System and method for multiple-factor selection

Abstract

The disclosed subject matter provides techniques for multiple-factor selection. The factors can be features or elements that are jointly associated with one or more outcomes by their joint presence or absence. There may be a non-causative correlation between the factors, features, or elements and the outcomes. In some embodiments, Entropy Minimization and Boolean Parsimony (EMBP) is used to identify modules of genes jointly associated with disease from gene expression data, and a logic function is provided to connect the combined expression levels in each gene module with the presence of disease. The smallest module of genes whose joint expression levels can predict the presence of disease can be identified.

1 . A method for predicting an effect of a drug on a subject by selecting factors from a data set of measurements, the measurements including values of the factors and outcomes, comprising:
 identifying two or more factors that are jointly associated with one or more outcomes from the data set, wherein the two or more factors include two or more genes and the one or more outcomes include the effect of the drug;   analyzing each of the two or more factors to determine at least one interaction therein with respect to an outcome;   identifying interactions, if any, from the determined at least one interaction that correlate the data from the two or more factors with the effect of the drug; and   predicting the effect of the drug on a subject by assessing the presence or absence of the interactions correlated with the effect of the drug on the subject.   
     
     
         2 . The method of  claim 1 , wherein the data includes gene expression data comprising expression levels for each of the two or more genes. 
     
     
         3 . The method of  claim 1 , wherein the data includes the presence or absence of single nucleotide polymorphisms (SNPs) in each of the two or more genes. 
     
     
         4 . The method of  claim 1 , wherein the two or more factors comprise a module of factors. 
     
     
         5 . The method of  claim 4 , wherein the at least one interaction comprises a structure of interactions. 
     
     
         6 . The method of  claim 4 , wherein the at least one interaction comprises a logic function. 
     
     
         7 . The method of  claim 6 , wherein the two or more genes comprise a module of genes. 
     
     
         8 . The method of  claim 7 , wherein the module of genes comprise a smallest module of genes with joint expression levels that can be used for the prediction of the effect of the drug on the subject with high accuracy. 
     
     
         9 . The method of  claim 8 , wherein the logic function comprises a simplest logic function connecting the genes to achieve the prediction. 
     
     
         10 . A method for predicting an effect of a drug on a subject by selecting two or more genes from gene data, comprising:
 discretizing the gene data;   identifying two or more genes from the selected two or more genes having a minimal conditional entropy;   identifying an interaction, if any, that correlates the gene data for the two or more identified genes with the effect of the drug; and   predicting the effect of the drug on the subject by assessing the presence or absence of the interaction correlated with the effect of the drug on the subject.   
     
     
         11 . The method of  claim 10 , wherein the gene data includes gene expression data comprising expression levels for each of the two or more genes. 
     
     
         12 . The method of  claim 10 , wherein the gene data includes the presence or absence of SNPs in each of the two or more genes. 
     
     
         13 . The method of  claim 10 , wherein the gene expression data is derived from at least one microarray of gene expression data. 
     
     
         14 . The method of  claim 10 , wherein the two or more genes comprise a module of genes. 
     
     
         15 . The method of  claim 10 , wherein the interaction is modeled using a most parsimonious Boolean function. 
     
     
         16 . A system for predicting an effect of a drug on a subject by selecting two or more genes from gene data, comprising:
 at least one processor, and   a computer readable medium coupled to the at least one processor, having stored thereon instructions which when executed cause the processor to:   discretize the gene data;   choose a single threshold for each of the two or more genes;   identify the two or more genes from the selected two or more genes having a minimal conditional entropy; identify an interaction, if any, that correlates the gene data for the two or more genes with the effect of the drug; and   predict the effect of the drug on the subject by assessing the presence or absence of the interaction correlated with the effect of the drug on the subject.   
     
     
         17 . The system of  claim 16 , wherein the gene data includes gene expression data comprising expression levels for each of the two or more genes. 
     
     
         18 . The system of  claim 16 , wherein the gene data includes the presence or absence of SNPs in each of the two or more genes. 
     
     
         19 . The system of  claim 16 , wherein the gene expression data is derived from a microarray of gene expression data. 
     
     
         20 . The system of  claim 16 , wherein the two or more genes comprise a module of genes. 
     
     
         21 . The system of  claim 16 , wherein the interaction comprises a most parsimonious Boolean function. 
     
     
         22 . A system for predicting an effect of a drug on a subject by selecting factors from a data set of measurements, each measurement comprising values of the factors and outcomes, comprising:
 at least one processor, and   a computer readable medium coupled to the at least one processor, having stored thereon instructions which when executed cause the at least one processor to:   identify two or more factors that are jointly associated with one or more outcomes from the data, wherein the two or more factors include two or more genes and the one or more outcomes comprise the effect of the drug;   analyze each of the two or more factors to determine at least one interaction therein with respect to an outcome;   identify interactions, if any, from the determined at least one interaction that correlate the data from the two or more factors with the effect of the drug; and   predict the effect of the drug on the subject by assessing the presence or absence of the interactions correlated with the effect of the drug on the subject.   
     
     
         23 . The system of  claim 22 , wherein the data includes gene expression data comprising expression levels for each of the two or more genes. 
     
     
         24 . The system of  claim 22 , wherein the data includes the presence or absence of SNPs in each of the two or more genes. 
     
     
         25 . The system of  claim 22 , wherein the two or more factors comprise a module of factors. 
     
     
         26 . The system of  claim 25 , wherein the at least one interaction comprises a structure of interactions. 
     
     
         27 . The system of  claim 25 , wherein the at least one interaction comprises a logic function. 
     
     
         28 . The system of  claim 22 , wherein the two or more genes comprise a module of genes. 
     
     
         29 . The system of  claim 28 , wherein the module of genes comprises a smallest module of genes with joint expression levels that can be used for the prediction of the effect of the drug with high accuracy. 
     
     
         30 . The system of  claim 29 , wherein the logic function comprises the simplest logic function connecting the genes to achieve the prediction.