Introduction We know that translating new knowledge from research into change in health care delivery is not a simple process. This thesis examines this process for a new technology applied to primary health care in tropical countries: including RDTs in clinical guidelines for treating fever in children Method The thesis examines the question: ―does implementing policy of using RDTs to target treatment instead of presumptive treatment of fever result in better quality patient care under experimental conditions as well as in routine practice?‖ Three methodological approaches are used to delineate translation to change in the field. A Cochrane review of randomised trials examines effects on quality of care in a trial, where delivery conditions are usually optimal. An analysis of a dataset from an effectiveness trial from Uganda examines effects of the policy on quality of care delivered within the context of a trial through routine health services. And third, a survey of current practice assesses implementation of an RDT-based guideline when it is introduced into the health system for routine use in selected districts. Across all three components, the thesis examines implementation of the guideline. In addition, both the systematic review and the effectiveness trial measure effects of the intervention on prescribing of antimalarials and antibiotics, and clinical outcomes (primary outcomes).The effectiveness trial evaluates effects of the policy on incremental cost, and the survey of current practice also assesses adequacy of essential health systems inputs and support services. Results The systematic review showed that HWs prescribed antimalarials to as many as 40% to 80% of cases with negative RDTs under experimental conditions. Use of RDTs was associated with 29% decline in prescribing of antimalarial drugs. Prescribing of antibiotics did not change in one trial but increased by 19% in another. Data from the effectiveness trial show that HWs used RDTs and adhered to RDT results almost all the time. This reduced antimalarials usage by 60.2% (high), 48.9% (medium) and by 22.1% (low). The data show no significant change in usage of antibiotics. Both the review and the pragmatic trial detected no significant difference in clinical outcomes between RDT and clinical diagnosis arms. Data from the effectiveness trial shows that use of RDTs is associated with a cost-saving of US$ 0.50 per case of fever (24.5% decline) in low transmission setting, and a cost-saving of US$ 0.33 per case of fever (17.7% decline) in medium transmission. Use of RDTs did not lead to a significant change in cost in high transmission settings: US$ +0.02 (95% CI: US$ -0.97 to US$+1.06). Cost-savings were accrued exclusively in older children and adults. The survey found inadequate implementation of all components of the guideline in both districts. Essential supplies, equipment and in-service training were inadequate in both districts. Discussion and conclusion Antimalarial use is lower when RDTs are used to guide treatment of fever instead of presumptive treatment. This results in savings from drugs costs in older children and adults with fever in low and medium transmission areas. This research does not confirm whether or not use of RDT-based guidelines has any effects on usage of antibiotics or clinical outcomes. A case study of Uganda shows that when delivered through routine services, none of the components of an RDT-based guideline is implemented to acceptable standards. There is insufficient evidence to suggest that the policy is superior to presumptive treatment of fever in terms of clinical outcomes. However, it can save money for medicines in low and medium transmission settings if its use is restricted to older children and adults.

Access to appropriate antimalarial treatment is essential to reduce the impact of malaria - a major cause of morbidity and mortality in Malawi. This study responds to recent calls to improve access to malaria treatment, by providing insight into existing barriers and making suggestions for improvement. It is among a limited number of studies which explore multiple dimensions of access. A mixed-methods approach was taken to investigate the influence of a number of factors on access to appropriate treatment among children in the Chikhwawa district of Malawi. Two cross-sectional household surveys enabled quantification of differences in health facility attendance between two study areas, which differed in terms of distance from health facility and ethnic and cultural backgrounds. Mean costs associated with childhood fever were calculated and compared between study areas. Qualitative methodologies provided insight into caregivers’ behaviour during treatment-seeking and after receiving antimalarials. The study found that individuals engaged in a three-phased approach to treatment-seeking. During Phase 1 caregivers assessed the illness in order to establish the need for care and appropriate action. Phase 2 involved seeking care outside of the home, in most cases from a health facility. However a number of barriers to health facility attendance were identified, including: geographic location, direct and opportunity costs; women's lack of decision-making and financial autonomy; and perceptions of available care. Health facility attendance did not guarantee appropriate treatment. Adverse events, such as vomiting; lack of understanding of the correct dosing schedule; and challenges associated with administering antimalarials to children impacted on adherence. Treatment failure led to Phase 3, which frequently involved returning to the health facility. However, repeated treatment failure, fear of receiving repeated antimalarial treatment, and advice from health facility staff led caregivers to seek alternative sources of care, such as traditional healers. This thesis provides a valuable addition to our knowledge of multiple dimensions of access to appropriate malaria treatment. Caregivers experience a number of barriers during treatment-seeking and utilisation of antimalarials, and equity of access remains a challenge. The issues raised within this thesis are specific and targetable, and the findings indicate the considerable potential impact of strengthening the existing community-based treatment and referral system, community based health education, and diagnosis and treatment at health facility level.

