Fig. 3: Mixed-Effect Random Forest (MERF) Regression Models Using Microbial Metabolic Pathways and Clinical Covariates to Predict Neuropsychiatric Outcomes. | Translational Psychiatry

Fig. 3: Mixed-Effect Random Forest (MERF) Regression Models Using Microbial Metabolic Pathways and Clinical Covariates to Predict Neuropsychiatric Outcomes.

From: Association between microbiome and the development of adverse posttraumatic neuropsychiatric sequelae after traumatic stress exposure

Fig. 3: Mixed-Effect Random Forest (MERF) Regression Models Using Microbial Metabolic Pathways and Clinical Covariates to Predict Neuropsychiatric Outcomes.

MERF models combining microbial metabolic pathway abundance data with clinical and demographic (highlighted in red) features demonstrate the importance of microbial pathways in predicting outcomes for (A) PTSD raw score, (B) depression normalized score, and (C) somatic symptoms count (yes/no). The top 15 features from analysis of permutated importance on model outcomes are displayed for PTSD and depression. Our model found only seven significant features for somatic symptoms.

Back to article page