While antibiotics have saved the lives of millions of people since the discovery of the first β-lactam, penicillin, their continued effectiveness is being increasingly threatened by resistant bacteria. Bacterial resistance to β-lactams is mainly achieved through the production of serine-β-lactamases (SBLs) and metallo-β-lactamases (MBLs). Although both types of β-lactamases are commonly isolated in clinical settings, MBLs represent the greatest threat to public health since they are resistant to SBL inhibitors and most β-lactams. However, aspergillomarasmine A (AMA), a fungal natural product synthesized by Aspergillus versicolor, was shown to be a rapid and potent inhibitor against two clinically relevant MBLs: NDM-1 and VIM-2. In bacteria possessing these enzymes, AMA could rescue the activity of meropenem, a broad-spectrum β-lactam that is usually reserved for the treatment of the most severe bacterial infections. However, many questions remain revolving around AMA's inhibitory potency and spectrum. Therefore, the activity of AMA in combination with six β-lactams from three subclasses (carbapenem, penam, cephem) was explored against 19 MBLs from three subclasses (B1, B2, B3). After determining that AMA activity was linked to MBL zinc affinity and that AMA was more potent when paired with a carbapenem, the efficacy of an AMA/meropenem combination was evaluated with and without avibactam, a potent SBL inhibitor. This study used ten Escherichia coli and ten Klebsiella pneumoniae laboratory strains as well as 30 clinical strains producing at least one MBL and one SBL. Once establishing that the AMA/avibactam/meropenem combination was effective against carbapenemase-producing Enterobacterales, new Acinetobacter and Pseudomonas shuttle vectors were created. With these shuttle vectors, it was determined that the AMA/avibactam/meropenem combination was effective against some of the bacteria topping the World Health Organization’s priority pathogen list. / Thesis / Doctor of Philosophy (PhD) / Bacteria are all around us. While some bacteria can promote human health, others can cause serious infections. These infections are typically treated with antibiotics. β-Lactam antibiotics, such as penicillins and cephalosporins, are especially important to medicine. Unfortunately, an increasing number of bacteria employ enzymes, known as β-lactamases, which negate the effects of β-lactam antibiotics. Previous studies demonstrated that a natural product, known as aspergillomarasmine A (AMA), could inhibit some β-lactamase enzymes. Consequently, the inhibitory power of AMA was further explored against a larger number of β-lactamase enzymes and in combination with different β-lactam antibiotics. After discovering that AMA had more inhibitory power when combined with a β-lactam antibiotic known as meropenem, the efficacy of the AMA/meropenem pairing was evaluated against resistant bacteria in the presence and absence of avibactam, another β-lactamase inhibitor. The AMA/avibactam/meropenem combination was shown to be effective against some of the world’s most antibiotic-resistant bacteria.

<p>Antibiotic resistance is often linked to human use of antibiotics. However, antibiotics and antibiotic biosynthetic pathways have been evolving for millions of years suggesting that antibiotic resistance is an ancient phenomenon. As of now, there has been no systematic survey of environmental microbes proven to exist in the absence of human influence and Lechuguilla cave offers such environment<em>. </em> Resistance diversity in strains isolated from this cave was analyzed by a phenotypic screen against a panel of 26 different antibiotics. Resistant strains were further investigated through determination of minimal inhibitory concentration (MIC) and inactivation studies. Of particular interest was strain LC044 (<em>Brachybacterium paraconglomeratum</em>), observed to inactivate macrolide antibiotics by phosphorylation. Genome sequencing and bioinformatics (BLAST analysis) identified a putative macrolide phosphotransferase (MPH) in strain LC044 and biochemical characterization of the purified recombinant protein confirmed its macrolide inactivating properties. To investigate if characterized MPH was unique to cave isolate, available terrestrial <em>Brachybacterium faecium</em> DSM 4810 genome was mined for presence of MPH-like protein. The top hit to the MPH from LC044 (a protein with 282 amino acids and 72% identity) was heterologously expressed and purified. Complete biochemical analysis of this enzyme revealed (i) MPH-activity, despite its annotation as aminoglycoside phosphotransferase (APH), and (ii) no significant differences in substrate specificities or kinetic parameters between these two enzymes suggesting that these two enzymes were equally effective resistance enzymes. This work highlights the prevalence of antibiotic resistance in a pristine, cave ecosystem and provides further support for the theory that antibiotic resistance is everywhere. Furthermore, the <em>mph</em> resistance determinant found in cave isolate and closely related terrestrial isolate show homology to clinical<em> mph</em> genes, suggesting that environmental <em>mph</em> genes could have served as reservoir of clinical determinants.</p> / Master of Science (MSc)

