This study characterized Salmonella enterica serovar Heidelberg from human, abattoir poultry and retail poultry isolates to examine the molecular relationships of cefoxitin resistance between these groups. A total of 147 S. Heidelberg (70 cefoxitin-resistant and 77 cefoxitin-susceptible) isolates were studied. All cefoxitin-resistant isolates were also resistant to amoxicillin-clavulanic acid, ampicillin, ceftiofur and ceftriaxone, and all contained the CMY-2 gene. Pulsed-field gel electrophoresis typing illustrated that 93.9% isolates clustered together with ≥ 90% similarity. Core genome analysis using whole genome sequencing identified 12 clusters of isolates with zero to four single nucleotide variations. These clusters consisted of cefoxitin-resistant and susceptible human, abattoir poultry and retail poultry isolates. Analysis of CMY-2 plasmids from cefoxitin-resistant isolates revealed all belonged to incompatibility group I1. Analysis of plasmid sequences using WGS revealed high identity (95-99%) to a previously described plasmid (pCVM29188_101) found in Salmonella Kentucky. When compared to pCVM29188_101, all sequenced cefoxitin-resistant isolates were found to carry one of ten possible variant plasmids. The discovery of several clusters of isolates from different sources with zero to four SNVs suggests that transmission between human, abattoir poultry and retail poultry sources may be occurring. The classification of newly sequenced plasmids into one of ten sequence variant types suggests transmission of a common CMY-2 plasmid amongst S. Heidelberg with variable genetic backgrounds. / October 2016

Norovirus is the commonest cause of viral gastroenteritis, affecting all age groups worldwide. Outbreaks frequently occur in semi-closed communities such as schools, cruise ships, prisons and hospitals. Within the healthcare environment, the economic and logistical burdens and the inconvenience caused by norovirus is significant, since ward closure remains central to infection control. The aim of this study was to investigate norovirus transmission dynamics during hospital outbreaks. The ultimate goal was to provide information that could, in future, lead to the development of novel, less disruptive approaches to curtailing the spread of infection. The study explored the application of 'next generation' high throughput DNA sequencing technologies to the determination of large numbers of norovirus genomes. Whole genome sequences provide the highest possible level of discrimination among viruses, information which is essential to the identification of linked and independent cases of infection. The approach exploits the high norovirus mutation rate, which is typical of RNA viruses. Consequently, viruses within a single ward which differ by more than a few SNVs can be considered to represent independent introductions, rather than a single outbreak. Whole genome sequence data (determined for noroviruses collected between 2009 and 2013) were combined with epidemiological data, providing further insights into transmission dynamics. These data identified multiple independent virus introductions during single ward outbreaks. The possible origin of such outbreaks in Oxfordshire hospitals were investigated using viruses originating in the local community, and in other healthcare environments distributed throughout the UK. Whole genome sequences of noroviruses from consecutive years were genetically divergent, confirming the rapid evolution of the virus over time and excluding the possibility of prolonged environmental contamination as a reservoir of infection. Such detailed information on norovirus transmission within the healthcare environment could inform alternative future approaches to optimising infection control within the healthcare setting.

