Abundant bacterial subcommunity is structured by a stochastic process in an agricultural system with P fertilizer inputs.

Soil microorganisms play an important role in agroecosystems and are related to ecosystem functioning. Nevertheless, little is understood about their community assembly and the major factors regulating stochastic and deterministic processes, particularly with respect to the comparison of abundant and rare bacterial subcommunities in agricultural systems. Here, we investigated the assembly of abundant and rare bacterial subcommunities in fields with different crops (maize and wheat) and phosphorus (P) fertilizer input at three different growth stages on the Loess Plateau. The high-throughput sequencing dataset was assessed using null and neutral community models. We found that abundant bacteria was governed by the stochastic process of homogenizing dispersal, but rare bacterial subcommunity was predominant by deterministic processes in maize and wheat fields due to broader niche breadths of abundant species. Soil nitrogen (N) and P also determined the assembly of abundant and rare soil subcommunities. The relative abundance and composition of the abundant and rare bacterial subcommunities were also influenced by soil nutrients (soil available P (AP) and NO-N) and agricultural practices (P fertilization and crop cultivation). In addition, the abundant bacterial community was more susceptible to P fertilizer input than that of the rare bacteria, and a higher relative abundance of abundant bacteria was observed in the P70 treatment both in maize and wheat soils. The microbial co-occurrence network analysis indicated that the maize field and low nutrient treatment exhibited stronger associations and that the abundant bacteria showed fewer interconnections. This study provides new insights toward understanding the mechanisms for the assembly of abundant and rare bacterial taxa in dryland cropping systems, enhancing our understanding of ecosystem diversity theory in microbial ecology.