Enterococcus faecalis is a Gram-positive bacterial species that is typically a member of the human gastrointestinal tract microbiota. However, E. faecalis is also a nosocomial pathogen, which is involved in urinary tract infections, soft tissue infections and endocarditis. In recent times, the occurrence of antibiotic resistance has complicated the treatment of these infections. One of the major differences between commensal and pathogenic strains of E. faecalis is that pathogens contain multiple mobile elements such as plasmids, transposons and integrative conjugative elements (ICE). These elements allow for the acquisition and transfer of virulence factors and resistance genes. Conjugative plasmids are a class of plasmids present in E. faecalis whose transfer to host cells is induced by a small pheromone peptide, cCF10 (LVTLVFV). This peptide is initially encoded as a 22-amino acid precursor (pre-cCF10) from the signal sequence of the chromosomal ccfA gene and is then proteolytically cleaved by signal peptidase II and Eep. Once pCF10 has been transferred a host E. faecalis cell, it is exceptionally stable. A replicon clone is maintained in greater than 85% of host cells over 100 generations in the absence of selection, suggesting the stability of pCF10 is intrinsic to the replicon. Three unique features of the replication initiation protein PrgW may contribute to this stability: (a) the interaction of PrgW with pre-cCF10, (b) disulfide bond formation at three conserved cysteines (C78, C275, and C307) in PrgW, and (c) processing of the nascent PrgW protein. Replication initiation proteins associated with theta replicons, such as pCF10, are often self-contained units. To initiate plasmid replication, the replication initiation protein (PrgW) binds to direct repeats (oriV) in its own coding sequence (prgW). In silico analysis of PrgW suggests the existence of three distinct domains within the protein. The first 122 amino acids are homologous to a conserved domain present in related replication initiation proteins, which includes a Helix-Turn-Helix (HTH) DNA binding domain. This suggests that this domain of PrgW has a DNA-binding function and binds to the oriV site in prgW. The following 61 amino acids are not similar to any known sequence, and are encoded by the DNA sequence containing the direct repeats in the oriV site. This domain may or may not have a distinct function. The remaining sequence forms a domain that contains cysteines C275 and C307, and is also similar to no known structure. It is hypothesized that this domain is related to the stability of pCF10. C307 appears to be critical, as previous experiments indicate that its mutation alone affects plasmid stability. Secondary structure analysis of this domain revealed the presence of multiple alpha-helices that contain distinct hydrophobic domains that possibly contribute to pre-cCF10 binding and PrgW tertiary structure. The positions of the conserved cysteines within these alpha-helices may stabilize a hydrophobic binding pocket that could potentially facilitate interaction with pre-cCF10. PrgW has a predicted molecular weight of 38.6 KDa and can be detected in Western blots as a band with an apparent approximate molecular weight (mw) of 36,000. Previous data from our lab indicates that, when overexpressed in E. faecalis, four bands of PrgW are present with observed molecular weights of 40,000, 36,000, 24,000 and 18,000. Time course experiments revealed that the 40,000 mw form is converted to a 36,000 mw form independent. The 40,000 mw form is unstable (with a complete turnover in 30 minutes) while the 36,000 mw form has a half-life of greater than 4 hours. The 24,000 mw band does not have a DNA binding motif and is likely a turnover product. When the three conserved cysteines (and only cysteines) in PrgW are replaced with alanine, the 40,000 mw form is still processed to the 36,000 mw form. However, the cysteine to alanine mutants accumulate the 36,000 mw form. / Microbiology and Immunology