Thesis advisor: Gabor T. Marth / To help improve the analysis of forward genetic screens, we have developed an efficient and automated pipeline for mutational profiling using our reference guided tools including MOSAIK and FREEBAYES. Studies using next generation sequencing technologies currently employ either reference guided alignment or de novo assembly to analyze the massive amount of short read data produced by second generation sequencing technologies; the far more common approach being reference guided alignment due to the massive computational and sequencing costs associated with de novo assembly. The success of reference guided alignment is dependent on three factors; the accuracy of the reference, the ability of the mapper to correctly place a read, and the degree to which a variant allele differs from the reference. Reference assemblies are not perfect and none are entirely complete. Moreover, read mappers can only map reads in genomic locations that are unique enough to confidently place reads; paralogous sections, such as related gene families, cannot be characterized and are often ignored. Further, variant alleles that drastically alter the subject's DNA, such as insertions or deletions (INDELs), will not map to the reference and are either entirely missed or require further downstream analysis to characterize. Most importantly, reference guided methods are restricted to organisms for which such reference genomes have been assembled. The current alternative, de novo assembly of a genome, is prohibitively expensive for most labs requiring deep read coverage from numerous different library preparations as well as massive computing power. To address the shortcomings of current methods, while eliminating the costs intrinsic to de novo sequence assembly, we developed RUFUS, a novel, completely reference-independent variant discovery tool. RUFUS directly compares raw sequence data from two or more samples and identifies groups of reads unique to one or the other sample. RUFUS has at least the same variant detection sensitivity as mapping methods, with greatly increased specificity for SNPs and INDEL variation events. RUFUS is also capable of extremely sensitive copy number detection, without any restriction on event length. By modeling the underlying k-mer distribution, RUFUS produces a specific copy number spectrum for each individual sample. Applying a Bayesian detection method to detect changes in k-mer content between two samples, RUFUS produces copy number calls that are equally as sensitive as traditional copy number detection methods with far fewer false positives. Our data suggest that RUFUS' reference-free approach to variant discovery is able to substantially improve upon existing variant detection methods: reducing reference biases, reducing false positive variants, and detecting copy number variants with excellent sensitivity and specificity. / Thesis (PhD) — Boston College, 2014. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Biology.

For over two decades, the study of genetics has been making significant progress towards understanding the causes of common disease. Across a wide range of complex disorders there have been hundreds of associated loci identified, largely driven by common genetic variation. Now, with the advent of next-generation sequencing technology, we are able to interrogate rare and low frequency variation in a high throughput manner for the first time. This provides an exciting opportunity to investigate the role of rarer variation in complex disease risk on a genome-wide scale, potentially o↵ering novel insights into the biological mechanisms underlying disease pathogenesis. In this thesis I will assess the potential of this technology to further our understanding of the genetics of complex disease, using inflammatory bowel disease (IBD) as an example. After first reviewing the history of genetic studies into IBD, I will describe the analytical challenges that can occur when using sequencing to perform case-control association testing at scale, and the methods that can be used to overcome these. I then test for novel IBD associations in a low coverage whole genome sequencing dataset, and uncover a significant burden of rare, damaging missense variation in the gene NOD2, as well as a more general burden of such variation amongst known inflammatory bowel disease risk genes. Through imputation into both new and existing genotyped cohorts, I also describe the discovery of 26 novel IBD-associated loci, including a low frequency missense variant in ADCY7 that approximately doubles the risk of ulcerative colitis. I resolve biological associations underlying several of these novel associations, including a number of signals associated with monocyte-specific changes in integrin gene expression following immune stimulation. These results reveal important insights into the genetic architecture of inflammatory bowel disease, and suggest that a combination of continued array-based genome- wide association studies, imputed using substantial new reference panels, and large scale deep sequencing projects will be required in order to fully understand the genetic basis of complex diseases like IBD.

