Transmission of human microbiome: Computational metagenomic tools and biomedical applications

The human gut microbiome is important for health and its disruption can lead to disease. Transmission from other people is the main way we obtain microbiomes in our gut. It has been established that the first microbes are acquired from mother upon birth and more recent findings show that people co-housing share a substantial portion of their microbiome. However, the impact of social contact on the developing gut microbiome after the initial seeding from the family remains unexplored. Fecal microbiota transplantation (FMT) is a clinical procedure to modulate a recipient's gut microbiome in order to improve health or response to treatment and can be viewed as a forced transmission of the whole microbial community. Despite many single cohort studies published, its success in treating Clostridioides difficile infection and promising preliminary results in cancer immunotherapy, the underlying mechanisms of microbial engraftment and the links to clinical outcomes remain to be understood, preventing informed clinical decisions. In my doctoral thesis I explore the development and application of computational methods of strain sharing to study microbiome transmission and engraftment in FMT. During my PhD work I was continuously collecting metagenomic assembled genomes (MAGs) and clustering them to define species-level genome bins (SGBs) that allow for detecting transmission of otherwise uncharacterized species in our cohorts. I developed and maintained a computational pipeline for strain sharing detection among metagenomic samples. In Chapter 2 I applied the strain sharing pipeline to over 1,000 samples from 43 babies attending nursery for the first time with weekly microbiome sampling along with their educators and family. My pipeline supported the data analysis, which highlighted the importance of social contact in the development of the infant’s microbiome by observing substantial strain transmission among babies after only one month of attending the nursery. It enabled the tracking of individual strains spreading among the nursery shedding light to the dynamics of microbial transmission. In Chapter 3 we pooled together all available FMT cohorts with shotgun metagenomic sequencing for the total of 226 donor-recipient pairs. I participated in the data and metadata collection, applied the strain sharing detection pipeline and analyzed the strain engraftment from donors to recipients. We showed that recipients using antibiotics or with an infectious disease both showed higher engraftment rates and discovered that the combined route of delivery (for example both capsules and colonoscopy) yields higher engraftment rates than a single route. Although the mechanism of action of FMT is disease specific, we found an overall link between engraftment rate and clinical success of the treatment. By looking at the engrafted species we observed a pattern of Bacteroidetes and Actinobacteria species engrafting at a higher rate than Firmicutes. My application of the strain sharing methodology and data analysis played an important role in better understanding the patterns of engraftment in FMT and by helping clinicians to make more informed decisions. Overall my PhD work contributed to successful applications of computational methods of strain sharing detection to gain new insights into microbial transmission and clinical applications.