Mycobacterium avium isolates resistant to clarithromycin and azithromycin have been recovered from patients undergoing antibiotic therapy. Comparison of DNA fingerprints of sensitive and resistant isolates showed that resistance resulted from mutation of the original, sensitive isolate in five of seven patients. In the other two patients, the clarithromycin-resistant isolates were unrelated to the sensitive isolate, suggesting that the resistant isolate resulted from either superinfection or selection of a resistant strain from a polyclonal population. Investigation of the mechanisms of clarithromycin and azithromycin resistance in M. avium showed that high-level resistance resulted from a point mutation at position A-2058 in the 23S rRNA. Based on this finding, a rapid screen for clarithromycin-resistance in M. avium was developed based on PCR. Twenty-three clinical isolates were analyzed, seven of which were clarithromycin-resistant. The target product was amplified only in clarithromycin-resistant strains, all of which had mutations at position 2058. A polyuridylic acid (poly U)-dependent in vitro translation system from M. avium was developed to investigate the effect of antibiotics on protein synthesis. Clarithromycin was an effective inhibitor of protein synthesis in cell-free extracts of a susceptible M. avium strain, whereas a high-level resistant strain was less susceptible to clarithromycin in vitro. Mixtures of extracts from sensitive and resistant strains showed a pattern of clarithromycin inhibition similar to the resistant strain, suggesting that resistance may be dominant in partial diploids. Three M. avium strains exhibiting step-wise, intermediate resistance to azithromycin were characterized in comparison to the sensitive parent. All strains were similar in hydrophobicity, growth medium requirements, and growth response to temperature. The azithromycin-resistant strains were resistant to several unrelated agents, including ciprofloxacin, rifabutin, and ethidium bromide. Addition of carbonyl cyanide m-chlorophenylhydrazone (CCCP) did not lower minimal inhibitory concentrations (MICs) for ciprofloxacin or ethidium bromide. Cell-free extracts of the strains were as sensitive to azithromycin in vitro as the parent strain. The results rule out inactivation, efflux, and mutations in the target as resistance mechanisms, and suggest intermediate resistance may be due to altered permeability of the cell wall or membrane. / Ph. D.

The emergence of antibiotic resistance in riverine systems has become a growing issue worldwide. The use of watershed-scale models is popular with many other water quality issues but not in the case of antibiotic resistance. In this study, we introduce a watershed-scale bacteria fate and transport model to simulate antibiotic resistance in E. coli. This model was developed through amendments to an existing watershed-scale physically based hydrological model (SWAT), and the newly modified model was called SWAT-ARB. The SWAT-ARB model was employed in the receiving environment of a WWTP in the Adyar River basin, India. The SWAT-ARB model simulations of resistant fractions (resistant E. coli concentration/E. coli concentration) in stream water were analyzed by the flow levels with the application of a range of parameter values. It is concluded that the model can be used to test prevailing hypotheses and evaluate the current state of knowledge. For instance, model simulations suggest that the influx of ARB can be a primary driver of antibiotic resistance in rivers compared to ambient antibiotic concentrations. We used the SWAT-ARB model to quantify the impact of climate change on antibiotic resistance. Six climate models were used to obtain the future climates in two distinct scenarios. The model was applied to three watersheds as Adyar basin- India, Crab Creek basin- USA, and upper Viskan basin- Sweden. It was concluded that temperature increase may greatly affect the colder climates (Crab Creek and Viskan) with higher simulated resistant fractions. In case of Adyar basin, resistant fractions are alleviated in high flow conditions, while aggravated in low flow conditions. / Doctor of Philosophy / The antibiotic resistance occurs when bacteria no longer responds to antibiotics. Hence, the diseases that caused by resistant bacteria are harder to treat. These antibiotic resistant bacteria end up in our rivers because of our heavy use of antibiotics in human and animal treatments. Thus, the spread of antibiotic resistance has become a water quality issue in the rivers worldwide. Scientists generally use computer models to understand water quality issues in rivers. These computer models are important because of high cost of monitoring and their use in finding how environment works. Up to the date of this publication, there is no sophisticated enough model to simulate antibiotic resistance in rivers. Hence, we created a river basin scale model to simulate antibiotic resistance. We found that the influx of ARB can be a primary driver of antibiotic resistance in rivers compared to ambient antibiotic concentrations. The model was applied to three watersheds as Adyar basin- India, Crab Creek basin- USA, and upper Viskan basin- Sweden. It was concluded that temperature increase may greatly affect the colder climates (Crab Creek and Viskan) with higher antibiotic resistant bacteria compared to susceptible bacteria.