Le botulisme est une maladie nerveuse, commune à l’homme et aux animaux, due à l’action de la toxine botulique produite par Clostridium botulinum. Il existe 8 types de toxines dénommées A à H. Les bactéries capables de produire cette toxine se différencient en six groupe sur la base de leurs caractéristiques phénotypiques et biologiques. Les souches de C. botulinum responsables du botulisme humain appartiennent aux groupes I et II selon qu’elles soient protéolytiques ou non. Elles produisent les toxines A, B, E et F, ainsi que le nouveau type H récemment découvert. C. butyricum et C. baratii sont également capables de produire les toxines botuliques de type F et E et appartiennent au groupe V et VI. C. argentinense appartient au groupe IV et est capable de synthétiser la toxine de type G. Elle a été soupçonnée d’être impliquée dans des cas de botulisme infantile en Argentine. Les souches de C. botulinum responsables du botulisme animal appartiennent au groupe III (C. novyi sensu lato) et produisent les toxines C, D et leurs formes mosaïques C/D et D/C. La toxine botulique est le poison le plus puissant connu à ce jour. La dose létale nécessaire pour tuer une personne en bonne santé par intoxication alimentaire est de 70 µg seulement. C’est pourquoi cette toxine a fait l’objet d’études particulièrement approfondies, notamment celles impliquées dans des cas de botulisme humain. Elle peut également être utilisée pour le traitement de certaine pathologie ou la chirurgie esthétique (Botox). Malheureusement, elle peut également être utilisée à mauvais escient, en tant qu’arme de guerre ou à des fins de bioterrorisme. C’est pourquoi l’emploi de la toxine botulique ou de sa bactérie productrice fait l’objet d’une législation particulièrement stricte. Mon projet de doctorat s’est organisé autour de plusieurs projets de recherche visant à développer des méthodes de détection et de typage de du germe et de sa toxine (projets Européens BIOTRACER et AniBioThreat ; projets NRBC-bio ; LNR botulisme aviaire en France). Lors de mes recherches j’ai concentré mon travail sur le développement de méthodes capable de suivre et remonter à la source d’une contamination, qu’elle soit délibérée, accidentelle ou naturelle. Afin d’y parvenir j’ai investigué les gènes des flagellines de C. botulinum groupe I à III, responsables du botulisme humain et animal. L’analyse des gènes flaA et flaB a mis en évidence 5 groupes majeurs et 15 sous-groupes, certain étant spécifiques de régions géographiques. FlaB s’est montré spécifique de C. botulinum type E. Les gènes flagellines fliC, spécifiques à C. botulinum du groupe III, se divisent 5 groupes, avec fliC-I et fliC-IV associés aux types mosaïques C/D et D/C. J’ai étudié la prévalence des souches productrices de toxine de type mosaïques chez les volailles et les bovins. Les résultats montrent que les types C/D et D/C sont majoritaires en Europe. Enfin, j’ai séquencé 17 génomes provenant de souches responsables de botulisme animal en France (14 types C/D et 3 types D/C). Leur analyse montre que ces souches sont très proche génétiquement, entre elles et avec les souches Européennes. Grâce à ces données j’ai mis en évidence un large contenu extra chromosomique dans les souches C/D, qui peut être utilisé pour créer une carte d’identité génétique. D’autre part, l’étude des séquences Crisps à des fins de typage ne s’est pas avérée suffisamment résolutive, du fait de système Crispr-Cas déficient chez les souches C/D. Enfin, un très haut degré de discrimination a été atteint par typage SNP, qui a permis de distinguer jusqu’à l’origine de chaque souche. L’ensemble de ces résultats est développé dans le présent manuscrit / Clostridium botulinum is the etiologic agent of botulism, a deadly paralytic disease that can affects both human and animals. Different bacteria, producing neurotoxins type A to H, are responsible for the disease. They are separated into different groups (I to VI) on the basis of their phenotypical and biological characteristics. Human botulism is mainly due to Groups I and II producing neurotoxins A, B, E and F, with type H recently discovered. Also C. butyricum and C. baratii species (Groups V and VI), producing toxins type F and E respectively, are scarcely reported. C. argentinense Group IV, producing toxin type G, which has been suspected to be associated with infant botulism in Argentina. Animal botulism is mainly due to Group III, which is constituted by C. novyi sensu lato species. They produce toxin types C, D and their mosaic variants. Botulinum neurotoxins are the most powerful toxin known to date with as little as 70 µg enough to kill a person by food poisoning. Therefore, it received a great deal of attention. Botulinum neurotoxins have been deeply studied, especially human related toxins compared to animal. The toxins found to be useful for medical or cosmetic (Botox) treatments, but it was also used as a biological warfare agent, and for bioterrorism. Its extreme potency is equal to its dangerousness. Therefore, governments show concerns of its potential misuse as a bioterrorism weapon; research programs