Although brain structural covariance network (SCN) abnormalities were associated with suicidal thoughts and behaviors (STB) in individuals with major depressive disorder (MDD), previous studies reported inconsistent findings based on small sample size and underlying transcriptional patterns remained poorly understood.Using a multicenter MRI dataset including 218 MDD patients with STB (MDD-STB), 230 MDD patients without STB (MDD-nSTB) and 263 healthy controls (HC), we established individualized SCN based on regional morphometric measures and assessed network topological metrics using graph theoretical analysis. Machine learning methods were applied to explore and compare the diagnostic value of morphometric and topological features in identifying MDD and STB at the individual level. Brain-wide relationship between STB-related connectomic alterations and gene expression were examined using partial least square regression.Group comparisons revealed that SCN topological deficits associated with STB were identified in the prefrontal, anterior cingulate, and lateral temporal cortices. Combining morphometric and topological features allowed for individual-level characterization of MDD and STB. Topological features exhibited greater contribution to distinguishing between patients with and without STB. STB-related connectomic alterations were spatially correlated with the expression of genes enriched for cellular metabolism and synaptic signaling.These findings revealed robust brain structural deficits at network level, highlight the importance of SCN topological measures in characterizing individual suicidality, and demonstrate its linkage to molecular function and cell types, providing novel insights into the neurobiological underpinnings and potential markers for prediction and prevention of suicide.Copyright © 2024. Published by Elsevier Inc.