The single mitochondrion of Plasmodium falciparum has some unusual functional features and unconventional biochemical properties. The electron transport chain (ETC) has critical roles in generating the mitochondrial membrane potential (Δψm) and driving pyrimidine biosynthesis. Enzymes of the respiratory chain became attractive targets for antimalarial drugs such as atovaquone, an inhibitor of the bc1 complex and lethal to malaria parasites. Although much progress has been made, there are a number of knowledge gaps in our understanding of the underlying biochemical mechanisms of the respiratory chain and some major components are still not completely characterised. In addition, recent studies have generated conflicting data and hypotheses with regards to mitochondrial function. This thesis, using a multidisciplinary approach, was undertaken to improve our understanding of the mitochondrion and its function and to clarify some of the confusing and contradictory data in this area. Due to high sequence divergence, two ETC complexes remain to be clearly annotated in Plasmodium species: the succinate dehydrogenase (or complex II) and ATP synthase (or complex V). A pyramid-shaped bioinformatic strategy based on structural fingerprints and motif patterns was applied to identify candidate genes encoding the membrane anchors of complex II (SdhC and SdhD) and the ATP synthase F0 sector (subunits a and b). In order to validate the genes predicted, 2D (non-)gradient BNE/SDS-Page approaches were established to separate respiratory chain complexes from solubilised P. falciparum membranes and identify their subunits via NanoLC-MS/MS analysis. Although the proteomic strategy was validated with bovine mitochondrial preparation, no positive identification of the genes of interest could be obtained with P. falciparum extract due to low-expression of target proteins and difficulties in the preparation of isolated mitochondria of high purity. For the first time, the direct activities (or ubiquinone reduction activity) of the five dehydrogenases delivering ubiquinol to the bc1 complex were compared. Due to a higher rate of turnover, the type II NADH dehydrogenase (PfNdh2) appears to be the main enzyme for feeding the electron transport chain with activity two to three times greater than the other dehydrogenases (dihydroorotate dehydrogenase, malate quinone oxidoreductase, glycerol-3-phosphate dehydrogenase). Additionally, spatiotemporal confocal imaging of parasite mitochondria revealed that loss of PfNDH2 function provoked a collapse of the mitochondrial transmembrane potential (Δψm). This observation reinforced the interest of targeting PfNdh2 as a chemotherapeutic strategy for drug development. The catalytic properties of the P. falciparum complex II were also examined using various electron donors and acceptors. Although complex II has been previously considered as a succinate dehydrogenase, results obtained indicate that complex II functions as a ubiquinol-fumarate reductase (QFR) forming succinate from fumarate in the asexual stages of P. falciparum. The capacity of menaquinone, an alternative electron carrier of anaerobic species and recently detected in malaria parasites was also evaluated to replace ubiquinone within the respiratory chain in anaerobic conditions. Data obtained demonstrated that the menaquinone pool is not involved in the ETC and only the ubiquinone pool interacts with the different enzymes. Artemisinin and its derivatives are frontline drugs employed in the treatment of uncomplicated malaria usually in the form of combination therapies. The ETC has been previously implicated in the mode of action of artemisinin and its derivatives. In this chapter, this hypothesis was tested using a single-cell imaging and enzyme-assay based approach. Data presented reveal that endoperoxide drugs provoke a rapid collapse of mitochondrial and plasma membrane potentials, both essential for parasite survival. Addition of the iron chelator desferrioxamine or the superoxide scavenger Tiron drastically reduces the depolarization highlighting the role of ferrous ions and oxidant stress in the artemisinins activation process and membrane damaging activity. Deoxyartemisinin, which lacks the endoperoxide bridge, has no effect on membrane potential indicating that this peroxide functionality is the key pharmacophore responsible for the pharmacological activity of this class of compound. Thus, the results presented, suggest that artemisinin