Antibiotic resistance is currently a global threat spanning clinical, environmental, and geopolitical research domains. The environment is increasingly recognized as a key node in the spread of antibiotic resistance genes (ARGs), which confer antibiotic resistance to bacteria. Detecting ARGs in the environment is the first step in monitoring and controlling antibiotic resistance. In recent years, next-generation sequencing of environmental samples (metagenomic sequencing data) has become a prolific tool for the field of surveillance. Metagenomic data are nucleic acid sequences, or nucleotides, of environmental samples. Metagenomic sequencing data has been used over the years to detect and analyze ARGs. An intriguing instance of ARGs is the multi-functional ARG, where one ARG encodes two or more different antibiotic resistance functions. Multi-functional ARGs provide resistance to two or more antibiotics, thus should have evolutionary advantage over ARGs with resistance to single antibiotic. However, there is no tool readily available to detect these multi-functional ARGs in metagenomic data. In this study, we develop a computational pipeline to detect multi-functional ARGs in metagenomic data. The pipeline takes raw metagenomic data as the input and generates a list of potential multi-functional ARGs. A plot for each potential multi-functional ARG is also created, showing the location of the multi-functionalities in the sequence and the sequencing coverage level. We collected samples from three different sources: influent samples of a wastewater treatment plant, hospital wastewater samples, and reclaimed water samples, ran the pipeline, and identified 19, 57, and 8 potentially bi-functional ARGs in each source, respectively. Manual inspection of the results identified three most likely bi-functional ARGs. Interestingly, one bi-functional ARG, encoding both aminoglycoside and tetracycline resistance, appeared in all three data sets, indicating its prevalence in different environments. As the amount of antibiotics keeps increasing in the environment, multi-functional ARGs might become more and more common. The pipeline will be a useful computational tool for initial screening and identification of multi-functional ARGs in metagenomic data. / Master of Science / Antibiotics are the drug to fight against the infection of bacteria. Since the first antibiotic was discovered in 1928, many antibiotic drugs have been developed. At the same time, scientists discovered many genes responsible for the resistance of antibiotic drugs. Nowadays, antibiotic resistance is a global threat. Detecting antibiotic resistance genes in the environment is the first step toward monitoring and controlling antibiotic resistance. In recent years, next-generation sequencing has been widely used to get the DNA sequence from the environment. Metagenomics analysis has been used over the years to detect and analyze ARGs. In the literature, it has been reported that a single gene could carry two parts of sequences corresponding to two different ARGs, thus conferring resistance to two different antibiotics. This fusion might have some evolutionary advantages. In this study, we developed a novel computational tool to detect multi-functional ARGs. We collected data from three sources: the treatment plant water, the hospital wastewater, and the reclaimed water, and identified 19, 57, and 8 potential bi-functional ARGs in each source, respectively. After we manually inspected the result, we found three most likely bi-functional ARGs. We also found one bi-functional ARG that appears in all three datasets. The gene is responsible for aminoglycoside and tetracycline resistance. The tool will serve as the initial screening step to detect multi-functional ARGs.

