O objetivo deste trabalho foi investigar a divulgação de informações sobre operações de combinação de negócios, ocorridas no Brasil no ano de 2010 e verificar quais fatores influenciaram o nível de disclosure. A partir do ano de 2010 entrou em vigor o Pronunciamento Técnico CPC 15 - Combinação de Negócios, relacionado com a norma internacional IFRS 3 - Business Combination, a qual estabelece as informações que devem ser divulgadas numa combinação de negócios. Apesar do CPC 15 determinar quais informações devem ser divulgadas, observaram-se diferenças no nível de informação divulgada entre as empresas. Para se analisar o nível de divulgação de informações sobre operação de combinação de negócios, construi-se um índice de disclosure de combinação de negócios (INDCOMB), baseado nas informações a serem divulgadas pelas empresas, de acordo com o CPC-15. A partir desse índice, são analisados, por meio de regressão multivariada, quais fatores influenciavam o nível de disclosure. Tendo como variável dependente o índice INDCOMB, avalia-se se os seguintes fatores influenciavam no nível de disclosure: o porte da empresa adquirente, o percentual reconhecido do ágio por expectativa de rentabilidade futura (goodwill) reconhecido em relação ao valor da transação, a dispersão do capital da empresa adquirente, o porte da empresa de auditoria e a participação da adquirente em programas de ADRs, American Depositary Receipts. São utilizados como variáveis de controle a listagem da adquirente nos segmentos diferenciados da BM&FBOVESPA, o setor de atuação, a origem (estatal, privada de capital nacional ou privada de capital estrangeiro) e o porte relativo da empresa adquirida em relação à empresa adquirente. Analisam-se operações de combinação de negócios de 40 companhias abertas envolvidas em 76 operações de combinação de negócios. De acordo com os resultados, o porte da empresa de auditoria e o porte relativo da empresa adquirida são fatores que influenciam o nível de divulgação de informações sobre combinação de negócios. Os demais fatores não apresentam resultados conclusivos, conforme o estudo realizado. / The aim of this work has been to analyze the disclosure of information on business combination carried out in Brazil in the year 2010 and to assess the factors that influence the level of disclosure. From the year 2010 onwards the provisions of the CPC-15 - Business Combinations statement started its effects. CPC 15 is related to the IASB international standard IFRS 3 - Business Combination, that provides the information must be disclosed in a business combination operation. Although CPC-15 determines the compulsory information that should be disclosed, it was observed that there are differences in the level of information disclosed between companies. To analyze the level of disclosure of information about the operations an index of disclosure of a business combination (INDCOMB) has been built, based on information that companies should disclose according to the CPC-15. Then the factors which influenced the level of disclosure was analyzed using multivariate regression. The dependent variable is the index INDCOMB and the following factors has been analyzed to assess the influenced on the disclosure level: the acquirer size, the percentage of goodwill recognized relative to the value of the transaction, the acquirer free float, the auditing firm size and the acquirer participation on ADRs (American Depositary Receipts). As control variables the listing status of acquirer on the special segments of BM&FBOVESPA, its sector of activity, the source (state, private domestic capital or private foreign capital), the acquired relative size regarding the acquiring company was used. Business combination of 40 public companies involved in 76 operations has been analyzed. The results suggest that auditing firm size and the relative size of the acquired company are factors that influence the level of disclosure of information on a business combination. The other factors did not show conclusive results, according this work.

