Plants contain organelles called chloroplasts, which is where photosynthesis takes place. Chloroplasts also contain their own DNA, which is separate from the DNA in the nucleus. This DNA does not change much over evolutionary time, so it can be used to investigate relationships between organisms. Here we created a tool that makes it easier to analyze this chloroplast DNA as well making it easier to share complete chloroplast genomes on public databases. In addition, we also found a mobile element in the chloroplast DNA of a group of ferns, which appears to be driving structural changes in their genomes.

Thesis advisor: Gabor T. Marth / Structural variations (SVs), like single nucleotide polymorphisms (SNPs) and short insertion-deletion polymorphisms (INDELs), are a ubiquitous feature of genomic sequences and are major contributors to human genetic diversity and disease. Due to technical difficulties, i.e. the high data-acquisition cost and/or low detection resolution of previous genome-scanning technologies, this source of genetic variation has not been well studied until the completion of the Human Genome Project and the emergence of next-generation sequencing (NGS) technologies. The assembly of the human genome and economical high-throughput sequencing technologies enable the development of numerous new SV detection algorithms with unprecedented accuracy, sensitivity and precision. Although a number of SV detection programs have been developed for various SV types, such as copy number variations, deletions, tandem duplications, inversions and translocations, some types of SVs, e.g. copy number variations (CNVs) in capture sequencing data and mobile element insertions (MEIs) have undergone limited study. This is a result of the lack of suitable statistical models and computational approaches, e.g. efficient mapping method to handle multiple aligned reads from mobile element (ME) sequences. The focus of my dissertation was to identify and characterize CNVs in capture sequencing data and MEI from large-scale whole-genome sequencing data. This was achieved by building sophisticated statistical models and developing efficient algorithms and analysis methods for NGS data. In Chapter 2, I present a novel algorithm that uses the read depth (RD) signal to detect CNVs in deep-coverage exon capture sequencing data that are originally designed for SNPs discovery. We were one of the early pioneers to tackle this problem. In Chapter 3, I present a fast, convenient and memory-efficient program, Tangram, that integrates read-pair (RP) and split-read (SR) signals to detect and genotype MEI events. Based on the results from both simulated and experimental data, Tangram has superior sensitivity, specificity, breakpoint resolution and genotyping accuracy, when compared to other recently published MEI detection methods. Lastly, Chapter 4 summarizes my work for SV detection in human genomes during my PhD study and describes the future direction of genetic variant researches. / Thesis (PhD) — Boston College, 2013. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Biology.

Thesis advisor: Gabor T. Marth / During the last few years, numerous new sequencing technologies have emerged that require tools that can process large amounts of read data quickly and accurately. Regardless of the downstream methods used, reference-guided aligners are at the heart of all next-generation analysis studies. I have developed a general reference-guided aligner, MOSAIK, to support all current sequencing technologies (Roche 454, Illumina, Applied Biosystems SOLiD, Helicos, and Sanger capillary). The calibrated alignment qualities calculated by MOSAIK allow the user to fine-tune the alignment accuracy for a given study. MOSAIK is a highly configurable and easy-to-use suite of alignment tools that is used in hundreds of labs worldwide. MOSAIK is an integral part of our genetic variant discovery pipeline. From SNP and short-INDEL discovery to structural variation discovery, alignment accuracy is an essential requirement and enables our downstream analyses to provide accurate calls. In this thesis, I present three major studies that were formative during the development of MOSAIK and our analysis pipeline. In addition, I present a novel algorithm that identifies mobile element insertions (non-LTR retrotransposons) in the human genome using split-read alignments in MOSAIK. This algorithm has a low false discovery rate (4.4 %) and enabled our group to be the first to determine the number of mobile elements that differentially occur between any two individuals. / Thesis (PhD) — Boston College, 2010. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Biology.

