Cortical structural changes of morphometric similarity network in early-onset schizophrenia correlate with specific transcriptional expression patterns

BACKGROUND: This study aimed to investigate the neuroanatomical subtypes among early-onset schizophrenia (EOS) patients by exploring the association between structural alterations and molecular mechanisms using a combined analysis of morphometric similarity network (MSN) changes and specific transcriptional expression patterns. METHODS: We recruited 206 subjects aged 7 to 17 years, including 100 EOS patients and 106 healthy controls (HC). Heterogeneity through discriminant analysis (HYDRA) was used to identify the EOS subtypes within the MSN strength. The differences in morphometric similarity between each EOS subtype and HC were compared. Furthermore, we examined the link between morphometric changes and brain-wide gene expression in different EOS subtypes using partial least squares regression (PLS) weight mapping, evaluated genetic commonalities with psychiatric disorders, identified functional enrichments of PLS-weighted genes, and assessed cellular transcriptional signatures. RESULTS: Two distinct MSN-based EOS subtypes were identified, each exhibiting different abnormal MSN strength and cognitive functions compared to HC. The PLS1 score mapping demonstrated anterior–posterior gradients of gene expression in EOS1, whereas inverse distributions were observed in EOS2 cohorts. Genetic commonalities were identified in autistic disorder and adult schizophrenia with EOS1 and inflammatory bowel diseases with EOS2 cohorts. The EOS1 PLS1- genes (Z < -5) were significantly enriched in synaptic signaling-related functions, whereas EOS2 demonstrated enrichments in virtual infection-related pathways. Furthermore, the majority of observed associations with EOS1-specific MSN strength differences contributed to specific transcriptional changes in astrocytes and neurons. CONCLUSIONS: The findings of this study provide a comprehensive analysis of neuroanatomical subtypes in EOS, shedding light on the intricate relationships between macrostructural and molecular aspects of the EOS disease. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12916-023-03201-1.