Introduction: Colonization of the anogenital tract of pregnant women by Group B Streptococcus (GBS) is the main risk factor for early-onset neonatal sepsis. Identification of maternal colonization allows antimicrobial prophylaxis and prevention of vertical transmission. Objectives: To identify the genetic diversity and phenotypic characteristics of GBS using molecular biology techniques and culture in specific medium, and the epidemiological aspects of pregnant women colonized by this bacterium. Methods: It was performed a cross-sectional study, anogenital samples of 316 pregnant women were collected between 35 and 37 weeks and submited to culture in specific medium, antimicrobial susceptibility test, multiplex PCR, multi locus sequence typing (MLST) and multi locus variable number of tandem repeat analysis (MLVA). The epidemiological aspects of colonized and non-colonized pregnant women were also investigated. Results: It was obtained a prevalence of 36.4% by culture and 38.6% by PCR. Multiplex PCR had sensitivity of 100%, specificity of 96.5%, positive predictive value of 94.3% and negative predictive value of 100%. The most common serotypes were Ia, V, II and III. Resistance genes were identified in 34 samples. Sensitivity to penicillin was universal, 24.3% presented resistance to erythromycin and 14.8% to clindamycin. Evaluating the epidemiological variables, there were no differences between the colonized and non-colonized pregnant women. Diversity index and number of genotypes found by MLST was 0.608 and 6, and by MLVA was 0.840 and 15, respectively. Conclusion: We found a high prevalence of maternal GBS colonization, with serotype distribution similar to the Americas region. Multiplex PCR was more accurate than culture, and maternal epidemiological variables showed no difference when evaluating presence or absence of bacterial colonization, its serotypes and antimicrobial resistance. MLVA showed a higher discrimination capacity among the unrelated strains than MLST. / Submitted by Otto Henrique May Feuerschuette (otto.feurschuette@unisul.br) on 2018-04-26T00:54:22Z No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) OTTO TESE CORRIGIDA PÓS BANCA (1).pdf FINAL.pdf: 3081595 bytes, checksum: afad5beba2687558aafe2f1705d36e01 (MD5) / Approved for entry into archive by Silvane Cauz (silvane.cauz@unisul.br) on 2018-04-26T12:14:53Z (GMT) No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) OTTO TESE CORRIGIDA PÓS BANCA (1).pdf FINAL.pdf: 3081595 bytes, checksum: afad5beba2687558aafe2f1705d36e01 (MD5) / Made available in DSpace on 2018-04-26T12:14:53Z (GMT). No. of bitstreams: 2 license_rdf: 811 bytes, checksum: e39d27027a6cc9cb039ad269a5db8e34 (MD5) OTTO TESE CORRIGIDA PÓS BANCA (1).pdf FINAL.pdf: 3081595 bytes, checksum: afad5beba2687558aafe2f1705d36e01 (MD5) Previous issue date: 2018 / Introdução: A colonização do trato anogenital de gestantes pelo Estreptococo do grupo B (EGB) é o fator de risco primário para a sepse neonatal precoce. A identificação da colonização materna permite a profilaxia antimicrobiana e prevenção da transmissão vertical. Objetivos: Identificar a diversidade genética e as características fenotípicas do EGB utilizando-se técnicas de biologia molecular e cultura em meio específico, e avaliar os aspectos epidemiológicos de gestantes colonizadas por essa bactéria Métodos: Foi realizado um estudo transversal com 316 amostras anogenitais de gestantes entre 35 e 37 semanas para realização de cultura em meio específico, teste de sensibilidade aos antimicrobianos, PCR multiplex, multi locus sequence typing (MLST) e multi locus variable number of tandem repeat analysis (MLVA). Pesquisou-se também os aspectos epidemiológicos dessas gestantes. Resultados: Foi encontrada prevalência de 36,4% pela cultura e 38,6% pela PCR. A PCR multiplex apresentou sensibilidade de 100%, especificidade de 96,5%, valor preditivo positivo de 94,3% e valor preditivo negativo de 100%. Os sorotipos mais encontrados foram Ia, V, II e III. Os genes de resistência foram identificados em 34 amostras. A sensibilidade à penicilina foi universal, 24,3% apresentaram resistência à eritromicina e 14,8% à clindamicina. Avaliando-se as variáveis epidemiológicas, não se identificaram diferenças entre