O gênero Heliura Butler contava, no início deste trabalho, com 53 nomes e 40 espécies válidas. Foi realizada uma análise cladística com o intuito de testar o monofiletismo do gênero e construir uma hipótese de relações filogenéticas entre suas espécies. A análise mostrou que o conceito prévio de Heliura era polifilético, o que também se revelou verdadeiro para todos os gêneros estudados que tiveram mais de uma espécie incluída nas análises. Este gênero, como aqui redefinido, é composto por 66 espécies no sensu stricto, dentre as quais 16 são espécies novas, e 76 no sensu lato (incluindo as espécies incertae sedis). Tal rearranjo conta com dois novos sinônimos para Heliura, Ptychotricos Schaus, sin. nov. e Mesocerea Hampson, sin. nov. Todas as espécies que pertencem a Heliura no senso revisado foram redescritas e ilustradas, e tiveram sua distribuição geográfica mapeada. As demais foram realocadas de acordo com o que foi possível apurar a respeito de suas relações filogenéticas. Dentre as que foram realocadas com sucesso, estão Eucereon baleris Dyar, comb. nov. e Pseudaethria cosmosomodes Dognin, comb. nov. Dois gêneros novos são criados para realocar outras espécies que não pertencem a Heliura: Bus, gen. nov. e Dus, gen. nov. Entretanto, não foi possível realocar todas elas, de modo que as demais receberam o status de incertae sedis. Onze novos sinônimos foram descobertos: Heliura cadroe Schaus (= Acridopsis lucis Butler), Pseudaethria cessogae Schaus (= Heliura cosmosomodes Dognin), Pseudohyaleucerea manicorensis Rego Barros & Machado (= Heliura quadriflavata Kaye), Delphyre nilammon Schaus (= Eucereon inconspícua Kaye), Heliura klagesi meridionalis Rothschild, Delphyre lemoulti Draudt (= Neacerea rhodocrypta Druce), Automolis oviplaga Rothschild (= Delphyre subapicalis Dukinfield-Jones), Theages quadricolor Walker, Eucereon quadricolor boreale Rothschild e E. quadricolor meridionale Rothschild (estes três = Chelonia punctata Guérin-Meneville) e Eucereon tigrisoma Rothschild (= Galethalea pica Walker). Outras duas espécies também tratadas aqui em Heliura, H. pierus Cramer e H. dares Cramer, são declaradas species inquirendae. Heliura distincta Rothschild passa a ser conhecida como Heliura rothschildi nom. nov., uma vez que ,Teucer distincta Rothschild, um ano mais antiga, também passa a fazer parte de Heliura. A combinação nova Heliura elongata (Schaus), comb. nov. é mais antiga que H. elongata Rothschild, e, portanto, este último nome passa a ser conhecido como H. umbrimaculodes nom. nov. São apresentadas notas sobre Delphyre Walker e Eucereon Hübner, com a revalidação de alguns de seus sinônimos (Neacerea Druce, gen. revalid. e Erithales Poey, gen. revalid.), além da criação de um gênero novo, Aus, gen. nov., para algumas espécies previamente alocadas em Delphyre. As identidades de Eucereon archias e E. punctatum são discutidas à luz de novas descobertas. Novas combinações são propostas em Galethalea Butler, Pseudohyaleucerea Rego Barros & Machado, Diabaena Felder, Pseudopharus Hampson, Eucereon Hübner e Rhipha Walker. Outras duas espécies novas são descritas, em Delphyre e Erithales. Lectótipos foram designados quando apropriado para todos os nomes descritos ou presumivelmente descritos a partir de mais de um espécime / The genus Heliura Butler had 53 names and 40 valid species at the beginning of this study. A cladistic analysis was performed to test its monophyletism, which results showed that it is polyphyletic, as well as all other genera included in the analysis and represented by more than one taxon. Heliura, as defined here, comprises 66 species in its sensu stricto, 16 of which are new, and 76 in its sensu lato (which includes incertae sedis species). This arrangement counts with two new synonyms for Heliura, Ptychotricos Schaus, sin. nov. e Mesocerea Hampson, sin. nov. All