and its derivatives act as a generator of additional reactive oxygen species that overcome the oxidative defenses of the malaria parasite and cause a widespread and rapid membrane potential depolarization leading to mitochondrial dysfunction and parasite death. Atovaquone is a bc1 inhibitor used in combination with proguanil (e.g. MalaroneTM) for the curative and prophylactic treatment of malaria. However, resistance to atovaquone associated with point mutations have been detected in the field. In this thesis, the first description of the effect of the Y268S mutation (harbored by the atovaquone-resistant field isolate TM902CB) on parasite bc1 catalytic turnover and stability has been reported. This mutation was shown to confer a 270-fold shift of the inhibitory constant (Ki) for atovaquone with a concomitant reduction in the Vmax of the bc1 complex of 40% and a 3-fold increase in the observed Km for ubiquinol. Western blotting analyses revealed a reduced iron-sulfur protein content in Y268S bc1 suggestive of a weakened interaction between this subunit and the cytochrome b. It was concluded that the reduced enzyme activity affects protein stability and should incur a fitness penalty to the parasite, features that were not fully discernable using the yeast model alone. Due to a small mitochondrial genome (mtDNA), the great majority of mitochondrial proteins, including those composing the ETC, are imported post-translationally from the cytosol into the organelle. Thus, it is essential to understand the import and processing machinery of mitochondrial proteins in malaria parasites. However, the description of the apicomplexan model including P. falciparum is fragmented across different studies and no general overview, including recent insights, has been proposed. An updated picture of the whole protein import and processing machinery in apicomplexa is presented. Novel putative components are revealed by comparison with five other apicomplexan species (Toxoplasma gondii, Cryptosporidium muris, Theileria parva, Babesia bovis and Neospora caninum). Aspects of the apicomplexan model are highly divergent from that seen in yeast, mammalians or plants.

The presence of some missing outcomes in randomized studies often complicates the estimation of measures of effect, even in well designed randomized controlled trials. The process may be complicated further when the efficacy rates are close to 0% or 100% as the standard binomial model is susceptible to model non-convergence. The main objective of this study was to compare the performance of multiple imputation (MI) and Complete Case analysis for dealing with missing binary outcomes when modeling a risk difference. Firstly, however, the binomial regression COPY method and the Cheung’s modified Ordinary Least Squares (OLS) method were examined using simulation processes for their appropriateness in risk difference modeling. It was found that the number of copies (for the COPY method) required to minimize non-convergence coincided with the number of copies that gave the most biased estimates of the true efficacy difference while increasing the number of copies made the problems of non-convergence and bias worse; using Cheung’s method, however, there was 100% convergence with unbiased estimates of effect size. Simulation methods were used to compare the performance of complete case (CC) analysis and several multiple imputation (MI) models for handling missing outcome data over a wide range of efficacy environments and missing value assumptions. When outcomes were missing at random (MAR) or completely at random (MCAR), MI analyses that included treatment group membership in the imputation calculations yielded unbiased estimates of efficacy differences. The CC method was found to be as good, and often better, than MI methods when outcomes were MAR or MCAR, with coverage close to 95% in many situations – but neither CC nor MI produced unbiased estimates of effect difference when outcomes were missing not at random (MNAR). It was concluded that CC and MI methods are equally good in terms of producing unbiased estimates of effect difference in most missing outcome situations, but applying the intention to treat principle (ITT) which requires all randomized patients to be included in the primary analysis of a RCT, MI should be adopted as the analysis method of first choice, accompanied by a secondary CC analysis for sensitivity purposes (i.e. to investigate the extent of any likely bias).

