Gut mycobiota dysbiosis in drug-naïve, first-episode schizophrenia

Bacterial dysbiosis has been demonstrated in patients with schizophrenia (SCH). The aim of the present study was to investigate alterations in mycobiota composition and fungi-bacteria correlation network in drug-naïve, first episode SCH. We recruited 205 SCH patients and 125 healthy controls (HCs), whose gut bacterial and fungal compositions were characterized by 16S and 18S ribosomal RNA gene amplicon sequencing, respectively. Fungal-bacterial relative correlation network analysis was performed using the Spearman's test and distance correlation. We also computed relative networks connectedness, which represents the ratio of significant interactions (edges) and taxa (nodes) in the network. SCH patients showed lower fungal α-diversity compared with that of HCs. Furthermore, we identified 29 differential fungal markers at multiple taxonomies between SCH patients and HCs. SCH patients also showed a significantly lower fungi-to-bacteria α-diversity ratio compared with that of HCs (p = 1.81 × 10^-8). In risk prediction models, we observed that combining bacterial and fungal markers achieved higher accuracy than that of bacterial markers alone (AUC = 0.847 vs AUC = 0.739; p = 0.043). Fungal-bacterial correlation network was denser in HCs than in SCH patients and was characterized by a high number of neighbors (p < 0.05). In addition, an increased abundance of Purpureocillium was associated with more severe psychiatric symptoms and poorer cognitive function in SCH patients (p < 0.05). Our study demonstrated a disrupted and weakened fungi-bacteria network in SCH patients, which might be associated with their clinical manifestations. Future research on fungal-bacterial correlation network is warranted to advance our understanding about the role of mycobiota in the etiology of SCH and to explore novel intervention approaches.