The VERA information environment

We present three components of a virtual research environment developed for the ongoing Roman excavation at Silchester. These components — Recycle Bridge, XDB cross-database search, and Arch3D — provide additional services around the existing core of the system, run on the Integrated Archaeological Database (IADB). They provide, respectively, embedding of legacy applications into portals, cross-database searching, and 3D visualisation of stratigraphic information.

Identification of non-zein proteins in BR473 maize protein bodies by LC-nanoESI-MS/MS

The nutritional value of maize seed is limited due to its high content of storage proteins (zeins), which are deficient in essential amino acids such as lysine and tryptophan. In a previous paper, we showed that protein bodies obtained from BR473 maize variety, developed by Embrapa (Brazilian Agricultural Research Corporation), were mainly constituted by Z27 and a smaller quantity of Z50 gamma-zeins. Besides zein proteins, other not identified protein band in the SDS/PAGE was also observed, which could indicate the presence of non-zein proteins additionally to gamma-zeins. In the present paper, we have demonstrated the presence of non-zein proteins in BR473 maize protein bodies by LC-nanoESI-MS/MS and database searching. This fact could be related to the excellent energetic value and higher protein quality of BR473 maize grains, since high lysine concentration in some maize varieties has been related to the presence of cytoskeleton proteins that are non-zeins. We have identified the following proteins: Brittle-1 protein (chloroplast precursor), Legumin-1, glyceroldehyde-3-phosphate dehydrogenase, and elongation factor 1-alpha.

Learning from large data : bias, variance, sampling, and learning curves

One of the fundamental machine learning tasks is that of predictive classification. Given that organisations collect an ever increasing amount of data, predictive classification methods must be able to effectively and efficiently handle large amounts of data. However, it is understood that present requirements push existing algorithms to, and sometimes beyond, their limits since many classification prediction algorithms were designed when currently common data set sizes were beyond imagination. This has led to a significant amount of research into ways of making classification learning algorithms more effective and efficient. Although substantial progress has been made, a number of key questions have not been answered. This dissertation investigates two of these key questions. The first is whether different types of algorithms to those currently employed are required when using large data sets. This is answered by analysis of the way in which the bias plus variance decomposition of predictive classification error changes as training set size is increased. Experiments find that larger training sets require different types of algorithms to those currently used. Some insight into the characteristics of suitable algorithms is provided, and this may provide some direction for the development of future classification prediction algorithms which are specifically designed for use with large data sets. The second question investigated is that of the role of sampling in machine learning with large data sets. Sampling has long been used as a means of avoiding the need to scale up algorithms to suit the size of the data set by scaling down the size of the data sets to suit the algorithm. However, the costs of performing sampling have not been widely explored. Two popular sampling methods are compared with learning from all available data in terms of predictive accuracy, model complexity, and execution time. The comparison shows that sub-sampling generally products models with accuracy close to, and sometimes greater than, that obtainable from learning with all available data. This result suggests that it may be possible to develop algorithms that take advantage of the sub-sampling methodology to reduce the time required to infer a model while sacrificing little if any accuracy. Methods of improving effective and efficient learning via sampling are also investigated, and now sampling methodologies proposed. These methodologies include using a varying-proportion of instances to determine the next inference step and using a statistical calculation at each inference step to determine sufficient sample size. Experiments show that using a statistical calculation of sample size can not only substantially reduce execution time but can do so with only a small loss, and occasional gain, in accuracy. One of the common uses of sampling is in the construction of learning curves. Learning curves are often used to attempt to determine the optimal training size which will maximally reduce execution time while nut being detrimental to accuracy. An analysis of the performance of methods for detection of convergence of learning curves is performed, with the focus of the analysis on methods that calculate the gradient, of the tangent to the curve. Given that such methods can be susceptible to local accuracy plateaus, an investigation into the frequency of local plateaus is also performed. It is shown that local accuracy plateaus are a common occurrence, and that ensuring a small loss of accuracy often results in greater computational cost than learning from all available data. These results cast doubt over the applicability of gradient of tangent methods for detecting convergence, and of the viability of learning curves for reducing execution time in general.

Knowledge discovery in multi-databases by analyzing local instances

Multi-databases mining is an urgent task. This thesis solves 4 key problems in multi-databases mining: Application-independent database classification - Local instance analysis model - Useful pattern discovery - Pattern synthesis.

Algorithms and applications of preference based ranking for information retrieval

In this thesis, the author designed three sets of preference based ranking algorithms for information retrieval and provided the corresponsive applications for the algorithms. The main goal is to retrieve recommended, high similar and valuable ranking results to users.

Efficient dimensionality reduction and one-class classification for content-based image retrieval

The thesis investigates various machine learning approaches to reducing data dimensionality, and studies the impact of asymmetric data on learning in image retrieval. Efficient algorithms are proposed to reduce the data dimensionality. Integration strategies for one-class classification are designed to address asymmetric data issue and improve retrieval effectiveness.

