High-resolution microbial network analysis defines biocontrol consortia in the wheat phyllosphere

Plant-associated microbiomes comprise diverse microbial species that coexist and interact, influencing community structure and host plant health. However, our understanding of these interactions in field conditions and at strain-resolution remains limited. This hinders the development of effective biocontrol applications, as laboratory findings often fail to translate to field settings due to insufficient insights into in situ interaction network structures. This study addresses this limitation by employing taxon-specific high-resolution amplicons to resolve a cross-kingdom co-occurrence network within the wheat phyllosphere microbiome. We performed in-depth monitoring of strains from the hub genus Pseudomonas, revealing a high degree of strain-specificity of Pseudomonas interactions both within and across kingdoms. Through negative interaction modelling, we identified a consortium of ten biocontrol taxa with the potential to suppress seven fungal pathogens in field conditions. Additional stabilizer strains were found to likely enhance persistence. We validated the strain-specific interactions of Pseudomonas with the major fungal pathogen Zymoseptoria tritici using co-inoculation experiments with genotypes retrieved from the same field. Consistent with our prediction, we identified a P. poae isolate as the most antagonistic towards the pathogen both in vitro and in planta. Our study demonstrates that taxon-specific high-resolution network inference can effectively map microbial interaction networks and predict strain-specific interaction patterns of biocontrol genotypes with high persistence under field conditions. Our novel approach supports the design of more effective and sustainable biocontrol strategies.