Background: Multiple sclerosis (MS) is the most common cause of disability in young people and yet there are no interventions available which reliably alter disease progression. This is despite several decades of research using the most common animal model of multiple sclerosis, experimental autoimmune encephalomyelitis (EAE). There is now emerging evidence across the neurosciences to suggest that limited internal validity (measures to reduce bias) and external validity (e.g. using a clinically relevant animal model) may influence the translational success. Aim and objectives: To provide an unbiased summary of the scope of the literature on candidate drugs for MS tested in EAE to identify potential reasons for the failures to translate efficacy to clinical trials. My objectives were, across all of the identified publications, to: (1) describe the reporting of measures to reduce bias and to assess their impact on measures of drug efficacy; (2) assess the relationship between treatment related effects measured using different outcome measures; (3) assess the prevalence and impact of any publication bias; (4) compare findings from the above with another disease with limited translational success (Parkinson’s disease; PD). Methods: I used systematic searches of three online databases to identify relevant publications. Estimates of efficacy were extracted for neurobehavioural scores, inflammation, demyelination and axon loss. For PD experiments, we searched for dopamine agonists tested in animal models of PD with outcome assessed as change in neurobehavioural scores. I calculated normalised mean difference or standardised mean difference effect sizes and combined these in a meta-analysis using a random effects model. I used stratified meta-analysis or meta-regression to assess the extent to which different study design characteristics explained differences in reported efficacies. These characteristics included: measures to reduce bias (random allocation to group and blinded assessment of outcome), the animal species, sex, time of drug administration, route of drug administration and the number of animals per group. Publication bias was assessed using funnel plotting, Egger regression and “trim and fill”. Results: I identified 1464 publications reporting drugs tested in EAE. Reported study quality was poor: 11% reported random allocation to group, 17% reported blinded assessment of neurobehavioural outcomes, 28% reported blinded assessment of histological outcomes, and < 1% reported a sample size calculation. Estimates of efficacy measured as the reduction in inflammation were substantially higher in unblinded studies (47.1% reduction (95% CI 41.8-52.4)) versus blinded studies (33.1% (25.8-40.4). Moreover, the same finding was identified for 121 publications on dopamine agonists tested in experimental PD models where efficacy was measured as change in neurobehavioural outcomes. For EAE studies we were unable to include data from 631 publications describing original research. Usually this was because the publication did not include basic details such as the number of animals in each group (115 publications), the observed variance (592) or suitable control data (49). For each category of outcome I found evidence of a substantial publication bias. Interventions were most commonly administered on or before the induction of EAE with shorter times to treatment associated with higher estimates of efficacy for the reduction in mean severity scores (a neurobehavioural outcome). Treatment related effects were found to vary across different outcome measures with the largest effect being for the reduction in axon loss. Where neurobehavioural scores and axon loss were measured in the same cohort of animals, the concordance between efficacies in these increased with later times to treatment. Conclusions: In this, the largest systematic review and meta-analysis of animal studies in any domain, I have found that a large number of publications present incomplete data. In addition, measures to reduce bias are seldom reported, the lack of which is associated with overstatements of efficacy for both a measure of drug efficacy in EAE and experimental PD studies. Translational success may have also been affected by the majority of studies administering drugs on or before EAE induction which is of limited relevance in the clinical setting where patients do not present at that stage of disease. Moreover, my analysis of the relationship between outcome measures provides empirical evidence from systematically identified studies to suggest that targeting axon loss as later time points is most strongly associated with improvements in neurobehavioural scores. Therefore drugs which are successfully able to target axon loss at these time points may offer substantial hope for clinical success. Overall, improvements in the conduct and reporting of preclinical studies are likely to improve their utility, and the prospects for translational success. While my findings relate predominately to the animal modelling of MS and PD it is likely that they also hold for other animal research.

