Changes in grassland soil types lead to different characteristics of bacterial and fungal communities in Northwest Liaoning, China.

Actinobacteria and Ascomycota were the dominant phyla of bacterial and fungal communities, respectively, but their relative abundances were not significantly different among different grassland soil types. The abundance, number of OTUs, number of species and diversity of both bacterial and fungal communities in warm and temperate ecotone soil were the highest, while the warm-temperate shrub soil had the lowest microbial diversity. Besides, environmental factors were not significantly correlated with soil bacterial Alpha diversity index. However, there was a highly significant negative correlation between soil pH and Shannon index of fungal communities, and a highly significant positive correlation between plant cover and Chao1 index as well as Observed species of fungal communities. Analysis of similarities showed that the structural composition of microbial communities differed significantly among different grassland soil types. Meanwhile, the microbial community structure of temperate steppe-sandy soil was significantly different from that of other grassland soil types. Redundancy analysis revealed that soil total nitrogen content, pH and conductivity were important influencing factors causing changes in soil bacterial communities, while soil organic carbon, total nitrogen content and conductivity mainly drove the differentiation of soil fungal communities. In addition, the degree of connection in the soil bacterial network of grassland was much higher than that in the fungal network and soil bacterial and fungal communities were inconsistently limited by environmental factors. Our results showed that the microbial community structure, composition and diversity of different grassland soil types in northwest Liaoning differed significantly and were significantly influenced by environmental factors. Microbial community structure and the observation of soil total nitrogen and organic carbon content can predict the health changes of grassland ecosystems to a certain extent.