are funded to study and raise awareness about both the toxins and the producing organisms. My PhD work was structured by the different projects I was involved in, which were related to C. botulinum detection and typing, like BIOTRACER and AniBioThreat European projects, the French national CBRN program, or the NRL for avian botulism. The main transversal objective I followed lead me to develop new methods to trace back the origin of C. botulinum contamination, in case of a deliberate, accidental or naturally occurring botulism outbreak. I investigated flagellin genes as potential genetic targets for typing C. botulinum Group I-II and III, responsible for human and animal botulism respectively. Flagellin genes flaA and flaB showed the investigated C. botulinum Group I and II strains to cluster into 5 major groups and up to 15 subgroups, some being specific for certain geographical areas, and flaB being specific to C. botulinum type E. Flagellin fliC gene investigated in C. botulinum Group III showed to cluster into five groups, with fliC-I and fliC-IV associated to type C/D and D/C respectively, being not discriminative enough to differentiate highly genetically related strains. I also studied the prevalence of mosaic toxin genes in C. botulinum Group III in animal botulism, mainly in poultry and bovine. The results brought out the mosaic toxin types C/D and D/C to be predominant in the samples investigated throughout Europe. Finally, I explored the full genome sequences of 14 types C/D and 3 types D/C C. botulinum Group III strains, mainly originating from French avian and bovine botulism outbreaks. Analyses of their genome sequences showed them to be closely related to other European strains from Group III. While studying their genetic content, I was able to point out that the extrachromosomal elements of strains type C/D could be used to generate a genetic ID card. Investigation of Crispr typing method showed to be irrelevant for type C/D, due to a deficient Crispr-Cas mechanism, but deserve more investigation for type D/C. The highest level of discrimination was achieved while using SNP core phylogeny, which allowed distinguishing up to the strain level. Here are the results I’m going to develop in this manuscript

Background: Infections caused by carbapenem-resistant Enterobacteriaceae (CRE) such as Escherichia coli and Klebsiella pneumoniae are among the most urgent threats of the infectious disease realm. The incidence of these infections has been increasing over the years and due to very limited treatment options, mortality is estimated at about 50%. By 2050, mortality from antimicrobial resistant infections is expected to surpass cancer at 10 million deaths annually. Methods: We evaluated 18 contemporary antimicrobials against 122 carbapenem-resistant Enterobacteriaceae using a variety of antimicrobial susceptibility testing methods according to Clinical Laboratory Standards Institute guidelines. Time-kill studies were performed on clinical isolates with variable resistance to meropenem, amikacin, and polymyxin B. Phenotypic expression assays were performed on all isolates and whole genome sequencing was performed on 8 isolates to characterize molecular resistance mechanisms. Pharmacodynamic modeling of meropenem and polymyxin B was also conducted. Results: CRE were primarily K. pneumoniae, and Enterobacter spp. 60% expressed Klebsiella pneumoniae carbapenemase (KPC) only, 16% expressed Verona Integron-encoded Metallo-beta-lactamase (VIM) only, 5% expressed KPC and VIM, and 20% expressed other mechanisms of resistance. Antimicrobial susceptibility testing indicated the most active antimicrobials against CRE were ceftazidime/avibactam, imipenem/relebactam, amikacin, tigecycline, and the polymyxins. Etest® strips did not reliably measure polymyxin B resistance. The automated testing system, BD Phoenix™, consistently reported lower MICs than the gold standard broth microdilution. Time-kill studies showed regrowth at clinically achievable concentrations of meropenem alone (4, 16, and 64 mg/L), polymyxin B alone (0.25 and 1 mg/L), or amikacin alone (8 and 16 mg/L), but combinations of meropenem with either polymyxin B or amikacin were bactericidal and synergistic. Meropenem administered simultaneously or prior to polymyxin B exhibited superior activity to polymyxin B administered first. Conclusions: Novel carbapenemase-inhibitor combinations (ceftazidime/avibactam and imipenem/relebactam) exhibit the best activity against KPC-producing CRE. The polymyxins, amikacin, and tigecycline exhibit the best activity against VIM-producing CRE. Meropenem in combination with polymyxin B is bactericidal and synergistic when the meropenem MIC is ≤32 mg/L, and meropenem should never be administered after polymyxin B. Meropenem and amikacin is bactericidal and synergistic when the amikacin MIC is ≤16 mg/L. Etest® strips should not be used for characterizing polymyxin B or colistin activity. Clinicians should be aware that automated testing systems may produce biased susceptibility results relative to the gold standard method, broth microdilution, which may influence interpretation of in vitro results.