Interventions to reduce sexual risk behavior in African American men

The purpose of this review was to describe characteristics of interventions that have been conducted for African American men and identify similarities and differences between interventions for subpopulations of African American men. Of the 473 studies identified through database searching, 8 studies met the eligibility criteria for this review. Target populations within these studies included heterosexual men, homosexual and bisexual men, persons with mental health problems, and injection and non-injection drug users. Little variation was found in the theories and methods used in interventions for different target populations. However, several different behavioral determinants were addressed within these interventions. Although interventions for drug users generally did not exclude individuals based on sexual orientation, these interventions were able to address HIV sexual risk behavior in heterosexual, homosexual, and bisexual-identified men. Given these findings, multi-faceted approaches to HIV prevention are recommended in order to reduce HIV risk in African American men.^

Cytomegalovirus in autoimmunity: T cell crossreactivity to viral antigen and autoantigen glutamic acid decarboxylase

Antigens of pathogenic microbes that mimic autoantigens are thought to be responsible for the activation of autoreactive T cells. Viral infections have been associated with the development of the neuroendocrine autoimmune diseases type 1 diabetes and stiff-man syndrome, but the mechanism is unknown. These diseases share glutamic acid decarboxylase (GAD65) as a major autoantigen. We screened synthetic peptide libraries dedicated to bind to HLA-DR3, which predisposes to both diseases, using clonal CD4+ T cells reactive to GAD65 isolated from a prediabetic stiff-man syndrome patient. Here we show that these GAD65-specific T cells crossreact with a peptide of the human cytomegalovirus (hCMV) major DNA-binding protein. This peptide was identified after database searching with a recognition pattern that had been deduced from the library studies. Furthermore, we showed that hCMV-derived epitope can be naturally processed by dendritic cells and recognized by GAD65 reactive T cells. Thus, hCMV may be involved in the loss of T cell tolerance to autoantigen GAD65 by a mechanism of molecular mimicry leading to autoimmunity.

The rise and fall of the medical mediated searcher

The relationship between the development of mediated online literature searching and the recruitment of medical librarians to fill positions as online searchers was investigated. The history of database searching by medical librarians was outlined and a content analysis of thirty-five years of job advertisements in MLA News from 1961 through 1996 was summarized. Advertisements for online searchers were examined to test the hypothesis that the growth of mediated online searching was reflected in the recruitment of librarians to fill positions as mediated online searchers in medical libraries. The advent of end-user searching was also traced to determine how this trend affected the demand for mediated online searching and job availability of online searchers. Job advertisements were analyzed to determine what skills were in demand as end-user searching replaced mediated online searching as the norm in medical libraries. Finally, the trend away from mediated online searching to support of other library services was placed in the context of new roles for medical librarians.

The Research Funding Service: a model for expanded library services

Traditionally, libraries have provided a modest amount of information about grants and funding opportunities to researchers in need of research funding. Ten years ago, the University of Washington (UW) Health Sciences Libraries and Information Center joined in a cooperative effort with the School of Medicine to develop a complete, library-based grant and funding service for health sciences researchers called the Research Funding Service. The library provided space, access to the library collection, equipment, and electronic resources, and the School of Medicine funded staff and operations. The range of services now includes individual consultation appointments, an extensive Web site, classes on funding database searching and writing grant applications, a discussion series that frequently hosts guest speakers, a monthly newsletter with funding opportunities of interest to the six health sciences schools, extensive files on funding sources, and referral services.

Teaching principles of genetics through SNP analysis

An understanding of inheritance requires comprehension of genetic processes at all levels, from molecules to populations. Frequently genetics courses are separated into molecular and organismal genetics and students may fail to see the relationships between them. This is particularly true with human genetics, because of the difficulties in designing experimental approaches which are consistent with ethical restrictions, student abilities and background knowledge, and available time and materials. During 2005 we used analysis of single nucleotide polymorphisms (SNPs) in two genetic regions to enhance student learning and provide a practical experience in human genetics. Students scanned databases to discover SNPs in a gene of interest, used software to design PCR primers and a restriction enzyme based assay for the alleles, and carried out an analysis of the SNP on anonymous individual and family DNAs. The project occupied eight to ten hours per week for one semester, with some time spent in the laboratory and some spent in database searching, reading and writing the report. In completing their projects, students acquired a knowledge of Mendel’s first law (through looking at inheritance patterns), Mendel’s second law and the exceptions (the concepts of linkage and linkage disequilibrium), DNA structure (primer design and restriction enzyme analysis) and function (SNPs in coding and non-coding regions), population genetics and the statistical analysis of allele frequencies, genomics, bioinformatics and the ethical issues associated with the use of human samples. They also developed skills in presentation of results by publication and conference participation. Deficiencies in their understanding (for example of inheritance patterns, gene structure, statistical approaches and report writing) were detected and guidance given during the project. SNP analysis was found to be a powerful approach to enhance and integrate student understanding of genetic concepts.

