Controlled Access. Restriction lift date: 2031-12-31
Intrinsic and extrinsic factors impacting the gut microbiome along the mother-infant axis
Loading...
Files
Date
2026-06-10
Authors
Joos, Raphaela
Journal Title
Journal ISSN
Volume Title
Publisher
University College Cork
Published Version
Abstract
The human gut microbiome has received substantial attention within the global research community in recent years. Commonly denominating the entirety of gastrointestinal microbial organisms, including fungi, archaea, viruses and bacteria, as well as all their genes encoding metabolic activity, it plays a crucial role in human functioning. Infancy especially is a critical window, as microbial assembly co-occurs with overall development, thereby potentially having long lasting effects on current and future health. Accordingly, the early-life gut microbiome has repeatedly been in the focus of modern biomedical research as a potential clinical biomarker. However, despite considerable enthusiasm, significant conceptual and methodological challenges remain, particularly in relation to its high inter-individual variability, rapid developmental dynamics, methodological heterogeneity, and its pronounced malleability in response to early-life exposures including prenatal factors, mode of delivery and infant feeding patterns. The aim of this thesis was to address these challenges by critically examining both the promise and limitations of the infant gut microbiome as a biomarker in development and health. In doing so, it focused not only on generating empirical findings but also on interrogating previous assumptions, definitions and analytical strategies that currently underpin early-life microbiome research.
First, an overview of current evidence on the role of the early-life gut microbiome in infant development is provided, alongside a critical discussion of methodological approaches used to generate these findings. Traditional statistical methods were compared with machine-learning approaches, highlighting both their demonstrated successes and persistent challenges within microbiome data analysis. Collectively, this work positions the infant gut microbiome as a dynamic and context-dependent system, underscoring both its promise and its limitations as a biomarker of early-life health. After, we addressed one of the most fundamental and pressing questions in contemporary microbiome research: what constitutes a “healthy” microbiome? Beyond discussing current approaches to defining microbiome health, i.e. the binary categorisation of disease states versus healthy controls, this chapter critically examined emerging conceptual frameworks and future perspectives. In doing so, it outlined a road map for how microbiome science may draw on principles of epidemiology to enhance interpretability, reproducibility and translational relevance, while also providing a conceptual foundation that informed the design, analytical choices, and interpretation of findings throughout the remainder of this thesis.
Building on this conceptual foundation, the first experimental chapter (Chapter 3) turned to methodological considerations and their critical influence on microbiome research outcomes. More specifically, we investigated how long-term frozen storage and DNA extraction methodology impact DNA yield and microbial community profiles samples from infant stool. Unanalysed infant stool samples collected in 2013 as part of a probiotic infant formula clinical trial were used to quantify differences in pre-extracted DNA and deep-frozen stool processed with an updated DNA extraction protocol. Whole-genome shotgun sequencing was applied to both extraction methods, respectively both timepoints, to estimate microbial abundances. Besides a high degradation rate of 92.6% observed in pre-extracted DNA aliquots over time, DNA quality metrics remained stable, while clear differences in DNA yield and microbial profiles were observed according to Gram stain. Notably, inter-individual variation exceeded methodological variation by approximately 42-fold, supporting the viability of both extracted DNA as well as stool samples for downstream analyses when stored under appropriate conditions.
In Chapters 4, 5, and 6 we aimed to explore the role of the early life microbiome in infant development, employing longitudinal data from the Cork-based COMBINE cohort. In brief, 456 infants were followed from birth to two years of age, resulting in over 1800 metadata variables collected at nine distinct study visits, and 541 stool samples collected at six visits up to one year of age, curated with whole-genome shotgun sequencing and the BioBakery suite. Our first aim was to examine how dietary transitions, i.e. the introduction of solid foods around 4 to 6 months of age, manifests in the infant gut microbiome – particularly on a functional level (Chapter 4). Using a repeated-measures study design, we selected infants with one sample collected during exclusive milk-feeding, and one during complementary feeding, and conducted a quantitative analysis of the impact of feeding patterns on microbial composition and function. Overall, community diversity increased by 33% over the first year of life, manifested in a shift in community composition from (facultative) aerobes to a more anaerobic gut environment, as well as a change from carbohydrate metabolism of simple, predominantly milk-derived, sugars to more complex, fibre-based substrates. Formula-fed infants, particularly, demonstrated high microbial volatility and diversity during early life, supporting a dose-response effect of breastfeeding on a more paced microbial maturation process.
In Chapter 5, differences in compositional and functional microbial profiles were investigated according to cognitive development outcomes assessed at age two years. A multivariate clinical analysis was paired with a covariate analysis of the microbiome to determine what early-life exposures covarying with cognitive outcomes also covary with microbial profiles, and what microbial differences between differential cognitive outcomes we can observe. We found cognitive outcomes to differ based on household income, infant sex, gestational age, maternal height, as well as with feeding practices in the first year of life, the latter also covarying distinctly with infant gut microbiota. Low cognitive outcome infants showed high microbial diversity early on, which progressed to low community diversity at one year of age, when functional, as opposed to compositional, taxa showed alterations based on cognitive outcome.
Finally in Chapter 6, we applied a machine-learning approach to build a microbiota age model, predicting chronological age from microbial community maturity. Overall, the model explained 52% of differences in age, successfully predicting age with a mean absolute error of 8.01 weeks and a root mean squared error of 11.05 weeks. The most important predictors of age were feeding-associated taxa including Faecalibacterium prausnitzii, Firmicutes bacterium CAG 41, Erysipelatoclostridium ramosum, Anaerostipes hadrus, Agathobaculum butyriciproducens, and Ruminococcus gnavus. While baselines were the most difficult to predict, most infants followed a normative microbiota maturation trajectory during the first year of life. A smaller proportion exhibited heterogeneous, non-normative patterns, which was only poorly explained by early-life exposures.
Taken together, the findings presented in this thesis extend our understanding of the role of the early-life microbiome in health and disease on a methodological and (meta-) scientific level. As such, the collective findings of this thesis suggest that the most valuable contribution of early-life microbiome research may not lie in identifying individual microbial taxa as clinical biomarkers, but rather in developing trajectory-based, probabilistic frameworks that integrate microbiome function, developmental timing, as well as individual context and exposures. By highlighting individuality and novel conceptual approaches, this work contributes to a more realistic roadmap for successfully translating infant microbiome research into meaningful health insights which could potentially be applied in a clinical setting.
Description
Controlled Access
Keywords
Microbiome , Infant gut microbiome , Microbial succession
Citation
Joos, R. 2026. Intrinsic and extrinsic factors impacting the gut microbiome along the mother-infant axis. PhD Thesis, University College Cork.