2.1.3. Carbohydrate Metabolism Study 1

There were no consistent changes in fasting glucose, fasting insulin, C-peptide, HOMA, or two-hour glucose response to the OGTT across the three treatment groups for either the overweight (*p =* 0.17, 0.34, 0.49, 0.15, and 0.86, respectively) or the obese groups (*p =* 0.86, 0.14, 0.14, 0.22, and 0.28, respectively) (ANOVA for fasting glucose and two-hour glucose; Kruskal–Wallis for remaining variables). However, there was a trend towards a reduction to the two-hour insulin response to the OGTT (−12%) in the 4 g dose for the overweight participants only (*p =* 0.05), as determined by a Kruskal–Wallis test (Figure 3).

**Figure 3.** The difference in the 2-h insulin response to the OGTT for overweight participants. N=9 for overweight 2-h insulin due to missing data (Placebo = 3, 2 g = 3, and 4 g = 3). Standard error bars shown.

#### 2.1.4. Microbiome Results Study 1

There were significant differences (*p* < 0.05) between the pooled active treatments groups (2 g SXRG84/day and 4 g SXRG84/day combined) compared to the placebo group for both overweight and obese participants as separate groups in terms of the most-changed composition and abundance of genera of bacteria from before and after the six-week intervention. This pattern was consistent for both the overweight and the obese participant groups (Figure S3). Below, Figure 4 depicts the change in composition and abundance of genera over a six-week period between the overweight and obese participants on placebo (enclosed circles), which was much smaller compared to those on the active treatment (open circles).

Permutational multivariate analysis is a powerful tool to demonstrate the overall ecological shifts of significance in the treatment and control communities of the microbiome, or the effective changes to the betadiversity. However, translating ecosystem shifts in local diversity compared to the global diversity, or betadiversity, to select species is only complicated by cascading interaction and external effects. Nevertheless, it is of interest to determine which species contributed to the significant change in betadiversity, as experiments can be designed to test cause-and-effect studies over time. Therefore, SIMPER analysis revealed 15 genera that contributed to at least 90% of the differences between placebo and pooled active group changes (Table 4). These 15 genera were examined individually to assess whether there were consistent changes in the two treatment groups compared to the placebo group. Five genera were identified that appeared to demonstrate a treatment effect. The individual genera that increased most in contrast to the placebo group were *Akkermansia*, *Clostridium*, *Pseudobutyrivibrio*, and *Bifidobacteria* (mostly *B. longum*) sp. Although not identified in the SIMPER analysis, only one genus seemed to decrease, and this was *Bilophila* sp. (Figure 5)—although the frequency of this change was not large enough to be considered one that contributed to 90% of the variation across treatments.

**Figure 4.** Principle component analysis of the change of composition of genera that was present in each treatment group, including Obese (light) and Overweight (dark) people on active treatments (open shapes) and placebo treatments (closed shapes). Obese Active = Obese participants on active (2 g/SXRG84 or 4 g/SXRG84), Overweight Placebo = Overweight participants on placebo, Overweight Active = Overweight participants on active (2 g/SXRG84 or 4 g/SXRG84), Obese Placebo = Obese participants on placebo.

**Table 4.** Genera as identified in SIMPER analysis that contributed to the differences between treatment groups. \* Indicates potential for treatment effect.


**Figure 5.** Univariate genera that were identified as those that contributed most to the difference between the active and placebo treatments consistently. Standard error bars are shown.

2.1.5. Dietary Data, Bowel Movements, Urinary F2-Isoprostanes—Blood Count Results Study 1

Dietary intake and urinary F2-isoprostanes were included as a tertiary outcome in the study. Dietary intake between the three treatment groups was not significantly different at baseline for total energy (kJ), all macronutrients, sugar (g), and dietary fiber (g) and all three groups had no difference in total diet score. In the 4 g treatment group, there was a trend towards a 10% reduction in saturated fat (*p =* 0.06), as determined by ANOVA, and a trend towards a 33% reduction in added sugar as a percent of total energy after the six-week treatment, as determined by ANOVA (*p =* 0.08, Table S1, Supplementary Material).

There was no significant difference between treatment groups in regard to a change in frequency in bowel movements throughout the trial (results not shown). Safety measures such as full blood-count data (Table S4 and urinary F2-isoprostanes remained stable throughout the trial, with no significant changes detected between groups (*p* > 0.05) (Table S2).