Wastewater treatment plants (WWTPs) receive a confluence of sewage containing antibiotics, antibiotic resistant bacteria, antibiotic resistance genes (ARGs), and pathogens, thus serving as key point of interest for the surveillance of antibiotic resistance (AR) dissemination. This thesis advances knowledge about the fate of AR indicators throughout treatment and reuse. The field study informs approaches for monitoring AR at a WWTP by characterizing the resistome (i.e., full profile of ARGs) and microbiome across eight sampling events via metagenomic sequencing, complemented by antibiotic data. The WWTP significantly reduced the total load of ARGs and antibiotics, although correlations between ARGs and antibiotics were generally weak. Quantitative polymerase chain reaction was applied to validate the quantitative capacity of metagenomics, whereby we found strong correlations. The influent and effluent to the WWTP were remarkably stable with time, providing further insight into the sampling frequency necessary for adequate surveillance. The laboratory study examined the effects of commonly applied disinfection processes (chlorination, chloramination, and ultraviolet irradiation [UV]) on the inactivation of antibiotic resistant pathogens and corresponding susceptible pathogens in recycled and potable water. Further, we evaluated their regrowth following disinfection by simulating distribution. Acinetobacter baumannii, an environmental opportunistic pathogen, regrew especially well following UV disinfection, although not when a disinfectant residual was present. Enterococcus faecium, a fecal pathogen, did not regrow following any disinfection process. There were no significant differences between water types. The findings of this study emphasize a need to move beyond the framework of assessing treatment efficacy based on the attenuation of fecal pathogens. / Master of Science / Wastewater treatment plants (WWTPs) have traditionally been designed and further enhanced to minimize environmental contamination caused by solid waste, fecal pathogens, nutrients (e.g., nitrogen), and organic matter. However, treatment has not been optimized to remove the contaminants of emerging concern (CECs) investigated in this thesis: antibiotic resistant bacteria (ARB), antibiotic resistance genes (ARGs), and antibiotics. WWTPs are key point of interest for local and global surveillance of antibiotic resistance as they can receive the aforementioned CECs (via human excretion or improper disposal) from various sources (e.g., residences, hospitals). Antibiotic resistant bacteria have caused 2.8 million infections and subsequently 35,000 deaths in the United States each year. Considering treated wastewater can serve as a route of exposure for humans, potential spread of antibiotic resistance by WWTPs is of high priority to mitigate from a public health perspective. In the first study utilizing a technology to assess the full complement of ARGs in a given sample, we observed that the total load of ARGs was removed by approximately 50% across wastewater treatment, on average; total antibiotic load exhibited a similar reduction. The second study demonstrated that antibiotic resistant environmental opportunistic pathogen (i.e., pathogens which take advantage of the "opportunity" to infect an immunocompromised host, especially thriving in low nutrient engineered systems), Acinetobacter baumannii, possesses the ability to regrow following disinfection in the absence of a disinfectant residual. In contrast, antibiotic resistant Enterococcus faecium, an opportunistic pathogen of fecal origin, was successfully inactivated and unable to regrow. The findings of this study emphasize a need to move beyond the framework of assessing treatment efficacy based on the attenuation of fecal pathogens.

Manure is a commonly used soil fertilizer, but there are concerns that this practice could affect the spread of antibiotic resistance genes (ARGs) from farm to fork. A microcosm-scale study evaluated the effect of prior antibiotic use (manure-based soil amendments generated from dairy and beef cattle with or without antibiotic administration), composting, and soil type on the quantity of ARGs and the microbial community composition of dairy and beef manure applied soil. ARGs were analyzed through novel metagenomic techniques and quantitative polymerase chain reaction of sul1, tet(W), and 16S rRNA gene, while the microbial community composition was determined via 16S rRNA amplicon sequencing. The results indicated that while prior antibiotic administration elevated the relative abundance of ARGs and changed the microbial community of raw manure applied soils, composting reduced this effect. However, compost applied soils still had a higher relative abundance of ARGs than the unamended soils and occasionally soil applied with raw manure of untreated cattle. Soil type may be a mediating factor as there were differences observed between the three soil types (sandy loam, silty clay loam, and silty loam) with sandy loam amended soils often having the least attenuation of ARGs. As the relative abundance of ARGs was still elevated and the microbial community composition still significantly different from the unamended soils after 120 days, these results suggest that 120 days is not a long enough waiting period between biological soil amendments and crop harvest for ARG dissipation. / Master of Science / Antibiotics are lifesaving drugs that kill infection-causing