as gestantes colonizadas e não colonizadas. O índice de diversidade e o número de genogrupos encontrados pelo MLST foi 0,608 e 6, e pela MLVA foi 0,840 e 15, respectivamente Conclusão: Constatou-se uma alta prevalência de colonização materna, com distribuição dos sorotipos semelhante à região das Américas. A PCR multiplex foi mais acurada que a cultura, e as variáveis epidemiológicas maternas não apresentaram diferença significativa ao avaliar-se a presença ou não de colonização bacteriana, seus sorotipos e a resistência aos antimicrobianos. A MLVA apresentou uma capacidade de discriminação entre as cepas não relacionadas maior do que a MLST

Long term longitudinal surveys have the advantage to enable several sampling of the studied phenomena and then, with the repeated measures obtained, find a confirmed tendency. However, these long term surveys generate large epidemiological datasets including more sources of noise than normal datasets (e.g. one single measure per observation unit) and potential correlation in the measured values. Here, we studied data from a long-term epidemiological and genetic survey of malaria disease in two family-based cohorts in Senegal, followed for 19 years (1990-2008) in Dielmo and for 16 years (1993-2008) in Ndiop. The main objectives of this work were to take into account familial relationships, repeated measures as well as effect of covariates to measure both environmental and host genetic (heritability) impacts on the outcome of infection with the malaria parasite Plasmodium falciparum, and then use findings from such analyses for linkage and association studies. The outcome of interest was the occurrence of a P. falciparum malaria attack during each trimester (PFA). The two villages were studied independently; epidemiological analyses, estimation of heritability and individual effects were then performed in each village separately. Linkage and association analyses used family-based methods (based on the original Transmission Disequilibrium Test) known to be immune from population stratification problems. Then to increase sample size for linkage and association analyses, data from the two villages were used together.

<p><a>The hemotrophic mycoplasma<i>, Mycoplasma ovis</i>, is found in sheep and goats throughout the world. This pathogenic bacterium is capable of causing an acute, life-threatening infection as well as chronic or subclinical infections in these animals. The purposes of the present studies were to develop <i>M. ovis</i>-specific assays for detection of this hemoplasma, and to better understand infection dynamics within pregnant ewes and lambs. </a>The first study describes the development and validation of a SYBR^® Green quantitative PCR (qPCR) assay, which was subsequently used to determine the prevalence of <i>M. ovis</i> infection within a population of goats and to evaluate risk factors for infection. This highly sensitive and specific assay consistently detected as few as 10 copies of plasmid/reaction. Convenience-based sampling of 362 goats from 61 farms located in Indiana revealed a prevalence of infection of 18% (95% confidence interval (CI), 14% to 22%). Bacterial loads of <i>M. ovis</i> ranged from 1.05 x 10³ to 1.85 x 10⁵copies/mL of blood with a mean of 1.31 x 10⁴copies/mL of blood. The only risk factor associated with hemoplasma infection was the production use of the goat; dairy goats had a 3.3 fold increase compared with the prevalence in goats used for meat. This study not only demonstrates that <i>M. ovis</i> infection is common in goats in Indiana, but shows the variability of bacterial loads that can be found in chronically-infected animals. While sub-clinically infected goats may have a bacteremia, levels are characteristically less than 2.0 x 10⁵copies/mL.