the species belonging to Heliura in the sense here defended were redescribed, illustrated and mapped. The other ones were rearranged according to the results obtained at the analysis. Among those successfully placed in genera already described are Eucereon baleris Dyar, comb. nov. and Pseudaethria cosmosomodes Dognin, comb. nov. Two new genera were created to place other species that do not belong in Heliura: Bus, gen. nov. and Dusi, gen. nov. However, it was not possible to place confidently all the species that do not belong in Heliura, and those which phylogenetic positions remain a mistery were given the status of incertae sedis. Eleven new synonyms were discovered: Heliura cadroe Schaus (= Acridopsis lucis Butler), Pseudohyaleucerea manicorensis Rego Barros & Machado (= Heliura quadriflavata Kaye), Delphyre nilammon Schaus (= Eucereon inconspicua Kaye), Heliura klagesi meridionalis Rothschild, Delphyre lemoulti Draudt (= Neacerea rhodocrypta Druce), Automolis oviplaga Rothschild (= Delphyre subapicalis Dukinfield-Jones), Theages quadricolor Walker, Eucereon quadricolor boreale Rothschild e E. quadricolor meridionale Rothschild (these three = Chelonia punctata Guérin-Meneville), and Eucereon tigrisoma Rothschild (= Galethalea pica Walker). Two other species here treated in Heliura were declared species inquirendae: H. Pierus Cramer and H. Dares Cramer. Heliura distinct Rothschild received a new name, Heliura rothschildi, nom. nov., because Teucer distinct Rothschild, which is one year older, is now also part of Heliura. At last, notes on Delphyre Walker and Eucereon Hübner are provided, with the revalidation of some of its synonyms (Neacerea Druce, gen. revalid. and Erithales Poey, gen. revalid.), plus the creation of a new genus, Aus, gen. nov., for some species previously placed in Delphyre. The identities of Eucereon archias and E. Punctatum are discussed based on new evidence. New combinations are proposed in Galethalea Butler, Pseudohyaleucerea Rego Barros & Machado, Diabaena Felder, Pseudopharus Hampson, Eucereon Hübner, and <i.Rhipha Walker. Two other new species are described, in Delphyre and Erithales. Lectotypes were designated when appropriated for all names described or supposedly described from more than one specimen

A leucemia linfóide aguda (LLA) é a neoplasia hematológica mais comum na infância e representa uma doença heterogênea em relação à biologia e ao prognóstico e seu tratamento consiste principalmente em quimioterapia. Apesar dos avanços no tratamento, cerca de 20% dos pacientes apresentam recaída da doença e/ou óbito indicando a necessidade de terapias diferenciadas para esse grupo. Recentemente, drogas epigenéticas como inibidores de DNA metiltransferases (iDNMTs) tem mostrado efeitos anti-neoplásicos promissores para o tratamento de diversos tipos de neoplasias incluindo a LLA. Nos tumores, a hipermetilação gênica é encontrada em vários genes, incluindo genes de reparo do DNA, reguladores do ciclo celular e apoptose. Sendo assim, drogas desmetilantes estão sendo apontadas como promissores agentes para o tratamento do câncer. A Zebularina (ZB) é um iDNMT análogo de citidina que inibe a metilação do DNA. Esta droga tem mostrado resultados animadores para o tratamento de diversas neoplasias, incluindo glioblastoma, leucemia mielóide aguda, câncer de mama, próstata e outros. O objetivo deste trabalho foi avaliar os efeitos do tratamento com a ZB, associada ou não à quimioterápicos, em linhagens celulares pediátricas de LLA, por meio de ensaios funcionais como proliferação celular, capacidade clonogênica, apoptose e ciclo celular. Além disso, foi analisada a capacidade desmetilante da droga e a expressão dos genes DNMT1, DNMT3a e DNMT3b após o tratamento com a ZB. A ZB inibiu a proliferação celular de maneira dose e tempo-dependente e agiu sinergicamente