The pathogenesis of major depressive disorder (MDD) is currently unclear and lacks objective diagnostic criteria. The complexity and heterogeneity of MDD also limit precise treatment. Using bioinformatics methods, we identified 18 gene signatures of MDD from the GSE98793 dataset, and validated them in an independent dataset GSE44593 (Area under the curve values were 0.92, 0.72, and 0.70 for the training, validation, and test sets, respectively). Among the gene signatures, TLR4 had the largest absolute coefficient value (coefficient = -6.13). We further identified three CD 14 + monocyte-associated gene signatures and two immune-related subtypes. The expression of TLR4 is significantly increased in subtype A of MDD (lower predicted probability), and is significantly correlated with the composition of multiple immune cells (P < 0.05). We validated that TLR4 acts as a protective factor in MDD (OR = 0.91, 95% CI = 0.85 to 0.98, P = 0.012), and its expression is driven by the same causal variants as MDD (H4/(H3 + H4) = 98.62%). Further analysis showed that the relationship between TLR4 and MDD is influenced by eight immune cell signatures. Our research provided genetic support that the immune factors may play an important role in MDD, and proposed a possible strategy for the diagnosis and treatment of MDD.