Les bactéries productrices de β-lactamase à spectre élargi (BLSE) ont été rapportées dans plusieurs pays, mais aucune information nest disponible sur les différents types de BLSE produits par les Entérobactéries au Cameroun. On distingue plus de 290 types de β-lactamases divisés en 4 classes (A, B, C, et D) et jusquaujourdhui il nexiste pas des inhibiteurs capables dinhibés toutes les classes de β-lactamase. Un total de 267 souches dEntérobactéries a été isolé entre 1995-1998 (259 souches) et 2002 (8 souches) des produits pathologiques (urines, pus et sang) des patients hospitalisés dans divers services de lHôpital Central de Yaoundé. Létude de la sensibilité in vitro de ces bactéries à 12 antibiotiques (amoxicilline, amoxicilline/acide clavulanique, pipéracilline, imipénème, céfazoline, céfoxitine, céfotaxime, ceftazidime, aztréonam, gentamicine, ofloxacine et triméthoprime/sulfaméthoxazole) par la méthode des disques a permis de noter un certains nombre de faits : limipénème, lofloxacine et la ceftazidime sont les antibiotiques les plus actifs sur les différentes Entérobactéries avec respectivement 100%, 91% et 88% de bactéries sensibles ; les différentes souches présentent un niveau de résistance élevé à lamoxicilline (90%), à la pipéracilline (79%), à la céfazoline (75%) et au triméthoprime/sulfaméthoxazole (74%). En vue de détecter les espèces bactériennes productrices de BLSE, 69 souches résistantes à au moins une céphalosporine de 3ème génération ou au monobactame ont été soumises au test de double synergie. Trente-huit (31 entre 1995-1998 ; 7 en 2002) souches ont montré un test positif. La prévalence des souches productrices de BLSE est de 12% (31/259). Ces souches présentant un phénotype BLSE sont observées chez toutes les espèces dEntérobactéries étudiées avec une prévalence de 18,8% (Klebsiella spp.), de 17,6% (Citrobacter spp.), de 14,3% (E. coli) et de 1,8% (Proteus spp.). Parmi les 38 souches potentiellement productrices de BLSE, 18 (47,4%) ont transféré le gène de résistance des oxyimino-céphalosporines aux souches E. coli HK 225, K12 et DH5α. Seize de ces gènes ont été transférés par conjugaison et deux par transformation. Le transfert de gène de BLSE a été co-transféré avec le gène de résistance à la gentamicine et/ou au triméthoprime/sulfaméthoxazole. Toutes les souches transconjugantes et transformantes ont présenté un phénotype BLSE et ont été toutes sensibles à la céfoxitine et à limipénème. Les extraits enzymatiques bruts obtenus par sonication des souches transconjugantes, transformantes et cliniques ont été capable de réduire les diamètres dinhibition autour des disques de pénicillines, céphalosporines 1ère à la 3ème génération ou monobactame en utilisant une souche dE. coli complètement sensible aux antibiotiques, mais nont pas eu deffet sur ces zones autour des disques dimipénème, de céfoxitine et damoxicilline/acide clavulanique. La détermination des points isoélectriques (pIs) des différentes souches après transfert du gène de résistance par la technique disoélectrofocalisation a montré que les souches transconjugantes et transformantes produisent des β-lactamases dont les pIs varient de 5,4 à 8,8. En effet, 12 souches produisent les BLSE de pI 8,2 ; deux souches en plus de cette BLSE produisent deux autres enzymes de pIs 5,4 et 8,5. Les souches de phénotypes CTX-M produisent soit une enzyme (pI 8,4 ; une souche) soit deux (pIs 7,3, 8,8 ; 2 souches) ou alors 3 types (pIs 5,4 ; 7,3 et 8,8 ; une souche). Les extraits dADN de 19 souches (16 transconjugantes, 2 transformantes et 1 souche clinique) soumis à la technique de PCR, PCR/NheI, en utilisant les amorces spécifiques des gènes blaSHV, blaCTX-M, blaTEM et blaOXA