quando combinada com o MTX em ambas as linhagens. Ela também diminuiu a capacidade clonogênica e aumentou a taxa de apoptose nas duas linhagens estudadas. Além disso, o tratamento com ZB causou uma parada na fase S do ciclo celular na linhagem ReH. A ZB foi capaz de desmetilar parcialmente o gene AhR e reduzir a expressão dos genes DNMT1, DNMT3a e DNMT3b. Todos os dados encontrados no presente trabalho sugerem que as drogas desmetilantes podem ser interessantes agentes para o tratamento da LLA pediátrica. / Acute lymphoblastic leukemia (ALL) is the most common hematologic malignancy in childhood and represents a heterogeneous disease regarding its biology and prognosis. Its treatment consists mainly of chemotherapy. Despite advances in treatment, about 20% of patients experience disease recurrence and/or death indicating the need for differentiated therapies for this group. Recently, epigenetic drugs such as DNA methyltransferases inhibitors (iDNMTs) has shown antineoplastic and promising results for several types of tumors including ALL. Gene hypermethylation is found in several genes in tumors cells, including genes responsible for DNA repair, cell cycle and apoptosis regulators. Therefore, demethylating agents may be promising agents for cancer treatment. Zebularine (ZB) an iDNMT is a cytidine analogue that inhibits DNA methylation. This drug has shown promising results for the treatment of many cancers, including glioblastoma, acute myeloid leukemia, breast and prostate cancer and others. The aim of this study was to evaluate the effects of ZB treatment, associated or not with chemotherapeutic agents, in childhood ALL cell lines through functional tests such as cell proliferation, clonogenic capacity, apoptosis and cell cycle. In addition, we examined the demethylating ability of ZB and the expression of DNMT1, DNMT3a and DNMT3b genes after treatment with this agent. ZB inhibited cell proliferation in a dose- and time-dependent manner and showed synergistic effects when combined with MTX in both cell lines. ZB treatment also reduced clonogenic capacity and increased the number of apoptotic cells in both cell lines studied. Furthermore, treatment with ZB caused an S phase cell cycle arrest in ReH cell line. ZB was able to partially demethylate AhR gene and reduce the expression of genes DNMT1, DNMT3a and DNMT3b. These results suggest that demethylating drugs may be interesting agents for the treatment of childhood ALL.

Breast cancer is the most common cancer among Canadian women and 14-20% will develop lethal metastases within 5 years. A potential novel therapeutic target is Focal Adhesion Kinase (FAK), a cytoplasmic tyrosine kinase. FAK’s expression is inversely correlated with survival and is known to regulate cell migration, proliferation and invasion. While tyrosine kinase inhibitors are historically ineffective as single agents, they are commonly used as part of combination therapies. Therefore, given its central role in tumor cell biology and cell signaling, we hypothesized that inhibiting FAK in combination with pharmacological agents commonly used to treat metastatic breast cancer patients will result in enhanced anti-tumor activity. We combined a commercial FAK inhibitor (PF-562271) with a range of chemotherapeutic agents commonly used to treat metastatic breast cancer and searched for synergistic partners. Only DNA topoisomerase inhibitors showed potential to synergistically reduce cell viability when paired with low doses of the FAK inhibitor. However, the combination does not induce an increase in cell death or apoptosis. It was then discovered that both agents in isolation and in combination produce increased levels of ROS, a toxic metabolite. This, along with other more preliminary data, provides clues for a novel proposed mechanism of action for this interaction.

