This study explores the potential of gut microbiome-based biomarkers for predicting immune-related adverse events (irAEs) associated with immune checkpoint inhibitors (ICIs), particularly anti-PD-1/PD-L1 therapy. The researchers analyzed microbiome data from 317 patients and in-house generated data from 115 patients with irAEs, using a machine learning approach, specifically the Random Forest (RF) algorithm, to construct a microbiome-based classifier that distinguishes between irAEs and non-irAEs. The RF classifier, developed using 14 microbial features, demonstrated robust discriminatory power (AUC=0.88) between non-irAEs and irAEs. The predictive score from the classifier showed significant discriminative capability in two independent cohorts. Functional analysis revealed that the altered microbiome in non-irAEs was characterized by increased menaquinone biosynthesis, accompanied by elevated expression of rate-limiting enzymes menH and menC. Targeted metabolomics analysis further highlighted a notably higher abundance of menaquinone in the serum of patients who did not develop irAEs compared to the irAEs group. The study concludes that microbial biomarkers have potential for predicting the onset of irAEs and highlights menaquinone as a possible selective therapeutic agent for modulating the occurrence of irAEs. The study also identifies specific microbial species that can distinguish between patients experiencing irAEs and non-irAEs, and provides new insights into preventing irAEs and uncovering potential mechanisms from a microbiological functional perspective. The study further validates the robustness of the model using external validation and shows that the gut microbiome-derived biomarker panel has excellent accuracy across studies. The results suggest that menaquinone, a metabolite derived from the microbiome community, may serve as a potential functional microbial metabolite for defending against the occurrence of irAEs. The study also highlights the potential of gut microbiome and their metabolites for effective synergistic antitumor response with ICI therapy and alleviating the toxicity induced by ICI drugs. The findings suggest that menaquinone may exert potential protective effects via inhibiting the pro-inflammatory signaling pathway such as NF-κB signaling pathway. The study provides a comprehensive analysis of the gut microbiome's role in predicting irAEs and highlights the potential of menaquinone as a therapeutic agent for modulating the occurrence of irAEs.This study explores the potential of gut microbiome-based biomarkers for predicting immune-related adverse events (irAEs) associated with immune checkpoint inhibitors (ICIs), particularly anti-PD-1/PD-L1 therapy. The researchers analyzed microbiome data from 317 patients and in-house generated data from 115 patients with irAEs, using a machine learning approach, specifically the Random Forest (RF) algorithm, to construct a microbiome-based classifier that distinguishes between irAEs and non-irAEs. The RF classifier, developed using 14 microbial features, demonstrated robust discriminatory power (AUC=0.88) between non-irAEs and irAEs. The predictive score from the classifier showed significant discriminative capability in two independent cohorts. Functional analysis revealed that the altered microbiome in non-irAEs was characterized by increased menaquinone biosynthesis, accompanied by elevated expression of rate-limiting enzymes menH and menC. Targeted metabolomics analysis further highlighted a notably higher abundance of menaquinone in the serum of patients who did not develop irAEs compared to the irAEs group. The study concludes that microbial biomarkers have potential for predicting the onset of irAEs and highlights menaquinone as a possible selective therapeutic agent for modulating the occurrence of irAEs. The study also identifies specific microbial species that can distinguish between patients experiencing irAEs and non-irAEs, and provides new insights into preventing irAEs and uncovering potential mechanisms from a microbiological functional perspective. The study further validates the robustness of the model using external validation and shows that the gut microbiome-derived biomarker panel has excellent accuracy across studies. The results suggest that menaquinone, a metabolite derived from the microbiome community, may serve as a potential functional microbial metabolite for defending against the occurrence of irAEs. The study also highlights the potential of gut microbiome and their metabolites for effective synergistic antitumor response with ICI therapy and alleviating the toxicity induced by ICI drugs. The findings suggest that menaquinone may exert potential protective effects via inhibiting the pro-inflammatory signaling pathway such as NF-κB signaling pathway. The study provides a comprehensive analysis of the gut microbiome's role in predicting irAEs and highlights the potential of menaquinone as a therapeutic agent for modulating the occurrence of irAEs.