Microbiome network remodeling is associated with soil lipid metabolism in safflower under different cropping system-location combinations: an omics-based dissection.

This study investigated differences in safflower rhizosphere soil microbiota and metabolic functions among three planting pattern-location combinations (PLMEs): soybean-safflower rotation at Chenghai (CH), tobacco-safflower rotation at Longpan (ZYHY), and apple orchard intercropping at Lijiang (LJ). Each PLME comprised 6 plots. Bacterial α-diversity was significantly higher in the CH system, whereas fungal α-diversity peaked under the ZYHY system. Correlation analysis showed that bacterial α-diversity was negatively correlated with available potassium, while fungal α-diversity was negatively correlated with catalase activity. Microbial community structures significantly varied among the different PLMEs, with redundancy analysis indicating that bacterial variation was mainly explained by electrical conductivity, while fungal variation was mainly explained by available phosphorus. Under the adopted network-construction procedure, sample-specific subnetwork analysis suggested that the bacterial subnetworks in CH were more topologically complex and stable, while the fungal subnetworks in ZYHY were more complex than those in other treatments; no marked differences in fungal subnetwork stability were observed. These findings remain exploratory. Although dominant microbial taxa were unchanged, their relative abundances varied notably. Non-targeted metabolomics analysis identified significant shifts in glycerophospholipid metabolism, with five key metabolites, including L-serine, phosphatidylethanolamine, and lecithin, serving as biomarkers strongly correlated with genera such as