Magister Public Health - MPH / Background Nigeria accounted for 2.95 million of the 22 million people globally living with HIV in 2008. In 2010, the HIV prevalence increased to 3.1 million, with 1.5 million people requiring anti-retroviral treatment (ART). ART is effective if patients adhere to treatment (taking 95% or more of drugs as prescribed) over a sustained period. Taking less than 95% of the medication can lead to drug resistance and treatment failure, which have dire individual and public health consequences. This study described adherence to ART and the factors that constrain and motivate adherence among patients on ART at the University of Abuja Teaching Hospital in the Federal Capital Territory (FCT), Nigeria. Methodology An observational, descriptive and analytical, cross-sectional survey of adherence among 502 adult ART patients (254 women and 248 men) from the University of Abuja Teaching Hospital was conducted. I collected sociodemographic and clinical characteristics of participants, and barriers and facilitators to adherence. For the prescription refill data, I utilized the updated pharmacy refill records from the ART dispensary. Bivariate and multivariate analysis was performed to analyse the factors that influence adherence to ART. Results Participants in this study had been on therapy for a mean of 43±27 months. Total optimal self-reported adherence over the previous three days (not missing a dose, taking correct doses in the correct frequency and correct schedule) was 53.6%, compared with 62.5% adherence calculated by prescription refill. However, most (80.3%) participants achieved virologic suppression at a level of <400 copies/μl3. Reported barriers to adherence were: forgot (43%); travelled away from home (21%); ran out of medication (16%); busy at work (13%); lack of food (5%) and medication snatched by armed robbers (2%). Self-reported adherence over the previous three days was positively associated with age and viral load. Younger respondents (under 30 years) were 3 times more likely to adhere to their regimen compared with those older than 30 years (OR = 2.5; 95% CI = 1.26-4.61; p =

Magister Public Health - MPH / In Zambia, there had been a large scaling up of new interventions to control malaria since 2003, which included the distribution of rapid diagnostic tests (RDTs), used to immediately determine if someone with symptoms suggestive of malaria actually has malaria; training of health workers in the use of the RDTs; and the prescription of artemisinin-based combination therapy (ACT) to which the malaria parasite is sensitive, rather than the old treatment regime of chloroquine to which the malaria parasite had become resistant. The use of RDTs to confirm the presence of malaria before treating for it with ACT became known as the „test and treat‟ policy. Previously, since the 1960s, in malaria endemic areas such as Zambia, children presenting with fever (the commonest symptom of malaria) without any obvious other cause for the fever, were assumed to have malaria and were hence treated for it with chloroquine. This was known as "presumptive treatment" of malaria. The combination of "presumptive treatment" and the use of a single medication led to the development of high levels of resistance to chloroquine, to the extent that it is now no longer an effective treatment for malaria. Years after the introduction of the "test and treat" policy, it was still unclear to what extent it was being implemented, as there was initial reluctance by health workers to test all children presenting with fever for malaria and if they did test they may not have followed the management guidelines of treating those who test positive with ACT and further investigating those who test negative for the cause of the fever. It seemed that staff had gotten used to the "presumptive treatment" approach to malaria over almost 4 decades and hence were quite reluctant to abandon it. The conflicting guidelines for malaria treatment in children between IMCI and "test and treat‟ has promoted a paradox between presumptive treatment for malaria and "test and treat" approach as IMCI teaches health workers to treat febrile children presumptively for malaria whereas the "test and treat" approach requires them to first make a definitive diagnosis before treating. Hence although the "test and treat" approach was instituted to overcome the problems with presumptive treatment approach it now had to contend with the competing and contradictory influence of the IMCI approach. This study therefore aimed to assess what proportion of children aged five years and younger who presented with fever were managed via the "test and treat" guidelines and which factors were associated with this, in Lusaka District, Zambia. Methodology: A cross sectional analytical study design was used based on a review of medical records. A sample size of 800 medical records of children presenting with fever was selected from 10 out of the 23 health care facilities in Lusaka, using a multistage stratified random sampling technique. Four hundred records were sampled from 2008 records (five years after commencement of the "test and treat" policy) and 400 from 2011 records (eight years after commencement of the "test and treat" policy). Trained data collectors used a data extraction tool to transcribe demographic and clinical data from the medical records in a standardized manner. Data Analysis: Univariate descriptive statistics analysis was performed using measures of central tendency and measures of dispersion to analyze numerical (continuous) variables such as age, weight and body temperature; and using frequencies for categorical variables such as gender, area of residence, RDTs/microscopy malaria tests conducted, received ACT if RDT positive, presence of an ACT treatment chart on the health centre wall and availability of a weighing scale. To determine the relationship between variables, bivariate analysis via the prevalence ratio was conducted. Results: Just over half (55%) of all children with fever were tested for malaria in 2008 and this gratifyingly increased to (73%) in 2011. Overall, the proportion of children correctly and appropriately treated with ACT, which means that those who tested positive for malaria were given ACT, was 85% in 2008 but regrettably dropped to 72% in 2011. Although "presumptive treatment" decreased from 24% in 2008 to 11% in 2011, the proportion of children with fever not tested for malaria, and although not treated for malaria, but left without a definitive diagnosis of their fever being made, remained high but dropping (22% in 2008 and 16% in 2011). Similarly the proportion of children who tested negative for malaria but then did not undergo any further investigation also unfortunately remained very high and rising (57% in 2008 and 89% in 2011). A combination of the above poor clinical management practises resulted in only 38% of children with fever in 2008 and unfortunately dropping to only 33% in 2011 being correctly managed (tested for malaria via RDT or microscopy and treated with ACT if positive, while further investigated for the cause of fever if negative). On preparedness of the health facility to implement the "test and treat" policy, it was noted that only 4 out of 10 health facilities were at least minimally prepared to do so, but paradoxically on bivariate analysis those minimally prepared were less likely (PR 0.62; 95% CI 0.41-0.94) to correctly manage the patients in 2011 than those who were unprepared. A similar paradox occurred for those correctly treated with ACT after testing positive, with facilities which were minimally prepared being less likely to do so (PR 0.28; 95% CI 0.14-0.58) in 2011 than those facilities which were unprepared to implement the "test and treat" policy. However these associations were inconsistent over time, as the associations were not present in 2008. Similarly all other factors such as staff category (doctor, nurse, clinical officer) and type of presenting symptoms besides fever (anorexia, lethargy, pallor) assessed, were not consistently associated with testing for malaria in both 2008 and 2011. The same applied for the other two main outcome variables of 'treated with ACT after test positive for malaria' and 'correctly managed child with fever', in that there were no factors that showed a consistent association with them in both 2008 and 2011. Conclusion: Testing of children with fever for malaria is at a low level but rose between 2008 and 2011. Paradoxically the proportion of those diagnosed with malaria who were correctly treated with ACT dropped between 2008 and 2011, as did the proportion of children with fever who were correctly managed. No factors assessed in this study were found to be consistently associated in both 2008 and 2011 with either testing for malaria, or treating confirmed malaria cases with ACT, or managing patients with fever correctly. Recommendations: In order for health workers to correctly implement the "test and treat" policy, which involves a series of complex steps, they ought to be formally trained to do so, mentored and constructively supervised. Additionally health facilities should be adequately equipped to enable health workers to fully implement the policy. Further studies to assess factors associated with the correct management of malaria via the "test and treat" policy are warranted.

Influenza viruses are respiratory pathogens that are responsible for annual influenza epidemics and sporadic influenza pandemics. Oseltamivir (Tamiflu((R))) is currently the only FDA-approved oral drug that is available for the prevention and treatment of influenza virus infection. However, its narrow therapeutic window, coupled with the increasing incidence of drug resistance, calls for the next generation of influenza antivirals. In this study, we discovered hesperadin, an aurora B kinase inhibitor, as a broad-spectrum influenza antiviral through forward chemical genomics screening. Hesperadin inhibits multiple human clinical isolates of influenza A and B viruses with single to submicromolar efficacy, including oseltamivir-resistant strains. Mechanistic studies revealed that hesperadin inhibits the early stage of viral replication by delaying the nuclear entry of viral ribonucleoprotein complex, thereby inhibiting viral RNA transcription and translation as well as viral protein synthesis. Moreover, a combination of hesperadin with oseltamivir shows synergistic antiviral activity, therefore hesperadin can be used either alone to treat infections by oseltamivir-resistant influenza viruses or used in combination with oseltamivir to delay resistance evolution among oseltamivir-sensitive strains. In summary, the discovery of hesperadin as a broad-spectrum influenza antiviral offers an alternative to combat future influenza epidemics and pandemics.