Genome-wide microarray technology has facilitated the systematic discovery of diagnostic biomarkers of cancers and other pathologies. However, meta-analyses of published arrays using melanoma as a test cancer has uncovered significant inconsistences that hinder advances in clinical practice. In this study a computational model for the integrated analysis of microarray datasets is proposed in order to provide a robust ranking of genes in terms of their relative significance; both genome-wide relative significance (GWRS) and genome-wide global significance (GWGS). When applied to five melanoma microarray datasets published between 2000 and 2011, a new 12-gene diagnostic biomarker signature for melanoma was defined (i.e., EGFR, FGFR2, FGFR3, IL8, PTPRF, TNC, CXCL13, COL11A1, CHP2, SHC4, PPP2R2C, and WNT4). Of these, CXCL13, COL11A1, PTPRF and SHC4 are components of the MAPK pathway and were validated by immunocyto- and immunohisto-chemistry. These proteins were found to be overexpressed in metastatic and primary melanoma cells in vitro and in melanoma tissue in situ compared to melanocytes cultured from healthy skin epidermis and normal healthy human skin. One challenge for the integrated analysis of microarray data is that the microarray data are produced using different platforms and bio-samples, e.g. including both cell line- and biopsy-based microarray datasets. In order to address these challenges, the computational model was further enhanced the stratification of datasets into either biopsy or cell line derived datasets, and via the weighting of microarray data based on quality criteria of data. The methods enhancement was applied to 14 microarray datasets of three cancers (breast, prostate, and melanoma) based on classification accuracy and on the capability to identify predictive biomarkers. Four novel measures for evaluating the capability to identify predictive biomarkers are proposed: (1) classifying independent testing data using wrapper feature selection with machine leaning, (2) assessing the number of common genes with the genes retrieved in independent testing data, (3) assessing the number of common genes with the genes retrieved in across multiple training datasets, (4) assessing the number of common genes with the genes validated in the literature. This enhancement of computational approach (i) achieved reliable classification performance across multiple datasets, (ii) recognized more significant genes into the top-ranked genes as compared to the genes detected by the independent test data, and (iii) detected more meaningful genes than were validated in previous melanoma studies in the literature.

This thesis provides the first meta-analysis investigating the effect of corporate tax competition among states, with special focus on the effect of the corporate tax rate change in competing country on the corporate tax rate in the home country. It examines 523 estimates from 20 published studies and working papers. Results of the meta-analysis show an evidence of substantial publication selectivity: researchers tend to discard negative and insignificant estimates, which overvalues the estimated effect size. Conducted precision effect test failed to find the evidence for the existence of a genuine effect of corporate tax competition. Empirical analysis shows that differences in the measurement of statutory and effective tax rate matter, thus the analysis was conducted on two separate sub-samples. Meta-regression analysis have found significant impact of variables related to publication bias for both sub-samples. Next to it, the results provide an evidence of significant influence of politically orientated controls, especially of the variable controlling whether or not there were elections in the particular year and state in case when the corporate tax rate changed. Powered by TCPDF (www.tcpdf.org)

Master of Science / Department of Agronomy / Allan Fritz / Fusarium head blight (FHB) is one of the most devastative diseases in wheat. Growing resistant cultivars is one of the most effective strategies to minimize the disease damage. Huangcandou (HCD) is a Chinese wheat landrace showing a high level of resistance to FHB spread within a spike (type II). To identify quantitative traits loci (QTL) for resistance in HCD, a population of 190 recombinant inbred lines (RILs) were developed from a cross between HCD and Jagger, a susceptible hard winter wheat (HWW) released in Kansas. The population was evaluated for type II resistance at the greenhouses of Kansas State University. After initial marker screening, 261 polymorphic simple-sequence repeats (SSR) between parents were used for analysis of the RIL population. Among three QTL identified, two from HCD were mapped on the short arms of chromosomes 3B (3BS) and 3A (3AS). The QTL on the distal end of 3BS showed a major effect on type II resistance in all three experiments. This QTL coincides with a previously reported Fhb1, and explained 28.3% of phenotypic variation. The QTL on 3AS explained 9.7% of phenotypic variation for mean PSS over three experiments. The third QTL from chromosome 2D of Jagger explained 6.5% of phenotypic variation. Allelic substitution using the closest marker to each QTL revealed that substitution of Jagger alleles of two QTL on 3AS and 3BS with those from HCD significantly reduced the PSS. HCD containing both QTL on 3AS and 3BS with a large effect on type II resistance can be an alternative source of FHB resistance for improving FHB type II resistance in wheat. Besides, meta-analyses were used to estimate 95% confidence intervals (CIs) of 24 mapped QTL in five previously mapped populations derived from Chinese landraces: Wangshuibai (WSB), Haiyanzhong (HYZ), Huangfangzhu (HFZ), Baishanyuehuang (BSYH) and Huangcandou (HCD). Nineteen QTL for FHB type II resistance were projected to 10 QTL clusters. Five QTL on chromosomes 1A, 5A, 7A, and 3BS (2) were identified as confirmed QTL that have stable and consistent effects on FHB resistance and markers in these meta-QTL regions should be useful for marker-assisted breeding.