Genome-based detection methods provide the most conclusive means for establishing the presence of microbial species. A prime example of their use is in the detection of bacterial species, many of which are naturally vital or dangerous to human health, or can be genetically engineered to be so. However, current genomic detection methods are cost-prohibitive and inevitably use unique sensors that are specific to each species to be detected. In this thesis we advocate the use of combinatorial and non-specific identifiers for detection, made possible by exploiting the sparsity inherent in the species detection problem in a clinical or environmental sample. By modifying the sensor design process, we have developed new molecular biology tools with advantages that were not possible in their previous incarnations. Chief among these advantages are a universal species detection platform, the ability to discover unknown species, and the elimination of PCR, an expensive and laborious amplification step prerequisite in every molecular biology detection technique. Finally, we introduce a sparsity-based model for analyzing the millions of raw sequencing reads generated during whole genome sequencing for species detection, and achieve significant reductions in computational speed and high accuracy.

Mycobacterium tuberculosis (Mtb) poses a global health catastrophe that has been compounded by the emergence of highly drug resistant Mtb strains. We used whole genome sequencing (WGS) to directly compare the accumulation of mutations in Mtb isolated from cynomolgus macaques with active, latent and early reactivation disease. Based on the distribution of single nucleotide polymorphisms (SNPs) observed, we calculated the mutation rates for these disease states. Our data suggest that during latency, Mtb acquires a similar number of chromosomal mutations as would be expected to emerge in a logarithmically growing culture over the same period of time despite reduced bacterial replication during latent infection. The pattern of polymorphisms suggests that the mutational burden in vivo is due to oxidative DNA damage. We next sought to determine why some strains of Mtb are preferentially associated with high-level drug resistance. We demonstrate that Mtb strains from the East Asian lineage acquire drug resistances in vitro more quickly than Mtb strains from the Euro-American lineage. Their higher drug resistance rate in vitro reflects a higher basal mutation. Moreover, the in vitro mutation rate correlates well with the bacterial mutation rate in humans as determined by whole genome sequencing of clinical isolates. Finally, using an agent-based model, we show that the observed differences in mutation rate predict a significantly higher probability of multi-drug resistance in patients infected with East Asian lineage strains of Mtb. Lastly, we sought to determine the mechanisms Mtb uses to proofread nascently polymerized DNA. Through fluctuation analysis of deletion mutants of two potential \(polIII\epsilon\) homologs, we demonstrate that neither is responsible for the maintenance of DNA replication fidelity. To explore the possibility that one of these homologs, Rv3711c, participates in an unknown redundant pathway, we used transposon capture and sequence (TraCS) to identify genes conditionally essential in an Rv3711c deletion mutant. Our analysis suggests that while Rv3711c does not participate in proofreading, it may act in an alternative novel DNA repair pathway. Taken together, our fluctuation analysis and TraCS data suggest that mycobacteria do not use canonical methods of proofreading to maintain genomic fidelity.

The pathogen Campylobacter fetus (CF) is classified into two subspecies, Campylobacter fetus subspecies fetus (CFF) and Campylobacter fetus subspecies venerealis (CFV). Even though CFF and CFV are genetically closely related, they exhibit differences in their host adaptation; CFF inhabits the gastrointestinal tract of both humans and several animal species, while classical CFV is specific to the bovine genital tract and is of particular concern with respect to international bovine trade regulation. Traditionally, differentiation between the two subspecies has been achieved using a limited number of biochemical tests but more rapid and definitive genetic methods of discrimination are desired. A recent study suggested that the presence of a genomic island only in CFV could discriminate between the two sub- species but this hypothesis could not be confirmed on a collection of isolates originating in Canada. To identify alternative gene targets that would support accurate subspecies discrimination, this study has applied several approaches including suppression subtractive hybridization and whole genome sequencing supplemented with optical mapping. A subtractive hybridization screen, using a well-characterized CFV isolate recovered during routine screening of bulls in an Artificial Insemination center in western Canada and that lacked much of the genomic island and a typical Canadian CFF isolate, yielded 50 clones; characterization of these clones by hybridization screening against selected CF isolates and by nucleotide sequence BLAST analysis identified three potentially CFV-specific clones that contained inserts originating from a second genomic island. Further screening using a larger CF sample set found that only Clone #35 was truly CFV-specific. Optical maps (NcoI digest) of the Canadian CFF and CFV isolates used for the subtractive hybridization showed that certain regions of these genomes were quite distinct from those of two reference strains. Whole genome sequencing of these two isolates identified two target genes (PICFV5_ORF548 and CFF_Feature #3) that appear to be selectively retained in the two subspecies. Screening of a collection of CF isolates by PCRs targeting these three loci (SSH_Clone #35, PICFV5_ORF548 and CFF_Feature #3) supported their use for subspecies discrimination. This work demonstrates the complex genomic diversity associated with these CF subtypes and the challenge posed by their discrimination using limited genetic loci.

