**3. Results**

—

Alpha diversity values of the environment, ingredients, and fermenting sauerkraut were measured by both the Shannon diversity index and Faith's phylogenetic diversity. These values reveal a reduction in bacterial diversity for fermenting sauerkraut as compared to the starting ingredients and environment (Figure 1). However, we observe contrasting alpha diversity patterns between the Shannon diversity and the Faith's phylogenetic diversity metrics during the fermentation process. The Shannon diversity index indicates a successive increase in alpha diversity of sauerkraut over time while Faith's phylogenetic diversity suggests a constant low alpha diversity. We attribute this to the fact that the Shannon diversity index segments closely related and possibly overlapping LAB into separate taxa. This generates a false appearance of diversity. By contrast, Faith's phylogenetic diversity index uses branch lengths as the basis for assigning diversity metrics, and does not separate LAB with the same level of granularity. Nevertheless, the low level of diversity in the fermenting product shown by both plots is likely the result of selective pressures in the fermentation environment, including low pH, anaerobic conditions, and high salinity. This indicates successful fermentation. ndex and Faith's iversity and the Faith's h's contrast, Faith's

) Faith's phylogenetic diversity (PD). Error bars represent standard error of the **Figure 1.** Alpha diversity measures of the sauerkraut, ingredient, and environment samples. (**A**) Shannon index; and (**B**) Faith's phylogenetic diversity (PD). Error bars represent standard error of the mean.

— Next we employed principal coordinates analysis (PCoA), with the unweighted UniFrac distance, to visualize the differences in community structures between our samples (Figure 2). The samples exhibited clear clustering by sample type—raw ingredient, environment, or fermentation time

point—indicating that the bacterial communities present have significant variation from one another. This is expected for the environmental samples versus raw ingredients or fermentation time points, but was surprising for the raw ingredients versus the fermenting sauerkraut. The Day 0 sample derived from the initial ingredient mixture is much more similar to the Day 14 sauerkraut community than it is to the raw ingredients (Figure 2). This suggests that the selective pressures intrinsic to fermentation have strong and immediate impacts on the bacteria found on and in the raw ingredients.

**Figure 2.** PCoA depicting the unweighted UniFrac distance between fermenting sauerkraut, environment, and ingredient microbiome samples.

To further characterize the bacterial community found in each of the collected samples, we examined the taxonomic structures of our bacterial communities at the order and genus levels (Figure 3). At the order level, we see differences in bacterial composition between the raw ingredient samples, the environmental samples, and the fermentation samples (Figure 3A). The raw ingredient and environmental site categorizations are fairly similar to each other. Sea salt differs slightly from the cabbage and caraway seed, which may be because salt is not a plant product and it likely presents a strong halophilic selection pressure. Overall, these two categories of samples are markedly different from the bacterial communities found during fermentation. The Day 0 sample contains significantly more bacterial taxa than the subsequent time point samples, and illustrates a precipitous drop in the number of bacterial species present over the first 48 hours of fermentation. The most abundant bacterial order in the Day 0 fermentation sample is Pseudomonadales, which is also a high abundance order in all of the environmental samples. This suggests that the environment plays some role in establishing the initial bacterial community of the combined ingredients. After two days, the most abundant bacteria present are of the Lactobacillales order, which is expected in the case of successful fermentation. This pattern persists throughout fermentation and jarring.

At the genus level, we find many of the same trends hold (Figure 3B). The three ingredient samples look the most similar at this level, with *Halomonas* common to and prevalent in all three of the samples. The environmental samples continue to show similarity, and the lack of similarity between the environmental samples and the Day 0 fermentation samples persists here as it did at the order level. LAB dominate the other fermentation samples as they did at the order level, with *Leuconostoc* and

*Lactobacillus* as the dominant genera. This is in line with the results published by Pederson and Albury 1969, which showed *Leuconostoc* and *Lactobacillus* as main players in the sauerkraut fermentation process [12].

**Figure 3.** Relative abundance of bacterial taxa in the fermenting sauerkraut, ingredient, and environmental samples at the (**A**) order and (**B**) genus levels. Only the top seven taxa from the fermenting and ingredients/environment sample groupings are shown.
