The human gut is colonized by a vast bacterial community that is currently rather well characterized at the species level. Yet, each of these species harbor a tremendous amount of individual genetic variations. Our understanding of the human gut microbiome, its dynamics, composition and impact on host health requires a deeper characterization of its
bacteria. The amount of publicly available shotgun sequencing data as well as development of computational tools allowed to reach strain-level resolution in metagenomic analysis. In this thesis, I present systematic approaches to study the strain-level variation using complementary phylogenetic and pangenomic methods aiming to address fundamental
questions about microbiome transmission in early life as well as impact of functions encoded by microbiome strains on host health. Across two different cohorts, I used a recently-developed strain-tracking method to assess the impact of delivery conditions on the initial seeding of the infant gut microbiome. While mode of delivery (vaginal or C-section) had a great impact on the amount of mother strains transmitted to the infant, place of delivery (home or hospital) and breastfeeding duration also had an impact on the ongoing development, strain replacement or persistence over the first year of life. In comparison, the father appeared as a stable source of strains independent of the delivery mode. This initial mother seeding, despite being reduced in C-section delivery, can be compensated by Fecal Microbiota Transfer, demonstrating the need of fecal microbiota exposure in seeding during vaginal delivery. In addition, strain dynamics was shown partially explained by differences in the carbohydrates degrading capacities, especially the ability to feed on Human Milk Oligosaccharides. These differences in metabolism between strains were also observed by their respective empirical growth rate that was seen associated with transmission and persistence in the infant gut. To further systematically assess the differences of metabolic capacities between strains and the impact on hosts, I developed a new method to identify gene groups (PanPhlAn Genomic Islands, PGIs) co-present across conspecific strains in metagenomic samples. By applying this method on a large collection of over 10,000 samples, I was able to build a set of 5,315 PGIs. Deeper characterization of these PGIs revealed horizontal gene transfer across species, high variation in carbohydrate metabolism capacities and association with the host lifestyle and health status. Together, these analyses demonstrated the complementary aspects of strain variation andstressed out the need to encompass both strain phylogeny and gene content to fully understand the microbiome at the strain-level.
| Identifer | oai:union.ndltd.org:unitn.it/oai:iris.unitn.it:11572/384731 |
| Date | 27 July 2023 |
| Creators | Dubois, Leonard |
| Contributors | Dubois, Leonard, Segata, Nicola, Valles Colomer, Mireia |
| Publisher | Università degli studi di Trento, place:Trento |
| Source Sets | Università di Trento |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/doctoralThesis |
| Rights | info:eu-repo/semantics/embargoedAccess |
| Relation | firstpage:1, lastpage:127, numberofpages:127 |
Page generated in 0.0016 seconds