</p><p> The second project utilized a combination of cross-sectional and longitudinal studies to estimate the prevalence of <i>M. ovis</i> infection from a cohort of naturally-infected pregnant ewes, assess changes in their bacterial loads, and determine the incidence of <i>M. ovis</i> in lambs pre- and post-weaning. The prevalence of <i>M. ovis</i> infection in ewes was not found to be significantly different during pregnancy, and before and after weaning of the lambs, with prevalence estimates of 45% (95% CI, 23.1 – 68.5), 36% (95% CI, 17.9 – 57.4), and 44%, (95% CI, 24.4 – 65.1), respectively. Bacterial loads of the ewes from the cross-sectional study ranged from 10⁴to 10⁹copies/mL of blood, with the median bacterial load at 10⁵ copies/mL of blood. While higher bacterial loads are typical of an acute infection, none of the ewes in this study had overt clinical signs. The data suggest that <i>M. ovis</i> loads may be higher in pregnant sheep, particularly in ewes half-way through pregnancy. Most of the <i>M. ovis</i> infections in the study lambs were detected post-weaning which suggests that transplacental or transmammary infection of <i>M. ovis</i> are unlikely routes.</p><p> In the third study, a subset of <i>M. ovis</i> genes for use in a multi-locus sequence typing assay (MLST) were evaluated. Next-generation sequencing was performed to generate data from pooled DNA amplicons in order to identify single nucleotide polymorphisms (SNPs) of <i>M. ovis </i>from five genes. Evaluation of the quality and depths of coverage for the reads and SNPs indicated that the pooled DNA amplicons produced reads and SNPs having high quality and sufficient depth. This pooling technique is a cost-effective alternative to whole-genome sequencing. While the MLST has good discriminatory power and may be used to identify genetically distant and divergent clusters of <i>M. ovis</i> from different geographical origins, within a herd the discrimination power is low, which may hamper its usefulness in transmission studies. </p><p> The fourth and final study was the development of a loop-mediated isothermal amplification (LAMP) assay targeting the dnaK gene of <i>M. ovis</i>, with comparison of the assay to conventional PCR (cPCR). The metal ion indicator hydroxynaphthol blue (HNB) was added prior to the reaction, which allowed for visual detection of LAMP-positive samples as indicated by a color change from violet to sky blue. <i>Mycoplasma ovis</i> was consistently detected in 45 minutes with the LAMP assay at a reaction temperature of 64°C, with more infected sheep being detected than by cPCR. Therefore, the LAMP assay is fast and reliable in the detection of <i>M. ovis</i>. The developed LAMP assay may have applications in diagnostics, surveillance and disease management as well as prevalence studies. However, a more robust molecular technique is necessary for <i>M. ovis</i> isolate or stain discrimination to investigate transmission or disease spread in an outbreak.</p><p> </p><p> In conclusion, three new molecular tools for the detection of <i>M. ovis</i> in goats and sheep were developed as results of these studies. We have shown that the qPCR assay is an efficient tool for detection and quantification of <i>M. ovis</i> loads in blood from both of these species. On the other hand, the value of the LAMP assay is for reliable detection of infection (not quantification), especially in resource-limited situations. The five-locus MLST protocol developed herein, a typing assay based on the polymorphism of five gene sequences, is a laborious technique requiring DNA extraction, PCR amplification, purification and sequencing of target loci. The value of this technique is not as a routine diagnostic, but rather it may be used to better understand the genetic diversity of <i>M. ovis</i> and investigate strain variations. Most importantly, the scheme is sufficiently robust to allow direct genotyping of <i>M. ovis</i> in total blood DNA extracts without culture isolation. The MLST approach may prove useful as a tool for future investigations of transmission and disease spread. These studies have also expanded our understanding of the infection dynamics of <i>M. ovis</i> in pregnant sheep and lambs. It is shown herein that despite the high prevalence and sometimes high bacterial loads in pregnant ewes, <i>M. ovis</i> does not appear to be transmitted to the lambs in utero or during the perinatal period. The lambs become infected mostly after weaning; this may suggest a protective effect during the pre-weaning period and/or subsequent exposure/infection from their environment. </p><br>