Over the past 30 years, HIV/AIDS has had a devastating impact on the lives of millions of people around the world. Despite successes in the roll out of HIV treatment programmes, HIV prevention programmes have been less successful in lowering incidence rates. Certain groups have been found to experience significantly higher prevalence than seen in the general population, with fishing communities representing one key group with significantly higher risk of HIV infection. In fishing communities, gender power relations and economic vulnerability intersect in ways that mean that individuals are particularly vulnerable to HIV infection. In the past decade, there has increasingly been a focus on structural drivers of HIV transmission such as gender power relations. Gender differences are fundamentally underpinned by power inequalities in society and can result in the subordination of women and their interests in a manner that favours men. This thesis set out to understand how gendered structural drivers shaped vulnerability to HIV in fishing communities, and to develop interventions to address these structural drivers. The research was conducted in two fishing communities in the rural district of Mangochi in Southern Malawi. The design of the study drew on social relations theory and second-wave feminism. The research methodology used qualitative and participatory methods to address four key research objectives: (1) to understand gender power relations in fishing communities in Southern Malawi; (2) to explore and document the key drivers and facilitators of participation in transactional sex in the study villages; (3) to document individual and community perceptions of HIV risk and transactional sex in the study villages; (4) to develop a HIV prevention strategy to address risk of HIV/AIDS among fishing communities in southern Lake Malawi. Key findings are that in fishing communities, transactional sex was common and took a variety of forms, ranging from gift-giving within relationships, to-sex-for fish exchanges, to sex worker encounters. Power differences between couples in transactional sexual encounters shaped individuals’ abilities to negotiate condom use. The context and motivations for transactional sex varied and were mediated by economic need and social position both of men and women. Microfinance is a tool that can alleviate poverty and potentially prevent HIV. However, in the fishing context microfinance loan repayment procedures often increased female fish traders’ vulnerability to HIV. Participants had a good understanding of HIV risk yet this did not result in the adoption of risk-reduction strategies. Building on these findings, the thesis presents suggested HIV interventions that were developed through participatory community workshops. Interventions identified included: working with men and boys to support transformation of gender roles and normative behaviour; enforcing legislation to improve the ecological environment to reduce the impact of decline fish stocks and environmental hazards on risk-taking behaviour; improving access to HIV testing and treatment services; and improving living and working conditions of men and women working in the fishing industry. In conclusion, this thesis demonstrates the urgent need to introduce structural interventions in fishing communities in Southern Malawi and provides clear recommendations for implementing potential interventions.

Malaria is the primary cause of death in Burkina Faso and affects mostly pregnant women and children under five years old. The control of Anopheles mosquitoes responsible for the transmission of the disease by universal coverage with long lasting insecticide treated bednets (LLINS) is one of the keys strategies adopted by many malaria endemic countries. Pyrethroids remain the only insecticide class approved to be used on nets. The proportion of mosquitoes resistant to this insecticide class has steadily increased since its first report in the rice fields of Vallée du Kou in 1999 and it is now rare to find any mosquitoes from this region of Burkina Faso surviving standard WHO susceptibility diagnostic dose assays. However, without data on the magnitude of resistance it is difficult to predict the impact that this may have on malaria vector control. This study assessed the strength of the pyrethroid resistance and the underlying mechanisms, before determining whether LLINs adequately kill wild resistant mosquitoes using both newly acquired LLINs and nets previously