Gene discovery for the system assignment of gene function

#This invention relates generally to genome-wide analysis and, more specifically, to a method of determining the function of a gene.

#The Human Genome Project, by cataloging the sequences of the estimated 100,000 human genes, provides a first step in understanding humans at the molecular level. However, with the completion of the sequencing phase of the project, many questions remain unanswered, including what roles most of these genes play in cells and how the genes work together to perform functions in cells. The answers to these questions will lead to important advances and developments in both research and medicine.

#Exemplified by genome sequencing projects, discovery science enumerates all the genes or encoded products of a genome without concern for their functional characteristics and cellular roles. The Human Genome Project and other large scale sequencing projects have fueled technological advances in discovery science. Large-scale gene sequencing, gene expression analysis methods, such as DNA microarrays, and proteomics methods have facilitated the accumulation of an enormous amount of data describing the sequences and expression levels of virtually every gene in organisms such as, the bacterium Escherichia coli, the yeast Saccharomyces cerevisiae, the worm Caenorhabditis elegans, as well as more complex organisms such as humans. Volumes of sequence and expression data can be obtained from virtually any cell or organism. However, standing alone, these volumes of sequence and expression data are difficult to interpret and apply to accurately predicting cellular functions of genes and their products, their interplay within a cell, or their dynamics in response to change.

#FIG. 1 shows a schematic diagram of a hypothetical network-associated reference data element region. The circles represent multidimensional coordinate points representative of the values of data elements (in arbitrary units) of three components of a biochemical system. Shown is a network-associated reference data element region of a reference biochemical system in three-dimensional space as a region of coordinate points and the coordinate point of a perturbed biochemical system that lies outside the network-associated reference data element region.

#FIG. 2 shows the coordinate points in two-dimensional space representative of the data element levels (in arbitrary units) of two components. The data set shows three states of a biochemical system that can be classified in three regions, corresponding to three reference network-associated data element regions.

#FIGS. 3A and B shows a data set for three networks under three states of a biochemical system and two component data element levels determined by logistic regression analysis, with FIG. 3B showing the coordinates of networks in biochemical systems “A” (x) and “B” (+).

#This invention is directed to methods for assigning a cellular function to a component of a biochemical system. The methods involve using multidimensional shape spaces to describe components of a biochemical system as they function in a particular state of the biochemical system. Multidimensional shape spaces contain multidimensional coordinate points, each of which represents a characteristic of one or more individual components of the biochemical system. For example, a multidimensional coordinate point can represent the expression of a gene, comparative expression of a gene with other genes, or expression of two or more genes. The ability to represent more than one gene using a multidimensional coordinate point is useful when analyzing global measurements of biochemical system components, such as when global expression analysis includes from some to substantially all expressed genes in a cell or tissue.

#Information about the cellular functions of genes in a biochemical system is obtained by perturbing a component of the biochemical system and determining the difference between the multidimensional shape spaces of the reference and perturbed systems. Because genes and their protein products function together in a biochemical system, such as a cell, perturbation of a gene will have some cellular consequence, such as altered expression or activity of another component of the system. Perturbation of a particular gene, for example, can result in a change in the function of the biochemical pathway of the gene, or can result in a global change in the function of one or more pathways or networks. For example, perturbation of a gene encoding an enzyme of a metabolic pathway can result in both altered function of down-stream enzymes in the same pathway, as well as up-regulation of genes in other pathways and networks that compensate for the loss of function of the perturbed metabolic pathway. Multidimensional shape space can provide a global view of the effect of perturbing a particular gene because it can contain multidimensional coordinate points corresponding to, for example, members of the perturbed pathway, compensatory pathways, and other pathways and networks possibly altered by the perturbation. Multidimensional shape space also can provide a focused view of the effect of perturbing a particular gene, depending on the contents of the multidimensional shape space, because multidimensional shape space can be used to describe systems, subsystems, networks, pathways and single components. For example, the method can be used to determine the function of components acting together as a system or subsystem, as well as used to dissect the function of such multi-component systems down to small groups, such as networks and pathways, as well as for determining the function of a single component within the analyzed system.

#One advantage of using multidimensional shape space is that the comparison allows detection and analysis of changes within the system that lack an overt phenotype or are otherwise difficult to reliably measure. For example, the effects of gene knock-out that lacks a detectable phenotype in an animal can be determined using multidimensional shape space.