Co-occurrence networks reveal candidate AMF–microbe assemblages for generalist and crop-specific inocula

Arbuscular mycorrhizal fungi (AMF) are widespread root symbionts increasingly used as bioinoculants. Because symbiotic efficiency depends on interactions with other plant-beneficial microbes, identifying compatible taxa and positive interaction patterns across hosts and niches could improve the design of crop-specific AMF-based inocula. Using amplicon sequencing and co-occurrence network analyses, microbial communities (AMF, fungi and bacteria) from the rhizosphere and roots of wheat were characterized and compared with those of two highly mycotrophic plant species: leek (monocotyledon) and clover (dicotyledon). Results showed that AMF diversity associated with wheat roots was 1.37- and 1.24-fold lower than that observed in leek and clover, respectively. Across all plant species, Glomus and Rhizophagus taxa dominated root-associated communities, whereas Diversispora prevailed in the rhizosphere. In contrast, wheat harboured the highest bacterial and fungal richness compared with leek and clover, whereas leek and clover roots were enriched in several functional bacterial groups, including nitrogen-fixing bacteria, denitrifying and nitrifying bacteria and plant growth promoting bacteria. Co-occurrence network analyses revealed niche partitioning with bacterial interactions predominant in roots and fungal interactions in the rhizosphere, but with fewer and less-positive connections in wheat than in leek and clover. AMF showed positive associations with beneficial bacterial taxa (e.g. Rhizobium, Pseudomonas, Streptomyces) and fungal taxa (e.g. Serendipita), with interaction patterns varying among plant species. Together, these results demonstrate that plant identity and niche jointly shape the diversity and interaction patterns of root and rhizosphere-associated microbial communities, and they highlight specific AMF-microbe assemblages as promising candidates for the development of generalist or crop-specific AMF-based inocula.