Average plasma glucose values in mg/dL (in grey) and average albumin glycation rate as a percentage of total albumin (in blue) from all the species used in this study (some of them with glucose values coming from ZIMs database; see methods) with the orders they belong to. Glucose and glycation values are standardized in order to be compared, with the dotted lines representing half the maximum and maximum values for each variable (as indicated by the axes in their corresponding colours), from inside out. Tree from a consensus on 10,000 trees obtained from “Hackett All species” on Birdtree.org, including 88 species from 22 orders (see methods).

Plasma glucose levels (in mg/dL) variation as a function of A species mean centred body mass and B residual maximum lifespan. Both glucose and body mass are log transformed. Maximum lifespan (in years) is given as the residues of a pGLS model with body mass (in grams), both log10 transformed, so the effects of body mass on longevity are factored out (see ESM6). Different bird orders, are indicated by symbols, as specified on the legends at the right side of the graphs. Figure 2.A uses the values and estimates from the glucose model without life history traits (n=389 individuals from 75 species), while Figure 2.B uses only the data points employed on the complete model (n=326 individuals of 58 species).

Final glucose model with diet, body mass, and life history traits as explanatory variables, including the significant quadratic effect of maximum lifespan. Posterior means, CI95 and pMCMC from a phylogenetic MCMC GLMM model including n=326 individuals of 58 species. Both glucose and body mass are log10 transformed and life history traits are residuals from a pGLS model of log10 body mass and log10 of the trait in question (see ESM6). Body mass was also centred to better explain the intercept, as 0 body mass would make no biological sense. The intercept corresponds to the Omnivore diet, being used as the reference as it is considered the most diverse and “neutral” group for this purpose. Significant predictors are indicated in bold.

Final glycation model with diet, body mass, glucose and life history traits as explanatory variables. Posterior means, CI95 and pMCMC from a phylogenetic MCMC GLMM model including n=316 individuals of 58 species. Both glucose and body mass are log10 transformed and life history traits are residuals from a linear model of log10 body mass and log10 of the trait in question (see ESM6). Body mass and glucose were also centred to better explain the intercept. The intercept corresponds to the Omnivore diet, being used as the reference as it is considered the most diverse and “neutral” group for this purpose. Significant predictors are indicated in bold and the confidence intervals are considered for making pairwise comparisons between the groups.

Albumin glycation rates (as a percentage of total albumin) variation as a function of individual plasma glucose values (mg/dL). Different bird orders are indicated by symbols, as specified on the legends at the right side of the graphs. The values and estimates used are from the glycation model without life history traits (n=379 individuals from 75 species).

Outcomes of the models on individual data on effects of diet on (A) plasma glucose levels and (B) albumin glycation in birds. Glucose levels are given in mg/dL, while glycation levels are a percentage of total plasma albumin which is found to be glycated. Terrestrial carnivores showed significantly higher glycation levels than omnivores (Estimate=22.01 %, CI95[18.77, 25.15], pMCMC=0.039). Models without life history traits, including more individuals, are represented, but the models with life history traits do not show differences in their qualitative predictions (i.e. higher albumin glycation in terrestrial carnivores than in omnivores; see ESM1).