Orientador: Suzete Aparecida Lanza Destéfano / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-20T14:37:32Z (GMT). No. of bitstreams: 1 Tonin_MarianaFerreira_D.pdf: 2371085 bytes, checksum: d7c915d664efeec24397b855164c65b9 (MD5) Previous issue date: 2012 / Resumo: Espécies pertencentes ao gênero Xanthomonas são responsáveis por doenças que podem causar grandes perdas econômicas em diversas culturas. Esses fitopatógenos têm sido objeto de diversos estudos taxonômicos resultando em significativas alterações na classificação em nível inter e infraespecífico. O presente estudo teve por objetivo caracterizar taxonomicamente bactérias do gênero envolvendo: (1) diferenciação e análises filogenéticas de espécies do gênero Xanthomonas; (2) esclarecimento da posição taxonômica de linhagens classificadas como Xanthomonas sp.; (3) diferenciação de patovares da espécie X. campestris; (4) desenvolvimento de primers específicos para X. campestris, X. translucens, X. cucurbitae e X. melonis. O par de primers rpoB2F/rpoB3R, desenhado a partir de sequências do gene rpoB, foi empregado em experimentos de amplificas;ao utilizando-se DNAs de 26 espécies do gênero Xanthomonas e os produtos de amplificas;ao (800 pb) foram digeridos com diversas endonucleases. Os perfis de restrição obtidos com a utilização da enzima Hae III permitiram a diferenciação da maioria das espécies, incluindo os patógenos de mesmo hospedeiro como X. albilineans e X. sacchari (patogênicas à cana-de-açúcar); X. cucurbitae . e X. melonis (patogênicas ao melão); X. vesicatoria, X. gardneri e X. euvesicatoria/X. perforans (patogênicas ao tomateiro). Ainda, os produtos de amplificação do gene rpoB foram seqüenciados e a árvore filogenética construída a partir destas sequências também possibilitou a diferenciação das espécies do gênero, indicando que o gene rpoB pode ser considerado um marcador molecular eficiente para o estudo das relações filogenéticas de espécies do gênero Xanthomonas. Os DNAs de linhagens classificadas como Xanthomonas sp. também foram analisados por meio de experimentos de amplificação com primers específicos para a espécie X. axonopodis, de hibridizas;ao DNA-DNA, e analise de multilocus utilizando-se sequências dos genes rpoB, rpoA, atpA, recA e região espaçadora 16S-23S DNAr. Os resultados mostraram que as linhagens de X. sp. pv. viticola, X. sp. pv. betae e X. sp. pv. paulliniae pertencem a espécie X. axonopodis, entretanto, não foi possível definir a posição taxonômica das linhagens de X. sp. pv. arracaciae, X. sp. pv. esculenti e X. sp. pv. eucalypti, embora várias ferramentas tenham sido utilizadas. Assim, somente estudos complementares deverão ser conduzidos visando esclarecer a classificação dessas linhagens. Nesse estudo também foram analisadas as espécies X. campestris, X. translucens, X. cucurbitae e X. melonis. Visando a diferenciação dos patovares da espécie X. campestris, os DNAs destas linhagens foram submetidos a experimentos de PCR-RFLP do gene rpoB e as digestões duplas utilizando-se Cfo I/Mbo I. Os resultados permitiram a diferenciação dos patovares X.c. pv. raphani, X.c. pv. barbariae, X.c. pv. incanae, X.c. pv. armoraciae e X.c. pv. campestris/ X.c. pv. aberrans. Ainda, os dados obtidos nas amilises do gene rpoB permitiram o desenvolvimento de primers específicos para as espécies X. campestris, patogênicas as crucíferas (rpoB2F/xcamR) e X. translucens, patogênicas a gramíneas e cereais (trans1F/trans2R). Alem disso, amilises de sequências da região espaçadora 16S-23S DNAr possibilitaram o desenho do par de primers mecF/mecR, especifico para X. cucurbitae e X. melonis, espécies patogênicas ao melão. A especificidade destes primers foi confirmada em experimentos de amplificação utilizando-se DNAs de algumas bactérias de diferentes gêneros isoladas de mesma espécie de plantas hospedeiras. O nível de sensibilidade da técnica de PCR utilizando-se os primers desenvolvidos foi de 0,1 pg para rpoB/xcam, de 0,01 ng para trans1F/trans2R e de 1 pg para mecF/mecR / Abstract: Xanthomonas species are responsible for diseases causing economic losses in many crops. These phytopathogens have been subject of several taxonomic studies resulting in significant changes at interespecific or infraspecific level. This study aimed the taxonomic characterization of Xanthomonas species including: (1) differentiation