Nous abordons, dans cette thèse, les méthodes de combinaison de systèmesde transcription de la parole à Large Vocabulaire. Notre étude se concentre surl’attelage de systèmes de transcription hétérogènes dans l’objectif d’améliorerla qualité de la transcription à latence contrainte. Les systèmes statistiquessont affectés par les nombreuses variabilités qui caractérisent le signal dela parole. Un seul système n’est généralement pas capable de modéliserl’ensemble de ces variabilités. La combinaison de différents systèmes detranscription repose sur l’idée d’exploiter les points forts de chacun pourobtenir une transcription finale améliorée. Les méthodes de combinaisonproposées dans la littérature sont majoritairement appliquées a posteriori,dans une architecture de transcription multi-passes. Cela nécessite un tempsde latence considérable induit par le temps d’attente requis avant l’applicationde la combinaison.Récemment, une méthode de combinaison intégrée a été proposée. Cetteméthode est basée sur le paradigme de décodage guidé (DDA :Driven DecodingAlgorithm) qui permet de combiner différents systèmes durant le décodage. Laméthode consiste à intégrer des informations en provenance de plusieurs systèmes dits auxiliaires dans le processus de décodage d’un système dit primaire.Notre contribution dans le cadre de cette thèse porte sur un double aspect : d’une part, nous proposons une étude sur la robustesse de la combinaison par décodage guidé. Nous proposons ensuite, une amélioration efficacement généralisable basée sur le décodage guidé par sac de n-grammes,appelé BONG. D’autre part, nous proposons un cadre permettant l’attelagede plusieurs systèmes mono-passe pour la construction collaborative, à latenceréduite, de la sortie de l’hypothèse de reconnaissance finale. Nous présentonsdifférents modèles théoriques de l’architecture d’attelage et nous exposons unexemple d’implémentation en utilisant une architecture client/serveur distribuée. Après la définition de l’architecture de collaboration, nous nous focalisons sur les méthodes de combinaison adaptées à la transcription automatiqueà latence réduite. Nous proposons une adaptation de la combinaison BONGpermettant la collaboration, à latence réduite, de plusieurs systèmes mono-passe fonctionnant en parallèle. Nous présentons également, une adaptationde la combinaison ROVER applicable durant le processus de décodage via unprocessus d’alignement local suivi par un processus de vote basé sur la fréquence d’apparition des mots. Les deux méthodes de combinaison proposéespermettent la réduction de la latence de la combinaison de plusieurs systèmesmono-passe avec un gain significatif du WER. / This thesis presents work in the area of Large Vocabulary ContinuousSpeech Recognition (LVCSR) system combination. The thesis focuses onmethods for harnessing heterogeneous systems in order to increase theefficiency of speech recognizer with reduced latency.Automatic Speech Recognition (ASR) is affected by many variabilitiespresent in the speech signal, therefore single ASR systems are usually unableto deal with all these variabilities. Considering these limitations, combinationmethods are proposed as alternative strategies to improve recognitionaccuracy using multiple recognizers developed at different research siteswith different recognition strategies. System combination techniques areusually used within multi-passes ASR architecture. Outputs of two or moreASR systems are combined to estimate the most likely hypothesis amongconflicting word pairs or differing hypotheses for the same part of utterance.The contribution of this thesis is twofold. First, we study and analyze theintegrated driven decoding combination method which consists in guidingthe search algorithm of a primary ASR system by the one-best hypothesesof auxiliary systems. Thus we propose some improvements in order to makethe driven decoding more efficient and generalizable. The proposed methodis called BONG and consists in using Bag Of N-Gram auxiliary hypothesisfor the driven decoding.Second, we propose a new framework for low latency paralyzed single-passspeech recognizer harnessing. We study various theoretical harnessingmodels and we present an example of harnessing implementation basedon client/server distributed architecture. Afterwards, we suggest differentcombination methods adapted to the presented harnessing architecture:first we extend the BONG combination method for low latency paralyzedsingle-pass speech recognizer systems collaboration. Then we propose, anadaptation of the ROVER combination method to be performed during thedecoding process using a local vote procedure followed by voting based onword frequencies.