bacteria. However, bacteria are living organisms and can adapt to stresses, such as antibiotics. When antibiotics are used, not all of the targeted bacteria are necessarily killed, and populations of resistant bacteria can survive. Resistant bacteria can not only continue to grow, but can also share their resistance capabilities with other unrelated bacteria through the transfer of antibiotic resistance genes (ARGs). ARGs are segments of DNA encoding mechanisms for the bacteria to survive antibiotic attack, such as pumping antibiotics outside of the cell or strengthening the cell wall so antibiotics cannot enter. The transfer of ARGs to human pathogens is of utmost concern, as it can cause once treatable diseases to turn deadly. Antibiotics are thus a double-edged sword because they can save lives on one hand, while their overuse or misuse can undermine their effectiveness by increasing antibiotic resistance. In the U.S. and many other countries, the biggest user of antibiotics is the livestock industry. Thus, there is growing interest in possible routes by which antibiotic resistance can spread from agriculture to humans. While some previous work has been done on direct contact with animals and meat products, less attention has been paid to the potential role raw produce grown in soils fertilized with manure-based amendments. This study thus sought to determine which factors impact ARG levels in soil. Questions of interest included: What is the effect of composting raw manure prior to soil application? Does prior treatment of cattle with antibiotics matter? Does the soil type influence the levels of ARGs? Do the ARG composition and microbial community composition respond similarly to such factors? These and other questions were evaluated in a controlled environment by simulating amended field conditions in small glass jars (microcosms) containing mixtures of different soils and manure-based amendments. Three different soils were amended with one of the following manure-based amendments: raw manure from antibiotic administered cattle, composted manure of antibiotic administered cattle, raw manure from cattle not given antibiotics, composted manure of untreated cattle, and no amendment. This experimental setup was done in duplicate, one for treatments from dairy cows and one for the beef steer treatments. The experiment lasted 120 days, as that is a current standard for how long organic farmers must wait between manure application and crop harvest. Samples were taken throughout the 120-day experiment, and the quantity of targeted ARGs was determined by analyzing the DNA through qPCR, while the overall ARG profile was studied using a new tool, called metagenomics. To identify the kinds of bacteria present in the samples (microbial community composition), the 16S rRNA gene, which is a universal gene in organisms, was targeted and sequenced via amplicon sequencing. The results of these analyses indicated that administering antibiotics to cattle and then subsequently amending soil with their manure was associated with the highest levels of ARGs compared to the other treatments, but composting reduced the effect of prior antibiotic use. Depending on the ARG, composting decreased ARG levels relative to the other treatments, but in some instances, it increased ARGs compared to soils with raw manure of untreated cattle. Even after composting, there were still higher levels of ARGs in the soil than unamended soils. Different soil types did react differently to the amendments, but more research is needed. All of the treatments resulted in different changes to the microbial community composition and did not return to the unamended soil’s community structure even after 120 days. Overall, based on these results, ARGs and the microbial community do not return to the initial condition within 120 days, which is a recommended wait time between amendment and harvest, while composting and soil type appear to be mediating factors. Additional research is needed to further advance understanding of potential mitigation options and to benchmark them to defined and measureable risk endpoints.

With a decline in antibiotic effectiveness, there is a renewed interest in using bacterial-specific viruses (bacteriophages or phages) to reduce bacterial loads, alone or with antibiotics. Most phages are therapeutically unsuitable because they are “temperate” and can integrate into the host genome, protecting the host from subsequent phage infections. However, dormant phages can be awakened by stressors such as antibiotics. Here we investigated whether antibiotics can uniquely interact with temperate phages to bias the phage away from dormancy. Model E. coli temperate phage and ciprofloxacin, a DNA-damaging antibiotic, exhibit a potent synergy, resulting in bacterial eradication at sublethal antibiotic concentrations, despite poor killing by the phage alone. Mechanistically, this synergy depletes survivors by awakening dormant phages. To broaden our findings, screening in the multi-drug-resistant pathogen P. aeruginosa, we identified phages that can synergize with four antibiotic classes, despite their widely differing targets - however, these are highly phage, antibiotic, and host-specific. Interestingly, ciprofloxacin also synergized with multiple phages, even in a ciprofloxacin-resistant clinical strain, functionally re-sensitizing the bacterium to the antibiotic. While some of these interactions operated through a mechanism independent