ont montré que les différents extraits contiennent les gènes des b-lactamases du type SHV (sulfhydryl variable), CTX-M (cefotaximase), TEM (Temoniera) et OXA (aoxacillinase). Lanalyse de la séquence des différents produits de PCR a permis de montrer que : 14 souches produisent la BLSE SHV-12, cinq la BLSE CTX-M (CTX-M-1, une souche ; CTX-M-15, 4 souches) ; quatre souches produisent en association avec les BLSE les β-lactamases TEM-1 (SHV-12, 2 souches ; CTX-M-15, 2 souches) OXA-30 (CTX-M-15, 4 souches). Lanalyse des profils de digestion des plasmides portant les gènes blaCTX-M par les enzymes de restriction, dhybridation et des profils génotypiques des souches cliniques productrice de ces enzymes par ERIC-PCR a montré que la dissémination de la BLSE CTX-M peut se faire soit par transfert de plasmide (on retrouve le même plasmide chez deux souches différentes, E. coli YC-14 et K. pneumoniae YC-17) soit par la même souche dun patient à un autre (E. coli YC-5 = E. coli YC-2). Létude de lenvironnement génétique de ce gène (blaCTX-M-15) par lanalyse des séquences a montré que CTX-M-15 est codée par deux plasmides multirésistants différents, parmi lesquels un (pYC-5b) porte lélément ISEcp1-blaCTX-M-15 flanqué par un site de réplication constituée de 5 paires de bases et inséré dans le transposon Tn2. Une forme tronquée de cet élément a été identifiée chez lautre plasmide (pYC-14). Les BLSE SHV-12, CTX-M-1 et CTX-M-15 sont décrites pour la première fois au Cameroun. Dans le but de rechercher de nouveaux inhibiteurs des β-lactamases plus actifs, nous avons étudié deux plantes médicinales, Garcinia lucida (Clusiaceae) et Bridelia micrantha (Euphorbiceae) en vue dévaluer leur activité anti- β-lactamase. Les concentrations inhibant 50 % de l'activité de l'enzyme (CI50) des extraits de G. lucida et B. micrantha sont respectivement de 0,01 mg/ml (β-lactamase P99) et de 0,019 mg/ml (β-lactamase OXA-10) indiquant ainsi une bonne inhibition des β-lactamases par les extraits de ces plantes. Le produit 4 issu de la purification par chromatographie haute performance de lextrait de G. lucida est très actif sur la β-lactamase P99 (CI50 = 0,038 mg/ml) alors que les produits 2' et 3' provenant de lextrait de B. micrantha sont actifs sur lenzyme OXA-10 (CI50 de 0,09 mg/ml et 0,11 mg/ml respectivement). La détermination de la structure des composés actifs de ces deux plantes pourrait être une voie importante dans le développement des inhibiteurs des β-lactamases. / Bacteria producing extended-spectrum β-lactamase (ESBL) have been reported in many countries, but there is no information on different type of ESBL-producing enterobacteria in Cameroon. More than 290 β-lactamases have been described and divided into four classes (A, B, C and D) and up to now, there is no good inhibitor which can inhibite all classes of β-lactamases. A total of 267 strains of Enterobacteriaceaae were isolated between 1995-1998 (259 isolates) and 2002 (eight isolates) from pathological products (urines, pus and blood) of patients at the Yaounde Central Hospital. The susceptibility of the isolates to 12 antibiotics (amoxicillin, amoxicillin/clavulanate, piperacillin, imipenem, cefazolin, cefoxitin, cefotaxime, ceftazidime, aztreonam, gentamicin, ofloxacin and trimethoprim/sulfamethoxazole) was determined using the agar diffusion disk method. Imipenem, ofloxacin and ceftazidime were the most active antibiotics against overall enterobacteria with susceptibility rates of 100%, 91% and 88% respectively. High resistance rates were observed to amoxicillin (90%), piperacillin (79%), cefazolin (75%) and trimethoprim/sulfamethoxazole (73%). Enterobacteria isolates (69) resistant to oxyimino-cephalosporin