Several wild-type poxviruses have emerged as potential oncolytic viruses (OVs), including orf virus (OrfV), and vaccinia virus (VV). Oncolytic VVs have been modified to include attenuating mutations that enhance their tumour selective nature, but these mutations also reduce overall viral fitness in cancer cells. Previous studies have shown that a VV (Western Reserve) with its E3L gene replaced with the E3L homologue from, OrfV (designated VV-E3LOrfV), maintained its ability to infect cells in vitro, but was attenuated compared to its parental VV in vivo. Our goal was to determine the safety and oncolytic potential VV-E3LOrfV, compared to wild type VV and other attenuated recombinants. VV-E3LOrfV, was unable to replicate to the same titers and was sensitive to IFN compared to its parental virus and other attenuated VVs in normal human fibroblast cells. The virus was also less pathogenic when administered in vivo. Viral replication, spread and cell killing, as measures of oncolytic potential in vitro, along with in vivo efficacy, were also observed.. The Parapoxvirus, OrfV has been shown to have a unique immune-stimulation profile, inducing a number of pro-inflammatory cytokines, as well as potently recruiting and activating a number of immune cells. Despite this unique profile, OrfV is limited in its ability to replicate and spread in human cancer cells. Various strategies were employed to enhance the oncolytic efficacy of wild-type OrfV. A transient transfection/infection screen was created to determine if any of the VV host-range genes (C7L, K1L, E3L or K3L) would augment OrfV oncolysis. Combination therapy, including the use of microtubule targeting agents, Viral Sensitizer (VSe) compounds and the addition of soluble VV B18R gene product were employed to see if they also enhance OrfV efficacy. Unfortunately, none of the strategies mentioned were able to enhance OrfV.

Nowell (1976) first proposed that unless cytotoxic cancer therapy eradicates all tumor cells, genetic or heritable variation within heterogeneous tumors will inevitably lead to the evolution of chemotherapeutic resistance through clonal selection. This evolutionary hypothesis was formalized by Goldie and Coldman (1979), who developed one of the earliest mathematical kinetic models of resistance evolution in neoplasms. Their model predicted that the likelihood of response and cure would be increased in combination vs single agent cytotoxic therapies. In a later study, Gardner (2002) developed a computational kinetic model to predict chemotherapeutic combinations, doses, and schedules most likely to result in patient response and prolonged life. This model predicts that combination therapy involving both cytotoxic and cytostatic drugs will be more effective than combination therapy involving only cytotoxic drugs. Thus far, no systematic evaluation of the Goldie and Coldman and Gardner hypotheses have been conducted in the metastatic clinical trial setting. Here I test these hypotheses using the results of over 700 phase II, III and II/III clinical trials. I show that, as predicted by Goldie and Coldman, both overall response rate and overall survival were greater in combination arms. Moreover, median duration of response – the key indicator of the rate of resistance evolution - was also greater in combination vs single agent arms. These results suggest that generally combination chemotherapy is more effective than single agent therapy for advanced solid tumors as predicted by Goldie and Coldman (1979) hypothesis and that, at least in the metastatic setting, the potential disadvantages of combination therapy with respect to accelerated resistance evolution are outweighed by the greater waiting times for resistance mutations to arise. By contrast, although combination cytotoxic and cytostatic therapy is associated with a greater average overall response rate than multi agent cytotoxic therapy, this is not the case for both median duration of response and overall survival. Hence, there is no evidence that, in contrast to the predictions of the Gardner (2002) model, combination cytotoxic and cytostatic therapy decreases the rate of resistance evolution relative to that obtaining under combination cytotoxic therapy.