used by householders. Mosquitoes collected in the rice field of Vallée du Kou from 2011 to 2013 were tested against the four insecticides classes approved in public health. The time taken to obtain 50 % mortality (LT50) using 0.05% WHO deltamethrin papers was determined in 2011 and 2012 while CDC bottle bioassays were used to establish the concentration of deltamethrin needed to achieve 50 % morality (LC50) in 2013. The data confirmed the high prevalence of resistance to DDT and pyrethroid in Anopheles gambiae s.l from Vallée du Kou. The LT50 increased from 98 mins to 1315 mins in just one year, representing an increase > 10 fold while the LC50 recorded in 2013 exceed 1000 fold compared to laboratory susceptible Kisumu strain Anopheles gambiae s.l populations from the bioassays were screened for target site mutations in the sodium channel. The dramatic increase in the strength of resistance was not accompanied by an increase in genes frequency, as the frequencies of the 1014F and 1575Y mutations remained stable at approximately 0.8 and 0.3 respectively in the three years. Microarray was used to identify alternative candidate genes associated with the increasing level of resistance to pyrethroids. Quantitative PCR was used to demonstrate that a subset of these candidate genes, including P450s, GSTs, CCEs, aquaporin and chymotrypsin increased significantly between 2011 and 2013 and may be contributing to the increase in resistance observed recently in this study site The bio-efficacy of used and new LLINs of six different brands, distributed as part of the National Control Malaria Program (NMCP) of Burkina Faso was assessed using cone bioassays. Whilst some of the nets gave acceptable levels of mortality against the laboratory susceptible Kisumu strain, none were able to kill more than 45 % of the resistant wild mosquitoes from Vallée du Kou. These results demonstrate that the high levels of resistance in An. gambiae s.l in Burkina Faso is having a real impact on the ability of LLINs to kill mosquitoes which will have serious implications for malaria control in this region.

Background: Herpes simplex virus encephalitis (HSVE) continues to be one of the most devastating infections of the central nervous system (CNS) despite effective antiviral treatment. The pathogenesis of the disease has not been completely studied, and little is known about the molecular mechanisms underlying cellular death and tissue damage. Better understanding of the HSVE pathogenesis is required to develop better treatment and reduce of patients’ morbidity and mortality. Method: Post-mortem brain tissue from adult HSVE (n=3) and road traffic accident (RTA; n=5) cases were examined to characterise the neuropathological changes, immune activation and infiltration of the inflammatory cells, and changes in transcript abundance during herpes simplex virus type 1 (HSV-1) infection. Patients with HIV encephalitis (n=3) provided further control tissue. Results: There was extensive neuropathological change and widespread necrosis in temporal and frontal regions of the brain among the HSVE cases. CNS cells showed characteristic signs of lytic damage. Infection was associated with microglial activation and infiltration by macrophages and lymphocytes. Genome-wide gene-expression micro-array analysis of the brain tissue demonstrated 287 host transcripts with significantly lower abundance in HSVE compared to RTA cases. Mitochondrial DNA encoded transcripts were significantly over-represented in the set of low abundant transcripts (p<0.0001), with 28 out of 33 mitochondrial genes represented on the array exhibiting lower abundance. Reflecting the array findings, immune staining for cytochrome c oxidase subunit 1, a mitochondrial encoded protein, was reduced in HSVE compared to RTA cases. Furthermore, there was a reciprocal pattern of staining among the HSVE patients, with reduced staining for cytochrome c oxidase subunit 1 in areas of high HSV1 staining. Conclusion: The neuropathological findings confirm previous findings for HSVE. However, the transcript data demonstrates a new finding. There is a preferential loss of mitochondrial transcripts in brain tissue of HSVE patients. The immuno-histochemical results support the array findings. The tissue results also imply that the mitochondrial damage occurs in response to HSV infection. The findings suggest that adjunctive treatments which preserve mitochondrial function could be developed as a new therapeutic option in the treatment of HSVE.

