Gut microbiota ecology: Biodiversity estimated from hybrid neutral-niche model increases with health status and aging

θrare and θabundant, as well as using other traditional biodiversity scores (Hill1, Hill2, Simpson, Shannon and Pielou index). See the Materials and Methods section for details. Results show that θrare and θabundant have good predictive capability toward health status (AUC of ROC = 0.701), with θrare showing the best performances (Table 2). For all the other classical biodiversity indices, the obtained predictive accuracy is always lower (Table 2) and their AUC of ROC is comparable to the one obtained when only sex and the total number of reads are used as predictive variables (AUC of ROC = 0.644). To further evaluate the relationship between GM biodiversity and healthy aging, we com- pared the performance of the model in which health status is explained by GM biodiversity with a model in which the explanatory variables are a set of parameters traditionally associated with the health status of elderly people and available for the ELDERMET T0 data set: calf cir- cumference, BMI and the inflammatory markers IL-6, IL-8, IL-10, TNFa and CRP. The rela- tionships between such parameters, health status and healthy aging are shown in the correlation heat map of S6 Fig. As expected, the model based on the traditional aging parameters (adjusted for sex and total number of reads) has good predictive performances towards health status (AUC of ROC = 0.759). Such performances are better than those obtained by any biodiversity index (Table 2). Finally, when considering a model in which both GM biodiversity and the aging

parameters are included, the increase in performances is limited (Table 2, column “With aging parameters”), and an improvement is only obtained when biodiversity is quantified by θrare Fig 4. GM biodiversity and health status. Box-plots representing the distribution of GM biodiversity (θrare on the left and θabundant on the right) in the three groups of Healhty Hadza, Healthy Italian controls and DS Italian subjects. Mann-Whitney U-test p-values are reported for each pairwise comparison. Both biodiversity indices were adjusted for the total number of reads before computing the box-plots and hypothesis tests. https://doi.org/10.1371/journal.pone.0237207.g004 PLOS ONE Gut microbiota biodiversity and healthy aging PLOS ONE | https://doi.org/10.1371/journal.pone.0237207 October 30, 2020 10 / 23