Knowledge-based analysis of microarray gene expression data by using support vector machines

We introduce a method of functionally classifying genes by using gene expression data from DNA microarray hybridization experiments. The method is based on the theory of support vector machines (SVMs). SVMs are considered a supervised computer learning method because they exploit prior knowledge of gene function to identify unknown genes of similar function from expression data. SVMs avoid several problems associated with unsupervised clustering methods, such as hierarchical clustering and self-organizing maps. SVMs have many mathematical features that make them attractive for gene expression analysis, including their flexibility in choosing a similarity function, sparseness of solution when dealing with large data sets, the ability to handle large feature spaces, and the ability to identify outliers. We test several SVMs that use different similarity metrics, as well as some other supervised learning methods, and find that the SVMs best identify sets of genes with a common function using expression data. Finally, we use SVMs to predict functional roles for uncharacterized yeast ORFs based on their expression data.

Antagonism of WT1 activity by protein self-association.

Germline loss-of-function mutations at the Wilms tumor (WT) suppressor locus WT1 are associated with a predisposition to WTs and mild genital system anomalies. In contrast, germ-line missense mutations within the WT1 gene encoding the DNA-binding domain often yield a more severe phenotype consisting of WT, sexual ambiguity, and renal nephropathy. In this report, we demonstrate that the products of mutant alleles that impair DNA recognition can antagonize WT1-mediated transcriptional repression. We demonstrate that WT1 can self-associate in vitro and in vivo and that the responsible domain maps to the amino-terminal region of the protein. Oligomers of full-length protein form less efficiently or produce less stable complexes than oligomers between truncated polypeptides and full-length protein. Our data suggest a molecular mechanism to explain how WT1 mutations may act in deregulating cellular proliferation and differentiation.

Aod1, the immunoregulatory locus controlling abrogation of tolerance in neonatal thymectomy-induced autoimmune ovarian dysgenesis, maps to mouse chromosome 16.

Mice thymectomized at three days of age (D3Tx) develop during adulthood a variety of organ-specific autoimmune diseases, including autoimmune ovarian dysgenesis (AOD). The phenotypic spectrum of AOD is characterized by the development of anti-ovarian autoantibodies, oophoritis, and atrophy. The D3Tx model of AOD is unique in that disease induction depends exclusively on perturbation of the normal developing immune system, is T-cell-mediated, and is strain specific. For example, D3Tx A/J mice are highly susceptible to AOD, whereas C57BL/6J mice are resistant. After D3Tx, self ovarian antigens, expressed at physiological levels, trigger an autoimmune response capable of eliciting disease. The D3Tx model provides, therefore, the opportunity to focus on the mechanisms of self-tolerance that are relevant to disease pathogenesis. Previous studies indicate that the principal mechanisms involved in AOD susceptibility are genetically controlled and govern developmental processes associated with the induction and maintenance of peripheral tolerance. We report here the mapping of the Aod1 locus to mouse chromosome 16 within a region encoding several loci of immunologic relevance, including scid, Igl1, VpreB, Igll, Igl1r, Mtv6 (Mls-3), Ly-7, Ifnar, and Ifgt.