Thesis advisor: Lisa Barrett / Researchers have wondered how the brain creates emotions since the early days of psychological science. With the advent of neuroimaging techniques in the early 1990's and a surge of studies in affective neuroscience in recent years, scientists are now poised to answer this question. In this paper, I present the most up-to-date and statistically advanced meta-analytic summary of the human neuroimaging literature on emotion. I compare the locationist approach (i.e., that emotion categories consistently and specifically correspond to distinct brain regions) with the psychological construction approach (i.e., that emotions are constructed of more general brain networks not specific to emotions) to better understand the brain basis of emotion. I begin by outlining the set of brain regions consistently activated across all studies of emotion experience and perception. I next report findings from two sets of analyses probing the brain basis of discrete emotion categories. The first types of analysis demonstrates the brain regions that are consistently associated with the experience and perception of anger, disgust, fear, happiness and sadness across studies. The second type of analysis demonstrates the mental states (e.g., emotion experience or perception, cognitive load, locus of attention, mental response to methods, etc.) that are consistently associated with activity in given brain locations across studies. Overall, there was little evidence that discrete emotion categories can be localized consistently and specifically to individual brain regions. Instead, I found evidence that is consistent with a psychological construction approach to the mind: a set of common processes corresponding to interacting brain networks constitute emotion experience and perception across a range of emotion categories. / Thesis (PhD) — Boston College, 2010. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Psychology.

Schistosomiasis is a water-borne parasitic disease of great public health importance mainly in sub-Saharan African countries. The majority of current control programmes use the antihelminthic drug praziquantel to reduce disease burden in endemic areas. Praziquantel treatment has been reported to accelerate the development of protective immunity against re-infection that otherwise takes years to develop. To date, there is no licensed vaccine for schistosomiasis in humans but an attenuated schistosome parasite vaccine has been tested in animal models. Employing systematic review and meta-analysis approaches, my PhD research has four main objectives relating to attenuated schistosome vaccine and praziquantel treatment: 1) to identify predictors that determine protection levels after treatment with attenuated Schistosoma mansoni vaccines in the mouse model, 2) to quantify the influence of host and schistosome parasite species on attenuated parasite vaccine efficacy, 3) to explore the direction of change (increase/decrease) in schistosome parasite-specific antibody isotypes after praziquantel treatment in humans, 4) to identify predictors of praziquantel efficacy in humans. My analyses revealed three factors that have an influence on the protection levels provided by attenuated schistosome parasite vaccines: increasing numbers of immunizing parasites had a positive effect on the levels of protection whereas increasing the radiation dose and the time to challenge infection had negative effects. Analyses showed that the attenuated schistosome vaccine has the potential to achieve protection levels as high as 79% after a single dose in mice. Alongside this, baboon studies consistently reported protective effects of attenuated schistosome vaccines against re-infection. These results show there is a high potential for an attenuated schistosome parasite vaccine to be effective in humans. A meta-analysis of the influence of praziquantel treatment on the direction of change in schistosome-specific antibody isotypes was conducted. The analysis revealed considerable variability in the antibodies’ direction of change among populations. The results also demonstrated an increase of anti-worm IgA and IgE in the majority of studies. These antibodies have been reported to have a protective effect against re-infection. The combination of age and infection intensity, and the number of days after treatment were identified as influential predictors for some antibody isotypes, but there was no single predictor that consistently affected all antibody isotypes in the same way. Praziquantel efficacy levels in humans were investigated and the analyses revealed that cure rates for schistosomiasis increase with praziquantel dose, and were affected by the identity of the schistosome parasite species (S. mansoni vs. S. haematobium) and the age of the participants (children: 0-19 years old vs. adults: ≥ 20 years old). There has been no clear efficacy level reduction over the treatment years (1979-2013) suggesting that praziquantel is still effective in the treatment of schistosomiasis despite concerns about possible resistance. The development of a schistosome vaccine will benefit from a closer investigation into the mechanisms through which protection is acquired in attenuated schistosome parasite vaccine studies showing high potential efficacy in animal models. Nevertheless, it will take time to develop a schistosome vaccine for human use. The uptake of the vaccine will be made even more challenging by the lack of adequate infrastructure in schistosomiasis endemic areas. In the meantime, close monitoring of praziquantel efficacy levels is necessary to confirm the effectiveness of schistosomiasis control in endemic areas.