of the temperate nature of the phages, ciprofloxacin and piperacillin, a cell wall synthesis inhibitor, specifically reduced the frequency of phage dormancy events. Finally, in a Caenorhabditis elegans infection model, temperate phage-ciprofloxacin pairing increased the lifespan of drug-resistant P. aeruginosa infected worms compared to the uninfected control. Similar rescue was also observed for the phage-carrying strain treated with the antibiotic, supporting that the phage even in its dormant form can enhance antibiotic effectiveness. Overall, we show that temperate phages uniquely synergize with antibiotics at the level of biasing the phage away from dormancy. This is generalizable across phages, antibiotics, and hosts, and shows efficacy in vivo, thereby drastically expanding their therapeutic potential. / Thesis / Doctor of Philosophy (PhD) / Bacteriophages (phages) are bacterial viruses that present a promising solution to the antibiotic resistance crisis. They can kill bacteria even when antibiotics fail, alone or in combination with them. Most work to date focuses on phages that immediately lyse the bacteria. However, phages that can go dormant within the host are far more abundant but largely ignored in therapy. Once integrated, these can awaken to switch into lytic replication by external triggers that stress the bacterial host, including antibiotics. Supported by this idea, in this thesis I show that these kinds of phages can synergistically interact with antibiotics by biasing the phage away from dormancy. This phenomenon is generalizable across host, phages, and antibiotics and shows effectiveness in an animal model.

Metagenomics has become a reliable tool for the analysis of the microbial diversity and the molecular mechanisms carried out by microbial communities. By the use of next generation sequencing, metagenomic studies can generate millions of short sequencing reads that are processed by computational tools. However, with the rapid adoption of metagenomics a large amount of data has been generated. This situation requires the development of computational tools and pipelines to manage the data scalability, accessibility, and performance. In this thesis, several strategies varying from command line, web-based platforms to machine learning have been developed to address these computational challenges. Interpretation of specific information from metagenomic data is especially a challenge for environmental samples as current annotation systems only offer broad classification of microbial diversity and function. Therefore, I developed MetaStorm, a public web-service that facilitates customization of computational analysis for metagenomic data. The identification of antibiotic resistance genes (ARGs) from metagenomic data is carried out by searches against curated databases producing a high rate of false negatives. Thus, I developed DeepARG, a deep learning approach that uses the distribution of sequence alignments to predict over 30 antibiotic resistance categories with a high accuracy. Curation of ARGs is a labor intensive process where errors can be easily propagated. Thus, I developed ARGminer, a web platform dedicated to the annotation and inspection of ARGs by using crowdsourcing. Effective environmental monitoring tools should ideally capture not only ARGs, but also mobile genetic elements and indicators of co-selective forces, such as metal resistance genes. Here, I introduce NanoARG, an online computational resource that takes advantage of the long reads produced by nanopore sequencing technology to provide insights into mobility, co-selection, and pathogenicity. Sequence alignment has been one of the preferred methods for analyzing metagenomic data. However, it is slow and requires high computing resources. Therefore, I developed MetaMLP, a machine learning approach that uses a novel representation of protein sequences to perform classifications over protein functions. The method is accurate, is able to identify a larger number of hits compared to sequence alignments, and is >50 times faster than sequence alignment techniques. / Doctor of Philosophy / Antimicrobial resistance (AMR) is one of the biggest threats to human public health. It has been estimated that the number of deaths caused by AMR will surpass the ones caused by cancer on 2050. The seriousness of these projections requires urgent actions to understand and control the spread of AMR. In the last few years, metagenomics has stand out as a reliable tool for the analysis of the microbial diversity and the AMR. By the use of next generation sequencing, metagenomic studies can generate millions of short sequencing reads that are processed by computational tools. However, with the rapid adoption of metagenomics, a large amount of data has been generated. This situation requires the development of computational tools and pipelines to manage the data scalability, accessibility, and performance. In this thesis, several strategies varying from command line, web-based platforms to machine learning have been developed to address these computational challenges. In particular, by the development of computational pipelines to process metagenomics data in the cloud and distributed systems, the development of machine learning and deep learning tools to ease the computational cost of detecting antibiotic resistance genes in metagenomic data, and the integration of crowdsourcing as a way to curate and validate antibiotic resistance genes.