| Keywords [eng] |
Aphanizomenon flos-aquae, Cyanosphere, Microbiome, Genome-scale metabolic models, Metagenome-assembled genomes, Cross-feeding, Pangenome analysis, Metabolic interactions, Community modelling, Host–microbiome interactions |
| Abstract [eng] |
This thesis investigated metabolic interactions within Aphanizomenon flos-aquae-associated microbiomes and evaluated whether cyanobacterial genotype is linked to microbiome composition and predicted metabolic cross-feeding. Cyanobacterial microbiomes are increasingly recognized as important for host ecology, but their interaction structure remains less understood than their taxonomic composition. To address this, non-axenic A. flos-aquae cultures were analyzed using genome-resolved metagenomics, comparative genomics, and community-scale metabolic modelling. Metagenome-assembled genomes were reconstructed for cyanobacterial hosts and associated heterotrophic bacteria, taxonomic composition was compared across strain-associated microbiomes, and draft genome-scale metabolic models were built for selected microbiome members. Community-level analyses were then performed using MICOM, SMETANA, and metage2metabo. In parallel, host genomic diversity was assessed using Average Nucleotide Identity and pangenome analysis. The results showed that the microbiomes shared a recurring higher-rank taxonomic structure, dominated by Pseudomonadota and Bacteroidota, but varied substantially at lower taxonomic levels. For the AFA_2012_KM_D3 microbiome, as well as several other ones, a complete set of draft metabolic models was reconstructed for the host and seven associated heterotrophic bacteria. Community modelling identified both shared and microbiome-specific candidate exchange metabolites, while the number and type of predicted interactions differed among modelling frameworks. Comparative genomics showed substantial host genomic variation, including a large accessory and singleton fraction in the A. flos-aquae pangenome. Overall, the study showed that metabolic interaction modelling within A. flos-aquae-associated microbiomes is feasible and that predicted interaction networks are both community-specific and method-dependent. These results provide a genome-informed basis for future experimental validation of candidate metabolic interactions in cyanobacterial microbiomes. |