MAZIN, S. C.Metodos Bayesianos em Metanalise: Especicac~ao da Distribuic~ao a Priori para a Variabilidade entre os Estudos. 2009. 175f. Dissertac~ao (mestrado) - Faculdade de Medicina de Ribeir~ao Preto, Universidade de S~ao Paulo, Ribeir~ao Preto, 2009. Prossionais da saude, pesquisadores e outros responsaveis por polticas de saude s~ao frequentemente inundados com quantidades de informac~oes nem sempre manejaveis, o que torna a revis~ao sistematica uma maneira eciente de integrar o conhecimento existente gerando dados que auxiliem a tomada de decis~ao. Em uma revis~ao sistematica os dados dos diferentes estudos podem ser quantitativamente combinados por metodos estatsticos chamados metanalise. A metanalise e uma ferramenta estatstica utilizada para combinar ou integrar os resultados dos diversos estudos independentes, sobre o mesmo tema. Entre os estudos que comp~oem a metanalise pode existir uma variabilidade que n~ao e devida ao acaso, chamada heterogeneidade. A heterogeneidade e geralmente testada pelo teste Q ou quanticada pela estatstica I2. A investigac~ao da heterogeneidade na metanalise e de grande import^ancia pois a aus^encia ou a presenca indica o modelo estatstico mais adequado. Assim, na aus^encia desta variabilidade utilizamos um modelo estatstico de efeito xo e na presenca utilizamos um modelo de efeitos aleatorios que incorpora a variabilidade entre os estudos na metanalise. Muitas metanalises s~ao compostas por poucos estudos, e quando isso acontece, temos diculdades de estimar as medidas de efeito metanalticas atraves da teoria classica, pois esta e dependente de pressupostos assintoticos. Na abordagem bayesiana n~ao temos esse problema, mas devemos ter muito cuidado com a especicac~ao da distribuic~ao a priori. Uma vantagem da infer^encia bayesiana e a possibilidade de predizer um resultado para um estudo futuro. Neste trabalho, conduzimos um estudo sobre a especicac~ao da distribuic~ao a priori para o par^ametro que expressa a vari^ancia entre os estudos e constatamos que n~ao existe uma unica escolha que caracterize uma distribuic~ao a priori que possa ser considerada ~ao informativa\"em todas as situac~oes. A escolha de uma distribuic~ao a priori ~ao informativa\"depende da heterogeneidade entre os estudos na metanalise. Assim a distribuic~ao a priori deve ser escolhida com muito cuidado e seguida de uma analise de sensibilidade, especialmente quando o numero de estudos e pequeno. / MAZIN, S. C. Bayesian methods in meta-analysis: specication of prior distributions for the between-studies variability. 2009. 175s. Dissertation (master degree) - Faculty of Medicine of Ribeir~ao Preto, University of S~ao Paulo, Ribeir~ao Preto, 2009. Health professionals, researchers and others responsible for health policy are often overwhelmed by amounts of information that can not always be manageable, which makes the systematic review an ecient way to integrate existing knowledge generating information that may help decision making. In a systematic review, data from dierent studies can be quantitatively combined by statistical methods called meta-analysis. The meta-analysis is a statistical tool used to combine or integrate the results of several independent studies on the same topic. Among the studies that comprise the meta-analysis we have a variability that does not yield from the chance, called the heterogeneity. Heterogeneity is usually tested by Q or quantied by the statistic I2. The investigation of heterogeneity in meta-analysis has a great importance because the absence or presence indicates the most appropriate statistical model. In the absence of this variability we used a xed eect statistical model and a random eects model was used to incorporate the variability between studies in the meta-analysis. Many meta-analysis are composed of few studies, and in those cases, it is dicult to estimate the eect of meta-analytic measures by the classical theory because the asymptotic assumptions. In the Bayesian approach we do not have this problem, but we must be very careful about the specication of prior distribution. One advantage of Bayesian inference is the ability to predict an outcome for a future study. In this work, carried out a study about the specication of prior distribution for the parameter that expresses of the variance between studies and found that there is no single choice that features a prior distribution that would be considered uninformative at all times. The choice of a prior distribution uninformative depend heterogeneity among studies in the meta-analysis. Thus, the prior distribution should be examined very carefully and followed by a sensitivity analysis, especially when the number of studies is small.

