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

hunter-gatherers of Tanzania, that are known to have a life-style that is beneficial for the GM [35] (Healthy Hadza). Biodiversity data were adjusted for the total number of reads before per- forming the hypothesis tests, while sex balance among the groups was guaranteed by the origi- nal authors [33, 35]. In agreement with previous findings [35] and with the results obtained using other classical biodiversity indices (S8 Fig), the GM biodiversity turns out to be statistically significantly higher in the Healthy Hadza compared to the Healthy Italian control group, both when considering θrare (Mann-Whitney U-test p-value = 0.0004) and θabundant (Mann-Whitney U-test p-value = 0.0022), as shown in Fig 4. This result confirms the ability of θrare and θabundant to detect differ- ences in GM biodiversity in the presence of important diet and life-style differences. Regarding the comparison between Healthy Italian controls and DS Italian subjects, instead, the differences detected in the original paper [33] or using classical biodiversity indices such as Pielou, Shannon or Simspon (S8 Fig) are negligible. However, when estimating the GM biodi- versity from the 2NB model, we detect a statistically significant decrease of GM biodiversity (θrare) in DS subjects compared to the control group (Mann-Whitney U-test p-value = 0.0293), Table 1. Coefficients and t-test p-values for each parameter of the spline regression model. θrare θabundant Hill1 Hill2 Pielou Shannon Simpson Coef. P>|t| Coef. P>|t| Coef. P>|t| Coef. P>|t| Coef. P>|t| Coef. P>|t| Coef. P>|t| Global

model Intercept -0.104 0.3 0.071 0.465 -0.124 0.205 -0.144 0.147 -0.257 0.011 -0.178 0.068 0.159 0.114 spline1 -0.343 0.115 -0.522 0.014 -0.708 0.001 -0.647 0.003 -0.544 0.014 -0.708 0.001 0.526 0.017 spline2 0.015 0.9 0.007 0.954 0.096 0.416 0.14 0.242 0.172 0.161 0.151 0.2 -0.157 0.196 spline3 0.225 0.024 0.586 <0.001 0.488 <0.001 0.363 <0.001 0.115 0.256 0.379 <0.001 -0.209 0.038 Data set specific deviations from the global model Bian Intercept 0.049 0.643 -0.128 0.21 0.024 0.818 -0.001 0.992 0.19 0.075 0.068 0.507 -0.047 0.658 spline1 0.447 0.046 0.153 0.482 0.586 0.008 0.695 0.002 0.593 0.009 0.684 0.002 -0.648 0.004 spline2 -0.236 0.081 -0.139 0.288 -0.47 <0.001 -0.575 <0.001 -0.651 <0.001 -0.618 <0.001 0.66 <0.001 spline3 -0.163 0.153 -0.141 0.202 -0.092 0.411 -0.122 0.282 0.247 0.033 0.002 0.988 -0.06 0.605 Biagi Intercept -0.117 0.515 -0.36 0.04 -0.404 0.022 -0.33 0.065 -0.349 0.057 -0.377 0.033 0.304 0.095 spline1 -0.047 0.926 -0.702 0.15 -1.07 0.03 -1.01 0.042 -1.31 0.01 -1.097 0.025 1.011 0.046 spline2 0.151 0.594 0.278 0.312 0.713 0.01 0.716 0.011 0.755 0.009 0.636 0.022 -0.705 0.014 spline3 -0.221 0.244 0.064 0.728 -0.047 0.802 -0.036 0.85 0.207 0.284 0.085 0.648 -0.002 0.993 Kong Intercept 0.377 0.096 0.694 0.002 0.339 0.127 0.198 0.379 0.234 0.309 0.399 0.072 -0.184 0.422 spline1 1.119 0.038 1.943 <0.001 1.068 0.043 0.591 0.269 0.528 0.335 1.129 0.032 -0.536 0.325

spline2 -0.506 0.035 -0.821 <0.001 -0.368 0.117 -0.232 0.329 -0.219 0.368 -0.425 0.069 0.178 0.461 spline3 -0.235 0.23 -0.429 0.024 -0.36 0.061 -0.161 0.406 -0.075 0.707 -0.304 0.111 0.174 0.378 Odamaki Intercept 0.138 0.213 0.283 0.008 0.233 0.031 0.215 0.05 0.05 0.657 0.119 0.269 0.018 0.872 spline1 -0.851 <0.001 -0.466 0.041 -0.373 0.104 -0.255 0.273 0.117 0.622 -0.475 0.038 0.427 0.071 spline2 -0.125 0.405 -0.032 0.824 0.12 0.414 0.174 0.242 0.309 0.043 0.373 0.011 -0.527 0.001 spline3 1.114 <0.001 0.782 <0.001 0.486 0.002 0.296 0.068 -0.376 0.023 0.221 0.165 0.118 0.475 ELDERMET T0 Intercept -0.446 0.105 -0.489 0.067 -0.192 0.476 -0.082 0.764 -0.125 0.653 -0.21 0.434 -0.091 0.742 spline1 -0.668 0.224 -0.928 0.082 -0.21 0.697 -0.021 0.969 0.072 0.897 -0.241 0.653 -0.254 0.648 spline2 0.716 0.033 0.715 0.028 0.006 0.987 -0.083 0.802 -0.194 0.568 0.035 0.916 0.393 0.246 spline3 -0.494 0.082 -0.276 0.317 0.012 0.965 0.023 0.936 -0.003 0.991 -0.003 0.99 -0.23 0.421 Covariates Sex 0.228 <0.001 0.086 0.067 0.254 <0.001 0.259 <0.001 0.28 <0.001 0.287 <0.001 -0.264 <0.001 NumOfReads -0.008 0.844 0.271 <0.001 0.234 <0.001 0.262 <0.001 0.134 0.001 0.23 <0.001 -0.204 <0.001 The upper block of parameters refers to the global behavior (Global model), while in the following blocks show how each data set deviates from that trend. P-values are

bold if < 0.05. When statistically significantly different from zero, cells with coefficients and corresponding p-values of the Global model are shaded red (if positive) or blue (if negative). Due to its opposite trend, inverted colors are used for results relative to the Simpson index. https://doi.org/10.1371/journal.pone.0237207.t001 PLOS ONE Gut microbiota biodiversity and healthy aging PLOS ONE | https://doi.org/10.1371/journal.pone.0237207 October 30, 2020 7 / 23