Identification of microbiota components correlated with host lifestyle, molecular, biochemical, immunophenotypic measurements and genotype in a deeply phenotyped Sardinian cohort

The human gut microbiome plays a crucial role in health and disease, yet the relative contributions of host genetics, environmental exposures, and lifestyle factors to its variation remain incompletely understood. In this thesis, we investigated host–microbiome interactions in a large, well-characterized Sardinian cohort (ProgeNIA, N ≈ 2,650) through integrated analyses of genome-wide genotypes, shotgun metagenomic data, and deep phenotyping. Using shotgun metagenomic sequencing, we profiled taxonomic and functional features of the gut microbiota and assessed their associations with host traits. Correlation analyses identified age and sex as the strongest determinants of microbiome composition. Smoking emerged as a major lifestyle factor negatively affecting alpha diversity, with current smokers showing reduced microbial diversity compared to non-smokers, while former smokers exhibited levels comparable to never-smokers, suggesting partial recovery following smoking cessation. We also observed strong associations between white blood cell counts and reduced alpha diversity, as well as between wine consumption and increased beta diversity. We performed genome-wide association studies (GWAS) of microbial taxa, metabolic pathways, and diversity metrics using linear mixed models accounting for family structure. These analyses identified multiple genome-wide significant loci for microbial taxa and pathways, whereas fewer associations were detected for diversity metrics. Notably, one smoking-associated taxon mapped to a genomic locus previously implicated in nicotine dependence, highlighting potential links between host genetics, lifestyle, and microbiome composition. To investigate shared genetic architecture with complex diseases, we conducted Bayesian colocalization analyses using publicly available GWAS summary statistics and identified strong evidence of shared causal variants between Parabacteroides merdae and coronary artery disease. In parallel, we developed COLSTATS, a scalable web-based platform for harmonizing and interrogating large collections of GWAS summary statistics, enabling researchers to efficiently perform reproducible Bayesian colocalization analyses and systematically explore shared genetic architectures across a wide range of complex traits and diseases. Heritability analyses indicated that gut microbiome diversity is influenced by both genetic and environmental factors. While additive genetic effects explained a substantial proportion of variance, explicit modeling of cohabitation effects revealed that shared living environment contributes independently to microbiome similarity beyond genetic relatedness. Additional analyses showed that microbiome genetic associations are highly sensitive to covariate modeling, reinforcing the context-dependent nature of host–microbiome genetic effects. Overall, our results indicate that host genetic effects on the gut microbiome are modest and strongly modulated by lifestyle, environmental exposures, and shared living conditions, with important implications for clinical and public health research, as they emphasize the need to account for environmental context when interpreting microbiome–disease associations and translating them into robust biomarkers or risk models.