and phylogenetic analysis of Xanthomonas species; (2) clarification of taxonomic position of strains classified as Xanthomonas sp.; (3) differentiation of the pathovars of X. campestris; ( 4) development of specific primers for X. campestris, X. translucens, X. cucurbitae and X. melonis. The rpoB2F/rpoB3R primers, designed from rpoB gene sequences, were employed in amplification experiments using DNAs from 26 species of Xanthomonas genus and the products (800 bp) were digested with different restriction enzymes. Profiles using Hae III allowed to differentiate the most species of the genus, including pathogens the affect the same plant host as X. albilineans e X. sacchari (pathogenic to sugarcane); X. cucurbitae e X. melonis (pathogenic to melon); X. vesicatoria, X. gardneri and X. euvesicatoria/X. perforans (pathogenic to tomato). Amplification products of the rpoB gene were sequenced and the phylpgenetic tree constructed from these sequences also allowed the differentiation of the Xanthomonas species, indicating that the rpoB gene can be used as an efficient molecular marker for phylogenetic relationships studies within the Xanthomonas genus. DNA of the strains classified as Xanthomonas sp. were also analysed by the amplification experiments using specific primers for X. axonopodis species, DNA-DNA hybridization and multilocus sequence analysis of the genes rpoB, rpoA, atpA, recA and intergenic spacer region 16S-23S rDNA. Results showed that the strains of X. sp. pv. viticola, X. sp. pv. betae and X. sp. pv. paulliniae belong to X. axonopodis species, however, it was not possible to define the taxonomic position of X. sp. pv. arracaciae, X. sp. pv. esculenti and X. sp. pv. eucalypti, although different approaches have been used. Thus, further studies should be conducted in order to clarify their classification. In this study X. campestris, X. translucens, X. cucurbitae and X. melonis were also analyzed. In order to differentiate the pathovars of the X. campestris specie, the DNAs of these strains were submitted to PCR-RFLP analysis of the rpoB gene and the double digestions using Cfo I/Mbo I allowed the differentiation of the pathovars X. c. pv. raphani, X.c. pv. barbariae, X.c. pv. incanae, X.c. pv. armoraciae and X.c. pv. campestris/ X.c. pv. aberrans. Also, the data obtained in the rpoB gene analysis allowed the development of specific primers for the species X. campestris, pathogenic to crucifers (rpoB2F/xcamR) and X. translucens, pathogenic to grasses and cereals (trans1F/trans2R). Furthermore, intergenic spacer 16S-23S rDNA sequences analysis enabled the development of the mecF/mecR primers, specific for X cucurbitae and X melonis, both pathogenic to melon. The specificity of these primers was confirmed by amplification experiments using DNAs from bacteria belonging to different genera pathogenic to the same plant hosts. Sensitivity level of the PCR technique using the primers developed was 0,1 pg for rpoB/xcam, 0,01 ng for trans1F/trans2R and 1 pg for mecF/mecR / Doutorado / Genetica de Microorganismos / Doutor em Genetica e Biologia Molecular

The increasing number of Extended Spectrum Beta-Lactamase (ESBL) producing Escherichia coli (E. coli) associated with sepsis in China is the reason for designing the current study. During 2014-2016, thirty hospitals representing 10 different provinces in China was involved in collecting E. coli isolates causing blood stream infections. Early treatment with suitable antibiotics have been found to be of lifesaving importance in the case of care for septic patients. Thorough understanding of the pathogens involved is therefore crucial. Using antimicrobial susceptibility testing, PCR and Multi Locus Sequence Typing (MLST), the molecular characteristics of ESBL producing E. coli isolates could be determined. This study can report that the most common ESBL producing genes found were CTX-M-14 (51 isolates, 45,5%), CTX-M-55 (23 isolates, 20,5%) CTX-M-15 (22 isolates, 19,6%). In addition, 2 isolates (1,8%) were found to be SHV-11 positive which is another ESBL producing gene. As a side finding, 5 isolates harbored Metallo-beta-lactamase (MBL) encoding genes such as NDM-5 and NDM-1 which were found to coexist with CTX-M-55 and CTX-M-14 respectively. An MLST analysis resulted in the finding of 25 different and 17 previously unknown (16,2 %) sequence types. The most common sequence types were ST131 (18 isolates, 17,1 %) as reported previously.  No significant differences in antimicrobial susceptibility were identified whether ESBL producing genes such as SHV and CTX-M was present or not. This study indicates that there could be novel resistance mechanisms present among those isolates not encoding the genes of interest. However, this finding requires further research before it can be confirmed.

