The UCSC genome browser database: update 2007.

The University of California, Santa Cruz Genome Browser Database contains, as of September 2006, sequence and annotation data for the genomes of 13 vertebrate and 19 invertebrate species. The Genome Browser displays a wide variety of annotations at all scales from the single nucleotide level up to a full chromosome and includes assembly data, genes and gene predictions, mRNA and EST alignments, and comparative genomics, regulation, expression and variation data. The database is optimized for fast interactive performance with web tools that provide powerful visualization and querying capabilities for mining the data. In the past year, 22 new assemblies and several new sets of human variation annotation have been released. New features include VisiGene, a fully integrated in situ hybridization image browser; phyloGif, for drawing evolutionary tree diagrams; a redesigned Custom Track feature; an expanded SNP annotation track; and many new display options. The Genome Browser, other tools, downloadable data files and links to documentation and other information can be found at http://genome.ucsc.edu/.

The UCSC Genome Browser Database: update 2006.

The University of California Santa Cruz Genome Browser Database (GBD) contains sequence and annotation data for the genomes of about a dozen vertebrate species and several major model organisms. Genome annotations typically include assembly data, sequence composition, genes and gene predictions, mRNA and expressed sequence tag evidence, comparative genomics, regulation, expression and variation data. The database is optimized to support fast interactive performance with web tools that provide powerful visualization and querying capabilities for mining the data. The Genome Browser displays a wide variety of annotations at all scales from single nucleotide level up to a full chromosome. The Table Browser provides direct access to the database tables and sequence data, enabling complex queries on genome-wide datasets. The Proteome Browser graphically displays protein properties. The Gene Sorter allows filtering and comparison of genes by several metrics including expression data and several gene properties. BLAT and In Silico PCR search for sequences in entire genomes in seconds. These tools are highly integrated and provide many hyperlinks to other databases and websites. The GBD, browsing tools, downloadable data files and links to documentation and other information can be found at http://genome.ucsc.edu/.

Transcriptome and genome conservation of alternative splicing events in humans and mice.

Combining mRNA and EST data in splicing graphs with whole genome alignments, we discover alternative splicing events that are conserved in both human and mouse transcriptomes. 1,964 of 19,156 (10%) loci examined contain one or more such alternative splicing events, with 2,698 total events. These events represent a lower bound on the amount of alternative splicing in the human genome. Also, as these alternative splicing events are conserved between the human and mouse transcriptomes they should be enriched for functionally significant alternative splicing events, free from much of the noise found in the EST libraries. Further classification of these alternative splicing events reveals that 1,037 (38.4%) are due to exon skipping, 497 (18.4%) are due to alternative 3' splice sites, 214 (7.9%) are due to alternative 5' splice sites, 75 (2.8%) are due to intron retention and the other 875 (32.4%) are due to other, more complicated, alternative splicing events. In addition, genomic sequences nearby these alternative splicing events display increased sequence conservation. Both the alternatively spliced exons and the proximal intron show increased levels of genomic conservation relative to constitutively spliced exons. For exon skipping events both intron regions flanking the exon are conserved while for alternative 5' and 3' splicing events the conservation is greater near the alternative splice site.

The UCSC Genome Browser Database.

The University of California Santa Cruz (UCSC) Genome Browser Database is an up to date source for genome sequence data integrated with a large collection of related annotations. The database is optimized to support fast interactive performance with the web-based UCSC Genome Browser, a tool built on top of the database for rapid visualization and querying of the data at many levels. The annotations for a given genome are displayed in the browser as a series of tracks aligned with the genomic sequence. Sequence data and annotations may also be viewed in a text-based tabular format or downloaded as tab-delimited flat files. The Genome Browser Database, browsing tools and downloadable data files can all be found on the UCSC Genome Bioinformatics website (http://genome.ucsc.edu), which also contains links to documentation and related technical information.

The Gli2 transcription factor is required for normal mouse mammary gland development.

The hedgehog signal transduction network performs critical roles in mediating cell-cell interactions during embryogenesis and organogenesis. Loss-of-function or misexpression mutation of hedgehog network components can cause birth defects, skin cancer, and other tumors. The Gli gene family (Gli1, Gli2, and Gli3) encodes zinc finger transcription factors that act as mediators of hedgehog signal transduction. In this study, we investigate the role of Gli2 in mammary gland development. Mammary expression of Gli2 is developmentally regulated in a tissue compartment-specific manner. Expression is exclusively stromal during virgin stages of development but becomes both epithelial and stromal during pregnancy and lactation. The null phenotype with respect to both ductal and alveolar development was examined by transplantation rescue of embryonic mammary glands into physiologically normal host females. Glands derived from both wild type and null embryo donors showed ductal outgrowths that developed to equivalent extents in virgin hosts. However, in null transplants, ducts were frequently distended or irregularly shaped and showed a range of histological alterations similar to micropapillary ductal hyperplasias in the human breast. Alveolar development during pregnancy was not overtly affected by loss of Gli2 function. Ductal defects were not observed when homozygous null epithelium was transplanted into a wild type stromal background, indicating that Gli2 function is required primarily in the stroma for proper ductal development. DeltaGli2 heterozygotes also demonstrated an elevated frequency and severity of focal ductal dysplasia relative to that of wild type littermate- and age-matched control animals.

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.

Defects in mouse mammary gland development caused by conditional haploinsufficiency of Patched-1.

In vertebrates, the hedgehog family of cell signaling proteins and associated downstream network components play an essential role in mediating tissue interactions during development and organogenesis. Loss-of-function or misexpression mutation of hedgehog network components can cause birth defects, skin cancer and other tumors. The mammary gland is a specialized skin derivative requiring epithelial-epithelial and epithelial-stromal tissue interactions similar to those required for development of other organs, where these interactions are often controlled by hedgehog signaling. We have investigated the role of the Patched-1 (Ptc1) hedgehog receptor gene in mammary development and neoplasia. Haploinsufficiency at the Ptc1 locus results in severe histological defects in ductal structure, and minor morphological changes in terminal end buds in heterozygous postpubescent virgin animals. Defects are mainly ductal hyperplasias and dysplasias characterized by multilayered ductal walls and dissociated cells impacting ductal lumens. This phenotype is 100% penetrant. Remarkably, defects are reverted during late pregnancy and lactation but return upon involution and gland remodeling. Whole mammary gland transplants into athymic mice demonstrates that the observed dysplasias reflect an intrisic developmental defect within the gland. However, Ptc1-induced epithelial dysplasias are not stable upon transplantation into a wild-type epithelium-free fat pad, suggesting stromal (or epithelial and stromal) function of Ptc1. Mammary expression of Ptc1 mRNA is both epithelial and stromal and is developmentally regulated. Phenotypic reversion correlates with developmentally regulated and enhanced expression of Indian hedgehog (Ihh) during pregnancy and lactation. Data demonstrate a critical mammary role for at least one component of the hedgehog signaling network and suggest that Ihh is the primary hedgehog gene active in the gland.