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

species. This approach is not an attempt to accurately describe the exact structure of the GM, but rather to assess whether some of the properties of the GM and its relationship with aging could be described with a minimalistic model. Fitting the model to the data allows to infer a biodiversity index for each of the two popula- tions, given by Hubbell’s diversity index [30, 31] and named here θrare and θabundant. The code to derive θrare and θabundant from the OTU counts table is available at https:// github.com/UniboDIFABiophysics/AlphaDiversityPublic. Analysing six publicly available data sets, we showed that our estimate of the GM biodiver- sity is coherent with those obtained with classical indices. However, θrare and θabundant appeared to have a higher statistical power when analysing the pattern of GM biodiversity with age and the relationship between biodiversity and diet or health status. Our results unveiled that GM biodiversity increases with age, with the exception of the mid- dle ages, in which it remains constant. This finding was verified using θrare and θabundant as bio- diversity estimates, but also confirmed with other classical biodiversity indices (Hill1, Hill2, Pielou, Shannon and Simpson index). The trend of GM biodiversity across the lifespan is still a matter of debate, especially in the elderly [39, 40]. Some studies suggest that GM biodiversity increases with age in healthy elderly [10, 36, 37], in agreement with our result, while others do not reveal any change of GM biodi- versity throughout such age range [41–43].

Moreover, even when the attention is focused on extreme aging and the GM biodiveristy of centenarians is analysed, results are discordant [36, 39, 42]. Some studies observe a decrease of GM biodiversity in centenarians compared to younger elderly [42], while others detect an increase [36]. Such controversies are possibly due to the high variability of data between and within data sets. Investigating the global trend of five different data sets, here we aimed to achieve a higher statistical power and, as already mentioned, we found a general increase of GM biodiversity with age. Our results, however, also highlighted some discrepancies between data sets and did not allow to draw conclusions on narrow age ranges, specifically for centenarians. Hence, fur- ther studies will be needed to validate our result and to clarify how GM biodiversity varies with age and how it behaves in extreme aging. The analysis of three case studies, allowed us to conclude that GM biodiversity is also related to the host health status and to healthy aging. We found that GM biodiversity of subjects with Down Syndrome, a model of accelerated aging [31], is lower than that of healthy controls from the same population and matched for age and sex. Moreover, we found that healthy elderly have a higher GM biodiversity compared to young subjects, as expected from the result on the increase of GM biodviersity with age. However, unhealthy elderly have a biodiversity that is lower than both young subjects and healthy elderly, corroborating the hypothesis that an increase of GM biodiversity is related to

healthy aging rather than to chronological aging alone. This result is in agreement with previous findings in which a decrease of GM biodiveristy was found to be associated with frailty [41], biological age [43] and hospitalization [3], and pointed us towards the development of a predictive model of health status in old age based on GM biodiversity. Analysing healthy and unhealhty elderly from the ELDERMET data set, we found that θrare and θabundant are good predictors of healthy aging (AUC of ROC = 0.701). On the contrary, other biodiversity indices did not show any improve predictive ability compared to sex and total number of reads alone. This confirms the greater statistical power of θrare and θabundant indices compared to the traditional ones to address healthy aging. Overall, during the analysis of the three case studies we noticed that, while results were often confirmed using classical biodiversity indices, not all of them were able to detect all the PLOS ONE Gut microbiota biodiversity and healthy aging PLOS ONE | https://doi.org/10.1371/journal.pone.0237207 October 30, 2020 12 / 23