No description available.

Les études longitudinales sur une longue période permettent d’échantillonner plusieurs fois le phénomène étudié et ainsi, avec des mesures répétées, dégager une tendance confirmée. Mais, dès lors, elles produisent de très larges bases de données épidémiologiques accompagnées de plus de sources de bruit par rapport aux études à observation unique ; et souvent, contiennent de la corrélation dans les mesures. Ici, nous avons présenté à travers cette thèse une étude de long terme des facteurs épidémiologiques et génétiques du paludisme menée dans deux cohortes familiales du Sénégal, l’une dans le village de Dielmo suivi pendant 19 années consécutives (1990 – 2008) et l’autre dans le village de Ndiop suivi pendant 16 années consécutives (1993 – 2008). L’objectif de ce travail de thèse a été de développer des méthodes d’analyse statistique pour identifier des gênes de susceptibilité / résistance au paludisme prenant en compte les relations familiales, les mesures répétées et des potentielles interactions génotypes – environnement dans l’évaluation des phénotypes. Par la suite, de tels phénotypes corrigés des facteurs identifiés comme potentielles sources de confusion et/ou de bruit ont été alors utilisés pour les tests de liaison et d’association génétique. Le phénotype principal étudié chez chaque volontaire a été la survenue ou non d’accès palustre, attribué à une infection au parasite Plasmodium falciparum, durant chaque trimestre de présence (PFA). Les études ont été menées de manière indépendante dans chacun des deux villages, de même que les analyses descriptives, l’estimation de la contribution génétique humaine et des effets individuels. Les tests de liaison et d’association génétique ont été réalisés par des méthodes familiales basées sur l’analyse de la transmission d’allèles des parents aux enfants (Transmission Disequilibrium Test). Ces méthodes sont connues pour être robustes par rapport au problème de la stratification de population et donc nous permettent d’augmenter la taille de notre échantillon dans les études de liaison et d’association génétique en analysant les deux villages en même temps. / Long term longitudinal surveys have the advantage to enable several sampling of the studied phenomena and then, with the repeated measures obtained, find a confirmed tendency. However, these long term surveys generate large epidemiological datasets including more sources of noise than normal datasets (e.g. one single measure per observation unit) and potential correlation in the measured values. Here, we studied data from a long-term epidemiological and genetic survey of malaria disease in two family-based cohorts in Senegal, followed for 19 years (1990–2008) in Dielmo and for 16 years (1993–2008) in Ndiop. The main objectives of this work were to take into account familial relationships, repeated measures as well as effect of covariates to measure both environmental and host genetic (heritability) impacts on the outcome of infection with the malaria parasite Plasmodium falciparum, and then use findings from such analyses for linkage and association studies. The outcome of interest was the occurrence of a P. falciparum malaria attack during each trimester (PFA). The two villages were studied independently; epidemiological analyses, estimation of heritability and individual effects were then performed in each village separately. Linkage and association analyses used family-based methods (based on the original Transmission Disequilibrium Test) known to be immune from population stratification problems. Then to increase sample size for linkage and association analyses, data from the two villages were used together.

