Animal microbiota plays an essential role in keeping the physiological status of the host healthy. This microbiota consists of hundreds of different species that inhabit the host, and whose composition varies on a daily basis due to multiple factors as diets, medicine intakes, lifestyle or diseases. It is a highly dynamical system that shows variability even when the host is in a healthy condition, but visible changes happen when the system is perturbed. Thus, the main objective of this thesis aimed to study the temporal variability of human microbiota, and link it to the health status of the host. For this, we proposed a formal definition of stability in microbial systems based on the use of the Taylor#s law, an empirical law that captures the degree of variability of a community through time or space. With this tool in hand, we studied different scenarios in which the microbiome was perturbed. This law estimates the temporal variability of the microbial population and quantitatively characterises the path toward disease via a noise-induced phase transition. As the estimation of these systemic parameters may be of clinical utility, we applied this methodology to two cases of follow-up studies. The first one was related to the study of children suffering from acute infectious diarrhoea, one of the leading causes of mortality in children worldwide. This study aimed to correlate the dynamics of gut microbiota with the evolution of children who were suffering from acute infectious diarrhoea caused by a rotavirus. The experiment involved 10 children with acute infectious diarrhoea caused by a rotavirus and six healthy children. We observed that both alpha and beta diversities recovered with time in unhealthy children, something further supported by the analysis of their dynamics. The second case consisted of the study of the oral microbiomes from 26 volunteers during one month. We aimed to measure the correlations that could exist between microbial variability and the daily variability of five oxidative stress markers. We observed that all volunteers presented a different degree of variability, but constrained within a stable region. Furthermore, we also found that there was not a universal pattern of correlations between markers, or between OTUs and markers, pointing to a polymicrobial theory of disease. Finally, it is known that the behaviour of a microbial ecosystem is determined by the interactions among the participating species. We present a novel methodology where the experimental design must include alterations of the microbial ecosystem to uncover specific interactions. We demonstrate that microbial interactions can be inferred from noise-induced experiments in which the variability of a given species is modified. We illustrate our findings with two examples where a four-fold increase in the variability of a given taxon uncovers some of its interactions with other taxa in the ecosystem.

Animal microbiota plays an essential role in keeping the physiological status of the host healthy. This microbiota consists of hundreds of different species that inhabit the host, and whose composition varies on a daily basis due to multiple factors as diets, medicine intakes, lifestyle or diseases. It is a highly dynamical system that shows variability even when the host is in a healthy condition, but visible changes happen when the system is perturbed. Thus, the main objective of this thesis aimed to study the temporal variability of human microbiota, and link it to the health status of the host. For this, we proposed a formal definition of stability in microbial systems based on the use of the Taylor#s law, an empirical law that captures the degree of variability of a community through time or space. With this tool in hand, we studied different scenarios in which the microbiome was perturbed. This law estimates the temporal variability of the microbial population and quantitatively characterises the path toward disease via a noise-induced phase transition. As the estimation of these systemic parameters may be of clinical utility, we applied this methodology to two cases of follow-up studies. The first one was related to the study of children suffering from acute infectious diarrhoea, one of the leading causes of mortality in children worldwide. This study aimed to correlate the dynamics of gut microbiota with the evolution of children who were suffering from acute infectious diarrhoea caused by a rotavirus. The experiment involved 10 children with acute infectious diarrhoea caused by a rotavirus and six healthy children. We observed that both alpha and beta diversities recovered with time in unhealthy children, something further supported by the analysis of their dynamics. The second case consisted of the study of the oral microbiomes from 26 volunteers during one month. We aimed to measure the correlations that could exist between microbial variability and the daily variability of five oxidative stress markers. We observed that all volunteers presented a different degree of variability, but constrained within a stable region. Furthermore, we also found that there was not a universal pattern of correlations between markers, or between OTUs and markers, pointing to a polymicrobial theory of disease. Finally, it is known that the behaviour of a microbial ecosystem is determined by the interactions among the participating species. We present a novel methodology where the experimental design must include alterations of the microbial ecosystem to uncover specific interactions. We demonstrate that microbial interactions can be inferred from noise-induced experiments in which the variability of a given species is modified. We illustrate our findings with two examples where a four-fold increase in the variability of a given taxon uncovers some of its interactions with other taxa in the ecosystem.