or monobactam were screened for ESBL production by the double disk (DD) synergy test. Thirty-eight (31 from 1995-1998; seven in 2002) were proved positive to the DD synergy test, suggesting the presence of ESBL. The prevalence of ESBL was 12% (31/259) and ESBL-producing strains were Klebsiella spp. (18.8%), Citrobacter spp. (17.6%), Escherichia coli (14.3%) and Proteus spp. (1.8%). Of the 38 ESBL-producing strains, 18 (47.4%) were transferred resistance to oxyimino-cephalosporins to E. coli HK 225, K12 and DH5α. Sixteen of these strains were transferred ESBL gene by conjugation and two by transformation. Resistance to gentamicin and/or trimethoprim/sulfamethoxazole was co-transferred. All transconjugant and transformant strains exhibited ESBL phenotype but remained susceptible to cefoxitin and imipenem. Crude extracts of β-lactamase-producing transconjugants, transformants and clinical strains were able to reduce the diameters of inhibition zones around disks containing penicillins, 1st to 3rd generation cephalosporins or monobactam when tested against a fully susceptible E. coli strains, but had no effect on such zones around cefoxitin-, imipenem- and amoxicillin/clavulanate disks. The determination of isoelectric points (pIs) of the various strains after transfer of resistance determinant showed that transconjugants and transformants strains produced β-lactamases with pIs 5.4 to 8.8. In fact, 12 strains produced β-lactamase with pI 8.2, 2 strains in addition to the above enzyme produced two others enzymes with pIs 5.4 and 8.5. Strains with CTX-M- phenotype produced enzyme (pI 8.4, 1 strain), two (pIs 7.3, 8.8; 2 strains) or three types pIs (5.4, 7.3, 8.8; 1 strain). PCR, PCR/NheI was performed using total or plasmid DNA from 19 strains as templates, and the primers specific to blaSHV, blaCTX-M, blaTEM and blaOXA. Direct sequencing of PCR products revealed that: 14 strains produced SHV-12 ESBL, 5 CTX-M- ESBL (CTX-M-1, one strain; CTX-M-15, 4 strains); four other strains shared non-ESBL TEM-1 (2 producing-SHV-12 strains and 2 producing-CTX-M-strains) and OXA-30 (4 producing CTX-M-15 strains). Analysis of restriction patterns of plasmid carrying blaCTX-M gene, hybridization, and genotyping patterns of CTX-M-clinical strains by ERIC-PCR showed that dissemination of blaCTX-M gene could be done by plasmid transfer (same plasmid in different strains, E. coli YC-14 and K. pneumoniae YC-17) or by strains from patient to patient (same strain, E. coli YC-2=E. coli YC-5). Sequence analysis of genetic environment of blaCTX-M-15 revealed that CTXM-15 was encoded by two different multiresistant plasmids, of which one (pYC-5b) carried an ISEcp1-blaCTX-m-15 element flanked by five bp target site duplication and inserted within a Tn2-derived sequence. A truncated form of this element in the second plasmid (pYC-14) was identified. ESBLs SHV-12, CTX-M-1, CTX-M-15 are described for the first time in Cameroon. In our effort to find new active β-lactamase inhibitors, we investigated two medicinal plants, Garcinia lucida (Clusiaceae) and Bridelia micrantha (Euphorbiaceae) for anti- β-lactamase activity. The extracts from G. lucida and B. micrantha exhibited good inhibition of β-lactamase P99 (IC50, 0.01 mg/ml) and OXA-10 (IC50, 0.02 mg/ml) respectively. Purified products from these plants by high performance liquid chromatography showed that, product 4 from G. lucida is very active on β-lactamase P99 (IC50, 0.03 mg/ml) whereas products 2 and 3 from B. micrantha are active on OXA-10 (IC50, 0.09 mg/ml; 0.11 mg/ml respectively). The structural elucidation of the active constituents of these plants will provide useful leads in the development of β-lactamase inhibitors.