Dans la plupart des questions écologiques, les phénomènes aléatoires d'intérêt sont spatialement structurés et issus de l'effet combiné de multiples variables aléatoires, observées ou non, et inter-agissant à diverses échelles. En pratique, dès lors que les données de terrain ne peuvent être directement traitées avec des structures spatiales standards, les observations sont généralement considérées indépendantes. Par ailleurs, les modèles utilisés sont souvent basés sur des hypothèses simplificatrices trop fortes par rapport à la complexité des phénomènes étudiés. Dans ce travail, la démarche de modélisation hiérarchique est combinée à certains outils de la statistique spatiale afin de construire des structures aléatoires fonctionnelles "sur-mesure" permettant de représenter des phénomènes spatiaux complexes en écologie des populations. L'inférence de ces différents modèles est menée dans le cadre bayésien avec des algorithmes MCMC. Dans un premier temps, un modèle hiérarchique spatial (Geneclust) est développé pour identifier des populations génétiquement homogènes quand la diversité génétique varie continûment dans l'espace. Un champ de Markov caché, qui modélise la structure spatiale de la diversité génétique, est couplé à un modèle bivarié d'occurrence de génotypes permettant de tenir compte de l'existence d'unions consanguines chez certaines populations naturelles. Dans un deuxième temps, un processus de Poisson composé particulier,appelé loi des fuites, est présenté sous l'angle de vue hiérarchique pour décrire le processus d'échantillonnage d'organismes vivants. Il permet de traiter le délicat problème de données continues présentant une forte proportion de zéros et issues d'échantillonnages à efforts variables. Ce modèle est également couplé à différents modèles sur grille (spatiaux, régionalisés) afin d'introduire des dépendances spatiales entre unités géographiques voisines puis, à un champ géostatistique bivarié construit par convolution sur grille discrète afin de modéliser la répartition spatiale conjointe de deux espèces. Les capacités d'ajustement et de prédiction des différents modèles hiérarchiques proposés sont comparées aux modèles traditionnellement utilisés à partir de simulations et de jeux de données réelles (ours bruns de Suède, invertébrés épibenthiques du Golfe-du-Saint-Laurent (Canada)).

Methicillin-resistant Staphylococcus aureus (MRSA) is an important human pathogen that is endemic in hospitals all over the world. It has more recently emerged as a serious threat to the general public in the form of community-acquired MRSA. MRSA has been implicated in a wide variety of diseases, ranging from skin infections and food poisoning to more severe and potentially fatal conditions, including; endocarditis, septicaemia and necrotising pneumonia. Treatment of MRSA disease is complicated and can be unsuccessful due to the bacterium's remarkable ability to develop antibiotic resistance. The considerable economic and public health burden imposed by MRSA has fuelled attempts by researchers to understand the evolution of virulent and antibiotic resistant strains and thereby improve epidemiological management strategies. Central to MRSA transmission management strategies is the implementation of active surveillance programs, via which unique genetic fingerprints, or genotypes, of each strain can be identified. Despite numerous advances in MRSA genotyping methodology, there remains a need for a rapid, reproducible, cost-effective method that is capable of producing a high level of genotype discrimination, whilst being suitable for high throughput use. Consequently, the fundamental aim of this thesis was to develop a novel MRSA genotyping strategy incorporating these benefits. This thesis explored the possibility that the development of more efficient genotyping strategies could be achieved through careful identification, and then simple interrogation, of multiple, unlinked DNA loci that exhibit progressively increasing mutation rates. The baseline component of the MRSA genotyping strategy described in this thesis is the allele-specific real-time PCR interrogation of slowly evolving core single nucleotide polymorphisms (SNPs). The genotyping SNP set