FIG 2.

After adjusting for relative clinical covariates, microbiome composition differs at the genus level among elders with Alzheimer’s disease, no dementia, and other dementia types. We performed generalized mixed-effect modeling regression to predict genus-level proportions as a function of age, malnutrition, frailty, medications, and dementia state (no/other/AD). Patient ID was used as random effect to account for the repeated nature of our samples. Genera with greater than 0.1% mean relative abundance were significantly associated with Alzheimer’s disease (AD) or other dementia types (Other) in comparison to elders without dementia (NO). Only genera with a P value of <0.05 are presented, with relative abundance on the y axis.