Introduction: In recent years, the average age of first reproduction has risen significantly, the mean now standing at around 30 years in many countries. The adverse effects of maternal age on fertility and reproduction have been well documented. However, the influence of paternal age on fertility, reproduction and postnatal health is relatively poorly understood, and 50% of all male infertility cases are classed as idiopathic or unexplained infertility. Methods: The aim of this study was to investigate factors that contribute to male infertility, split into two main parts. The first part focused on analysing data collected from patients who had undergone fertility treatment to assess the influence of different factors on infertility, especially at the genome level. The second part attempted to deal with some of the technical challenges of screening and diagnostic methods to study the genome, with the aim of providing tools that would assist future studies in pinpointing genetic factors responsible for infertility, especially in cases of idiopathic infertility. Results: Based on data from the first part of the study, it was determined that advanced paternal age can affect sperm progressive motility, sperm DNA integrity and the fertilisation rate of in vitro fertilisation (IVF) cycles, as well as the development of embryos. Direct analysis of sperm DNA fragmentation (SDF) and degradation levels revealed an association between elevated SDF and impaired embryo development. Furthermore, a correlation was shown between chromosome aneuploidy and variance in SDF and sperm DNA degradation. Moreover, aneuploidy can influence abnormal sperm morphology and consequently also progressive motility. Also, embryo development rate of IVF cycles on day three, demonstrated a significant decline in cycles where the sperm used for fertilisation had a high aneuploidy rate, which can highlight the reduced developmental capacity of aneuploid embryos. From the lifestyle factors assessed, only alcohol consumption significantly correlated with the sperm DNA damage. Therefore, poor semen quality may highlight damage that has been incurred by the sperm DNA. When the semen quality is suboptimal, the intracytoplasmic sperm injection (ICSI) technique is suggested as a standard strategy to improve the prognosis of ART. However, when the progressive motility is poor, the ICSI approach is not as effective. Based on our findings and in line with other studies, the only sperm parameter that can be affected by paternal age is sperm motility, which could be an indicator of SDF. Therefore, the decline in ICSI fertilisation rate in patients with impaired sperm progressive motility could be due to sperm DNA damage, and even ICSI cannot improve the fertilisation rate considerably. Discussion: The aim of the second part of this project was to establish a robust workflow for whole- genome amplification (WGA) and whole-genome sequencing of single cells to improve the coverage rate and fidelity, with the aim of providing means of detecting any mutation in the genome that might be responsible for reduced embryonic developmental competence. Towards this end, the efficiencies of two different WGA protocols (REPLI-g and TruePrime) were compared. Multiple technical factors required optimisation in order to create a suitable protocol. Our results demonstrated the overall superiority of REPLI-g compared to TruePrime in almost all the assessed parameters. The amplification rate of REPLI-g was much faster than that of TruePrime, and prolonged incubation led to overamplification and an increased duplication rate. However, the TruePrime method has a slower amplification rate and therefore, by increasing the incubation time, it was possible to improve the quality of the data. The modified protocol with reduced volume also had the most promising outcome in terms of the data produced, and could fulfil our expectations by being fast, cost-effective and efficient. Conclusion: In conclusion, the results from the first part of this study confirmed the negative impact of male age on assisted reproductive treatments, which can result in decreased success rates of fertilisation. Other factors such as sperm DNA damage may also contribute to this age effect, suggesting that assessing this parameter prior to fertility treatment, and attempting to mitigate elevated levels of sperm DNA damage, may be of value to older patients. Additionally, overcoming the technical challenges in studying genetic contributory factors in infertility is a promising step toward better understanding of the mutations and variations that are involved in this phenomenon.