Introduction: The recent World Health Organization (WHO) report on trends in Maternal mortality (MM), from 1990 to 2013, ranks Malawi as one of the fifteen sub-Saharan countries with the highest Maternal Mortality Ratio (MM) of above 500 per 100,000live births (WHO 2014b). Malawi has no registration system for recording births and deaths. MM estimates are based on direct sisterhood methods, (used in Demographic and Health Surveys) and WHO modelled estimates, which are both highly susceptible to inaccuracies because they are both indirect methods which do not identify individual deaths within a defined population. The difficulties in obtaining accurate MMR estimates highlight the need to explore other methodologies that give more reliable data on levels as well as the cause of maternal deaths (MDs). A Reproductive Age Mortality Survey (RAMOS) is one such approach and can provide more direct and complete estimation of MMR in countries without reliable vital registration or other data sources. This is the first RAMOS used in Malawi. The aim of this study was to identify the magnitude, causes of, and factors associated with MDs in the Mangochi district in Malawi. Methods: Deaths of women of reproductive age (WRA), (15 to 49 years) that occurred from December, 2011 to November, 2012 in the district were identified. Multiple data sources were used to identify deaths, including; health facilities, communities, mortuary records and police records. Classification the death as a MD or not was done according to the ICD-10 definition. Facility based audit were conducted for all facility based MDs and verbal autopsies for all MDs. Cause of death attribution was done in three ways, 1) by a panel of experts in maternal health using the WHO application of ICD-10 to deaths during pregnancy, childbirth and puerperium (ICD-MM) (WHO 2012c), 2) by health professionals working in health facilities and 3) by using an InterVA-4 computer model. Cause of death attributed by the three methods was then compared. The three delays model was used to identify delays associated with MDs. The number of MDs identified in this study was compared to the official register in the district. MMR was calculated based on three proxy denominators; 1) number of babies who received BCG vaccine, 2) live births from the census report and 3) live births calculated from general fertility estimates. Results: A total of 424 deaths of WRA were identified and 151 of these (35.6%) were identified to be MDs. Based on the three denominators, the MMR for the Mangochi district was within the range of 341-363 per 100,000 live births (95% CI: 289-425 per 100,000 live births). Only 86 MDs had been reported via existing registers, giving an underreporting rate of 43%. The highest MMR was in age group 25-29 years (494/100,000 live births (95% CI: 349-683 per 100,000). Most MDs (62.3% (94/151)) occurred in health facilities. Based on ICD-MM cause classification, 74.8% were direct MDs, 17.3% were indirect and 7.9% were due to unknown causes. The leading cause of direct MDs (n=113) was obstetric haemorrhage (35.8%) followed by pregnancy related infections (14.4%) and hypertensive disorders (12.6%). The most frequent indirect cause of MD (n=26) was malaria (56.7%). There was low level of agreement over the cause of death between the panel of experts and health the professionals (κ= 0.37), while a substantial level of agreement was observed between the panel of experts and the InterVA-4 model (κ= 0.66). Based on ICD-MM, health professionals identified contributory factors (morbidity group) to 15.1% of MDs (n=86) as the underlying cause of death. Substandard care for obstetric emergencies, lack of blood, lack of transport, failures to recognize the severity of a problem at community level and delays in starting the decision-making process to seek health care were frequently factors associated with MDs. Conclusion: The current MD reporting system in Malawi needs strengthening. The high numbers of health facility deaths, cause of MDs and their contributing factors in Mangochi reflect serious deficiencies in the quality of maternal care that need to be urgently rectified. Urgent orientation of health workers on ICD-MM is required to obtain accurate information on cause of MDs that can be used to design effective interventions. There is need to strengthen the referral system and educate women on obstetric danger signs.

Heterogeneous malaria exposure may result in distinct clusters of higher malaria burden, or hotspots, across space and time. Targeting control programs to these areas may be highly efficient, however, operationally attractive approaches for identifying hotspots are needed for any such program to be sustained by local malaria control programs. The principal aim of this project was to investigate the ability of convenient sampling to identify hotspots of malaria transmission in a low endemic transmission setting in the western Kenyan highlands: 1) The boundaries of hotspots, and associated uncertainties, was determined using a large community survey; 2) The value of convenience sampling to estimate transmission in the community was assessed using cross-­‐sectional surveys of 4964 children in 46 government primary schools and 3042 individuals in five rural-­‐health facilities; 3) The value of compound-­‐level screening of sentinel age groups that are likely to have patent level infections was determined and; 4) The potential use for convenience sampling in hotspot targeted approaches was assessed using spatial information on residences collected during the school and health-­‐facility surveys. The community-­‐based approach was able to detect 77% of the parasite infections in selected hotspots of malaria exposure using field-­‐based tools in sentinel age groups. Both convenience-­‐sampling approaches tested produced similar estimates of malaria transmission to the community when restricted to those residing in the same catchment areas and those testing positive for malaria were more likely to reside in a hotspot. The findings suggest that operationally attractive approaches provide reliable information on malaria transmission and may play an important role in targeted malaria control strategies. Future research on ascertaining what coverage of the hotspot is needed to see sustainable reductions in transmission would provide a threshold with which to gauge the utility of this strategy.

