprising:

• a) providing a model built by a weighted voting scheme from gene expression data from samples which are determined to be either sensitive or resistant to more than one compound by growth inhibition;
• b) assessing the sample for the level of gene expression for more than one gene, to thereby obtain a gene expression value for each gene;
• c) using the model built with a weighted voting scheme, classifying the sample in a chemosensitive class or a chemoresistant class including comparing the gene expression values of the sample to that of the model, to thereby produce a classification of the sample;
• and
• d) providing an output indication of the classification.

In a computer system, a method for building a model for classifying at least one sample to be tested having a gene expression product, comprising:

• a) receiving a vector for gene expression values of two or more samples which are determined to be either sensitive or resistant to a compound by growth inhibition, the vector being a series of gene expression values for the samples;
• b) determining genes that are relevant for classification of said sample in a chemosensitive class or a chemosensitive class;
• and
• c) using a weighted voting routine, constructing the model for classifying the samples using at least a portion of the genes determined in step b).

A computer apparatus for classifying a sample into a chemosensitive class or a chemoresistant class, wherein the sample is obtained from an individual, wherein the apparatus comprises:

• a) a source of gene expression values of the sample;
• b) a processor routine executed by a digital processor, coupled to receive the gene expression values from the source, the processor routine determining classification of the sample by comparing the gene expression values of the sample to a model built with a weighted voting scheme from gene expression data from samples which are determined to be either sensitive or resistant to more than one compound by growth inhibition;
• and
• c) an output assembly, coupled to the digital processor, for providing an indication of the classification of the sample.

A machine readable computer assembly for classifying a sample into a chemosensitive class or a chemoresistant class, wherein the sample is obtained from an individual, wherein the apparatus comprises:

• a) a source of gene expression values of the sample;
• b) a processor routine executed by a digital processor, coupled to receive the gene expression values from the source, the processor routine determining classification of the sample by comparing the gene expression values of the sample to a model built with a weighted voting scheme from gene expression data from samples which are determined to be either sensitive or resistant to more than one compound by growth inhibition;
• and
• c) an output assembly, coupled to the digital processor, for providing an indication of the classification of the sample.

A method of identifying a set of informative genes whose expression correlates with a treatment-sensitive class or a treatment-resistant class across samples which are determined to be either sensitive or resistant to a treatment, wherein more than one compound is assessed, comprising the steps of:

• a) sorting genes by degree to which their expression in said sample which is determined to be either sensitive or resistant to a treatment correlates with the treatment-sensitive class or the treatment-resistant class;
• b) determining whether said correlation is stronger than expected by chance;
• and
• c) repeating steps a) and b) for each compound being assessed;
• wherein a gene whose expression correlates with either the treatment-sensitive class or the treatment-resistant class more strongly than expected by chance is an informative gene, thereby identifying a set of informative genes.

A method of identifying a set of informative genes whose expression correlates with a drug-sensitive class or a drug-resistant class across samples which are determined to be either sensitive or resistant to a compound, wherein more than one compound is assessed, comprising the steps of:

• a) sorting genes by degree to which their expression in said sample which is determined to be either sensitive or resistant to a compound correlates with the drug-sensitive class or the drug-resistant class;
• b) determining whether said correlation is stronger than expected by chance;
• and
• c) repeating steps a) and b) for each compound being assessed;
• wherein a gene whose expression correlates with either the drug-sensitive class or drug-resistant class more strongly than expected by chance is an informative gene, thereby identifying a set of informative genes.