A vast amount of gene expression data is available to biological researchers. As of October 2010, the GEO database has 45,777 chips of publicly available gene expression pro ling data from the Affymetrix (HGU133v2) GeneChip platform, representing 2.5 billion numerical measurements. Given this wealth of data, `meta-analysis' methods allowing inferences to be made from combinations of samples from different experiments are critically important. This thesis explores the application of localized pattern-mining approaches, as exemplified by biclustering, for large-scale gene expression analysis. Biclustering methods are particularly attractive for the analysis of large compendia of gene expression data as they allow the extraction of relationships that occur only across subsets of genes and samples. Standard correlation methods, however, assume a single correlation relationship between two genes occurs across all samples in the data. There are a number of existing biclustering methods, but as these did not prove suitable for large scale analysis, a novel method named `IslandCluster' was developed. This method provided a framework for investigating the results of different approaches to biclustering meta-analysis. The biclustering methods used in this work involve preprocessing of gene expression data into a unified scale in order to assess the significance of expression patterns. A novel discretisation approach is shown to identify distinct classes of genes' expression values more appropriately than approaches reported in the literature. A Gene Expression State Transformation (`GESTr') introduced as the first reported modelling of the biological state of expression on a unified scale and is shown to facilitate effective meta-analysis. Localised co-dependency analysis is introduced, a paradigm for identifying transcriptional relationships from gene expression data. Tools implementing this analysis were developed and used to analyse specificity of transcriptional relationships, to distinguish related subsets within a set of transcription factor (TF) targets and to tease apart combinatorial regulation of a set of targets by multiple TFs. The state of pluripotency, from which a mammalian cell has the potential to differentiate into any cell from any of the three adult germ layers, is maintained by forced expression of Nanog and may be induced from a non-pluripotent state by the expression of Oct4, Sox2, Klf4 and cMyc. Analysis of cMyc regulatory targets shed light on a recent proposition that cMyc induces an `embryonic stem cell like' transcriptional signature outside embryonic stem (ES) cells, revealing a cMyc-responsive subset of the signature and identifying ES cell expressed targets with evidence of broad cMyc-induction. Regulatory targets through which cMyc, Oct4, Sox2 and Nanog may maintain or induce pluripotency were identified, offering insight into transcriptional mechanisms involved in the control of pluripotency and demonstrating the utility of the novel analysis approaches presented in this work.

Meta-analysis is a statistical technique that combines results from multiple independent studies to make inferences about parameters of interest. Although it is popular for parameter estimation and hypothesis testing, meta-analytic approaches that incorporate heterogeneous studies have not been fully developed. For heterogeneous studies, we do not expect all of the studies to have the same true underlying effect and the use of the fixed-effects model in a meta-analysis in this situation violates the assumption of homogeneity of effect size. Heterogeneity among studies can arise from multiple sources such as differences in populations by ancestry, differences in study designs, and different impacts of environmental exposures on the effect of the variable of interest. In this thesis, we introduce an analytic strategy and statistical models for meta-analysis of potentially heterogeneous studies. First, we propose a two-stage clustering approach to account for heterogeneity in trans-ethnic meta-analysis of genome-wide association studies (GWAS). Specifically, we cluster studies in the two-stage approach using cohort-specific genetic information prior to meta-analysis to account for between-cluster heterogeneity as well as to bolster within-cluster homogeneity. An extensive simulation study shows that this approach improves power and diminishes computational intensity compared to existing methods for trans-ethnic meta-analysis. Next, under a meta-regression framework, we develop a likelihood ratio test (LRT) statistic to accommodate multiple random effects. We allow multiple sources of heterogeneity in terms of study characteristics and model the heterogeneities as random effects. We show that the proposed LRT maintains a similar or higher power than other existing methods in a simulation study especially when heterogeneity exists. We apply this new approach to meta-analyze genome-wide association data. Lastly, we derive a score test in the same context as our proposed new LRT and show the substantial advantage of the score test in computational efficiency compared to the new LRT. The introduced strategy and methodologies can effectively and efficiently aggregate the evidence from potentially heterogeneous studies in statistical genetics and other research areas.

In a systematic review and random effects meta-analysis, we evaluated whether obesity is associated with postoperative atrial fibrillation (POAF) in patients undergoing cardiac surgery. Eighteen observational studies that excluded patients with preoperative AF were selected until December 2011 (n=36,147). Obese patients had a modest higher risk of POAF in comparison to non-obese (OR 1.12, 95%CI 1.04-1.21, p=0.002). The association between obesity and POAF did not vary substantially by type of cardiac surgery, study design or year of publication. POAF was significantly associated with higher risk of stroke, respiratory failure, and operative mortality. / Revisión por pares