Systems biology studies are becoming increasingly important as the need to study organisms in a holistic manner, instead of looking at processes in isolation, is being recognised. This is especially true for the study of host-pathogen interactions where the responses from bacteria are complex and overlap extensively. This thesis explores the application of 1H NMR metabolomics to the study of bacteria host interactions and seeks to identify its strengths and weaknesses with a view to integrating this technique into a combined approach that can provide an unprecedented, sophisticated understanding of host-pathogen interactions that we believe is intractable by other methodologies. The BBSRC CASE studentship that supported this work was awarded in conjunction with Procarta Biosystems Ltd who have produced a new generation of antibiotics with a novel mechanism of action. The final objective for the studentship therefore, was to develop a validated systems approach capable of defining the mechanism of action of this new class of antibiotics; transcription factor decoys (TFDs). By understanding how the target bacteria respond to antibiotic threats, the future development of new targets, delivery systems and formulations can be undertaken in a rational manner. The thesis builds towards this ultimate goal in three stages by showing, in a stepwise manner, how three increasingly complex scenarios can be interrogated by NMR metabolomics, either as a standalone technique or in combination with biophysical or genomic tools. In the first stage we investigated how the growth of different Pseudomonas aeruginosa isolates from Cystic Fibrosis patients might influence airway secretions. Growth and NMR analysis of the spent media was technically challenging, highlighting the need for improved data pre-processing techniques and experimental design. Nevertheless multivariate analysis of changes in spent media composition could be related to univariate clinical measures of respiratory disease. In the second stage we undertook a murine faecal microbiome study to show how different gut microbial communities affect the host gut metabolome. Faecal pellets were extracted into aqueous buffer and 1H NMR spectra obtained in the solution state. Clear differences in the amino acid and short chain fatty acid complement of the mouse gut were related to divergence in the gastrointestinal microbiota in the mice. The study required comparison of two separate sets of multivariate data and showed how, with application of Hierarchical Cluster Analysis, relationships between microbiota can be simplified to generate hypotheses that can be tested using metabolomic approaches. In this study the metabolomic technique was capable of identifying a link between divergence of gut microbiota and the nutritional performance of the mouse gut. In the third and final stage, we investigated whether a whole organism view could provide a bacterial perspective to enable a better understanding of how bacteria respond to antibiotic challenges. Here we combined 1H NMR spectroscopy, now of solid, bacterial cell samples (using high resolution magic angle spinning), with electron microscopy and transcriptomics to characterise the effect on Escherichia coli of four structurally and physically related antimicrobial peptides with suspected differences in their mechanisms of action. Bacterial responses characterised by the NMR metabolomic study could be detected at sub-lethal antimicrobial peptide concentrations and were qualitatively dierent according to the antimicrobial peptide. The technique was sufficiently sensitive and highthroughput to allow both a range of antimicrobial peptide concentrations to be probed as well as the bacterial response to be followed over time. Using the NMR technique to identify optimal conditions for GeneChip experiments allowed the antimicrobial peptide mechanism of action to be inferred from analysis of the ontological profile of those genes whose expression is altered in response to the antibiotic challenge. This study provided a fresh, novel perspective for previous functional and biophysical studies and shows that, with better integration with transcriptomic and other systems data, 1H NMR metabolomics will have considerable value in the study of host-pathogen interactions.