Taxonomy-Based Glyph Design—with a Case Study on Visualizing Workflows of Biological Experiments.

Glyph-based visualization can offer elegant and concise presentation of multivariate information while enhancing speed and ease in visual search experienced by users. As with icon designs, glyphs are usually created based on the designers' experience and intuition, often in a spontaneous manner. Such a process does not scale well with the requirements of applications where a large number of concepts are to be encoded using glyphs. To alleviate such limitations, we propose a new systematic process for glyph design by exploring the parallel between the hierarchy of concept categorization and the ordering of discriminative capacity of visual channels. We examine the feasibility of this approach in an application where there is a pressing need for an efficient and effective means to visualize workflows of biological experiments. By processing thousands of workflow records in a public archive of biological experiments, we demonstrate that a cost-effective glyph design can be obtained by following a process of formulating a taxonomy with the aid of computation, identifying visual channels hierarchically, and defining application-specific abstraction and metaphors.

ArrayExpress--a public repository for microarray gene expression data at the EBI.

ArrayExpress is a public repository for microarray data that supports the MIAME (Minimum Information About a Microarray Experiment) requirements and stores well-annotated raw and normalized data. As of November 2004, ArrayExpress contains data from approximately 12,000 hybridizations covering 35 species. Data can be submitted online or directly from local databases or LIMS in a standard format, and password-protected access to prepublication data is provided for reviewers and authors. The data can be retrieved by accession number or queried by various parameters such as species, author and array platform. A facility to query experiments by gene and sample properties is provided for a growing subset of curated data that is loaded in to the ArrayExpress data warehouse. Data can be visualized and analysed using Expression Profiler, the integrated data analysis tool. ArrayExpress is available at http://www.ebi.ac.uk/arrayexpress.

[The HIV infection in pregnant women in French Guiana]. / L'infection par le VIH chez la femme enceinte en Guyane française.

OBJECTIVE: The aim of this study is to analyse the characteristics of HIV pregnant women in French Guiana then to evaluate the HIV mother to child transmission rate (MTCT) and determine the pronostic factors associated with MTCT. PATIENTS AND METHOD: An epidemiological study has been led including all deliveries in French Guiana from January 1998 to December 2000. For each case a standardized questionnaire has been gathered including epidemiological, clinical and biological data and an univariate analysis has been realized. A hundred and forty-eight women have been included in the study among 135 women came for delivery. RESULTS: The factors associated with increased MTCT in our study were no antiretroviral therapy before delivery, the lack of follow-up during pregnancy and no antiretroviral therapy in children. The HIV mother to child transmission rate was 6,5% despite the availability of antiretroviral therapies. DISCUSSION AND CONCLUSION: This rate may be explained by the difficulties of follow-up in HIV infected women. Much more needs to be done to improve access to care for women coming from foreign countries. This may be indispensable to reduce the HIV mother to child transmission rate in French Guiana.

High-resolution integrated map encompassing the breast cancer loss of heterozygosity region on human chromosome 16q22.1.

Loss of heterozygosity (LOH) on the long arm of human chromosome 16 is a common genetic alteration observed in both invasive ductal and invasive lobular breast carcinomas. We have generated a high-resolution integrated map encompassing the smallest region of LOH overlap within chromosome 16q22.1 (SRO2). Southern hybridization experiments using more than 140 probes resulted in the assembly of 152 bacterial large-insert clones into a 2.8-Mb contig covering SRO2. The structure of the contig was verified by long-range mapping using total human genomic DNA, and the contig orientation was determined by fluorescence in situ hybridization. A total of 68 transcripts have been identified in the map. One of the genes residing within SRO2 is the E-cadherin gene, CDH1, which has previously been shown to be mutated in lobular breast carcinomas, resulting in loss of E-cadherin expression. In most cases of ductal carcinoma, which is the major mammary cancer type, E-cadherin is normally expressed, suggesting that other genes within 16q22.1 are involved in the development of this tumor subtype. The high-resolution map presented in this study provides a valuable resource for identification of tumor suppressor genes expected to be involved in the etiology of breast carcinomas.

Mutation spectrum and genotype-phenotype analyses in Cowden disease and Bannayan-Zonana syndrome, two hamartoma syndromes with germline PTEN mutation.

The tumour suppressor gene PTEN , which maps to 10q23.3 and encodes a 403 amino acid dual specificity phosphatase (protein tyrosine phosphatase; PTPase), was shown recently to play a broad role in human malignancy. Somatic PTEN deletions and mutations were observed in sporadic breast, brain, prostate and kidney cancer cell lines and in several primary tumours such as endometrial carcinomas, malignant melanoma and thyroid tumours. In addition, PTEN was identified as the susceptibility gene for two hamartoma syndromes: Cowden disease (CD; MIM 158350) and Bannayan-Zonana (BZS) or Ruvalcaba-Riley-Smith syndrome (MIM 153480). Constitutive DNA from 37 CD families and seven BZS families was screened for germline PTEN mutations. PTEN mutations were identified in 30 of 37 (81%) CD families, including missense and nonsense point mutations, deletions, insertions, a deletion/insertion and splice site mutations. These mutations were scattered over the entire length of PTEN , with the exception of the first, fourth and last exons. A 'hot spot' for PTEN mutation in CD was identified in exon 5 that contains the PTPase core motif, with 13 of 30 (43%) CD mutations identified in this exon. Seven of 30 (23%) were within the core motif, the majority (five of seven) of which were missense mutations, possibly pointing to the functional significance of this region. Germline PTEN mutations were identified in four of seven (57%) BZS families studied. Interestingly, none of these mutations was observed in the PTPase core motif. It is also worthy of note that a single nonsense point mutation, R233X, was observed in the germline DNA from two unrelated CD families and one BZS family. Genotype-phenotype studies were not performed on this small group of BZS families. However, genotype-phenotype analysis inthe group of CD families revealed two possible associations worthy of follow-up in independent analyses. The first was an association noted in the group of CD families with breast disease. A correlation was observed between the presence/absence of a PTEN mutation and the type of breast involvement (unaffected versus benign versus malignant). Specifically and more directly, an association was also observed between the presence of a PTEN mutation and malignant breast disease. Secondly, there appeared to be an interdependent association between mutations upstream and within the PTPase core motif, the core motif containing the majority of missense mutations, and the involvement of all major organ systems (central nervous system, thyroid, breast, skin and gastrointestinal tract). However, these observations would need to be confirmed by studying a larger number of CD families.