was identified previously from the Multi-locus sequence typing (MLST) sequence database using an in-house software package named Minimum SNPs. As discussed in Chapter Three, the genotyping utility of the SNP set was validated on 107 diverse Australian MRSA isolates, which were largely clustered into groups of related strains as defined by MLST. To increase the resolution of the SNP genotyping method, a selection of binary virulence genes and antimicrobial resistance plasmids were tested that were successful at sub typing the SNP groups. A comprehensive MRSA genotyping strategy requires characterisation of the clonal background as well as interrogation of the hypervariable Staphylococcal Cassette Chromosome mec (SCCmec) that carries the β-lactam resistance gene, mecA. SCCmec genotyping defines the MRSA lineages; however, current SCCmec genotyping methods have struggled to handle the increasing number of SCCmec elements resulting from a recent explosion of comparative genomic analyses. Chapter Four of this thesis collates the known SCCmec binary marker diversity and demonstrates the ability of Minimum SNPs to identify systematically a minimal set of binary markers capable of generating maximum genotyping resolution. A number of binary targets were identified that indeed permit high resolution genotyping of the SCCmec element. Furthermore, the SCCmec genotyping targets are amenable for combinatorial use with the MLST genotyping SNPs and therefore are suitable as the second component of the MRSA genotyping strategy. To increase genotyping resolution of the slowly evolving MLST SNPs and the SCCmec binary markers, the analysis of a hypervariable repeat region was required. Sequence analysis of the Staphylococcal protein A (spa) repeat region has been conducted frequently with great success. Chapter Five describes the characterisation of the tandem repeats in the spa gene using real-time PCR and high resolution melting (HRM) analysis. Since the melting rate and precise point of dissociation of double stranded DNA is dependent on the size and sequence of the PCR amplicon, the HRM method was used successfully to identify 20 of 22 spa sequence types, without the need for DNA sequencing. The accumulation of comparative genomic information has allowed the systematic identification of key MRSA genomic polymorphisms to genotype MRSA efficiently. If implemented in its entirety, the strategy described in this thesis would produce efficient and deep-rooted genotypes. For example, an unknown MRSA isolate would be positioned within the MLST defined population structure, categorised based on its SCCmec lineage, then subtyped based on the polymorphic spa repeat region. Overall, by combining the genotyping methods described here, an integrated and novel MRSA genotyping strategy results that is efficacious for both long and short term investigations. Furthermore, an additional benefit is that each component can be performed easily and cost-effectively on a standard real-time PCR platform.

Ehrlichia ruminantium est l'agent causal de la cowdriose, une maladie tropicale mortelle des ruminantstransmis par les tiques Amblyomma. Jusqu'à présent, il n'existe pas de vaccin efficace dû à la faible protection croisée des souches vaccinales vis-à-vis des isolats de terrain. Ceci est principalement lié àdiversité génétique d'E. ruminantium au sein les zones géographiques. Par conséquent, la caractérisation de lastructure génétique de la population d'E. ruminantium à l'échelle mondiale et régionale est importante pour définir les meilleures stratégies de contrôle et améliorer les stratégies de surveillance de la cowdriose. / Ehrlichia ruminantium is the causal agent of heartwater, a ruminant tropical fatal diseasetransmitted by Amblyomma ticks. Up to now, no effective vaccine is available due to a limitedcross protection of vaccinal strains on field isolates mainly associated to a high geneticdiversity of E. ruminantium within geographical locations. Thus, both characterization of E.ruminantium genetic population structure at worldwide and regional scale and estimation of E.ruminantium tick prevalence are important to delimitate better control strategies and improveheartwater monitoring strategies