Targeted mass spectrometry methods for detecting oxidative post-translational modifications

Oxidative post-translational modifications (oxPTMs) can alter the function of proteins, and are important in the redox regulation of cell behaviour. The most informative technique to detect and locate oxPTMs within proteins is mass spectrometry (MS). However, proteomic MS data are usually searched against theoretical databases using statistical search engines, and the occurrence of unspecified or multiple modifications, or other unexpected features, can lead to failure to detect the modifications and erroneous identifications of oxPTMs. We have developed a new approach for mining data from accurate mass instruments that allows multiple modifications to be examined. Accurate mass extracted ion chromatograms (XIC) for specific reporter ions from peptides containing oxPTMs were generated from standard LC-MSMS data acquired on a rapid-scanning high-resolution mass spectrometer (ABSciex 5600 Triple TOF). The method was tested using proteins from human plasma or isolated LDL. A variety of modifications including chlorotyrosine, nitrotyrosine, kynurenine, oxidation of lysine, and oxidized phospholipid adducts were detected. For example, the use of a reporter ion at 184.074 Da/e, corresponding to phosphocholine, was used to identify for the first time intact oxidized phosphatidylcholine adducts on LDL. In all cases the modifications were confirmed by manual sequencing. ApoB-100 containing oxidized lipid adducts was detected even in healthy human samples, as well as LDL from patients with chronic kidney disease. The accurate mass XIC method gave a lower false positive rate than normal database searching using statistical search engines, and identified more oxidatively modified peptides. A major advantage was that additional modifications could be searched after data collection, and multiple modifications on a single peptide identified. The oxPTMs present on albumin and ApoB-100 have potential as indicators of oxidative damage in ageing or inflammatory diseases.

Responses of the Emiliania huxleyi Proteome to Ocean Acidification

Ocean acidification due to rising atmospheric CO2 is expected to affect the physiology of important calcifying marine organisms, but the nature and magnitude of change is yet to be established. In coccolithophores, different species and strains display varying calcification responses to ocean acidification, but the underlying biochemical properties remain unknown. We employed an approach combining tandem mass-spectrometry with isobaric tagging (iTRAQ) and multiple database searching to identify proteins that were differentially expressed in cells of the marine coccolithophore species Emiliania huxleyi (strain NZEH) between two CO2 conditions: 395 (~current day) and ~1340 p.p.m.v. CO2. Cells exposed to the higher CO2 condition contained more cellular particulate inorganic carbon (CaCO3) and particulate organic nitrogen and carbon than those maintained in present-day conditions. These results are linked with the observation that cells grew slower under elevated CO2, indicating cell cycle disruption. Under high CO2 conditions, coccospheres were larger and cells possessed bigger coccoliths that did not show any signs of malformation compared to those from cells grown under present-day CO2 levels. No differences in calcification rate, particulate organic carbon production or cellular organic carbon: nitrogen ratios were observed. Results were not related to nutrient limitation or acclimation status of cells. At least 46 homologous protein groups from a variety of functional processes were quantified in these experiments, of which four (histones H2A, H3, H4 and a chloroplastic 30S ribosomal protein S7) showed down-regulation in all replicates exposed to high CO2, perhaps reflecting the decrease in growth rate. We present evidence of cellular stress responses but proteins associated with many key metabolic processes remained unaltered. Our results therefore suggest that this E. huxleyi strain possesses some acclimation mechanisms to tolerate future CO2 scenarios, although the observed decline in growth rate may be an overriding factor affecting the success of this ecotype in future oceans.

Interactions between drugs and drug-nutrient in enteral nutrition: a review based on evidences

Introduction: Enteral nutrition (EN) provides calories, macronutrients and micronutrients in adequate quantity and quality to meet the patient's needs. Some drugs when crushed and diluted may have their properties altered, including the reduction of bioavailability causing the reduction of the serum concentration of the drug; tube obstruction; drug-drug interaction or drug-nutrient interaction. Methods: The study was conducted through review of submitted articles in the databases of the Virtual Health Library (VHL): MEDLINE (National Library of Medicine, USA), Lilacs (Latin American and Caribbean Literature on Health Sciences) PUBMED - NCBI (National Center for Biotechnology Information) and COCHRANE. Results: For this survey, 42 articles were identified during database searching. After applying the inclusion and exclusion criteria, 08 articles were selected, obtained from the MEDLINE and Lilacs. Discussion: Some interactions were found such as the aluminium hydroxide and lactulose with the enteral nutrition, which may result in a precipitation and reduction of drug bioavailability. Mineral oil will alter the absorption of fat-soluble vitamins and reduces the tube light. Others results were found as phenytoin, warfarin, captopril and furosemide with enteral nutrition may reduce the maximum serum concentration. Conclusion: Drug interactions are more common in day-to-day activities than health professionals may suppose. Knowledge on the matter may also assist in reducing cases of obstruction of tubes, through which enteral nutrition and medications are administered. Thus, the multidisciplinary team, acting together, may have more beneficial effects to the patient.