Nutritional composition of yogurt
Analysis of biochemical parameters
The quality and safety of yogurt depends on its nutritional composition and microbiome diversity as well as quality of the starter culture. To assess yogurt quality, the nutritional properties of yogurt in terms of biochemical parameters and trace minerals were examined in this study. Biochemical parameters of yogurt samples including pH, fat, moisture, total solid (TS), solid-not-fat (SNF) and ash contents are summarized in Table 1. During the commercial yogurt manufacturing process, pH needs to be maintained in order to prevent the generation of rancid smell, taste and flavor. A high pH indicates improper fermentation and allows many undesirable microorganisms to grow in the medium. We found significant differences (P < 0.05) in pH values of all yogurt varieties studied in this report. The mean pH values of the yogurt samples remained slightly acidic and, on an average, ranged from 5.28 to 6.33. The Sweet Brand 6 (SwV5) had the highest average pH value of 6.33 ± 0.13 whereas the Sour Brand 7 (SoV5) showed the lowest pH (5.28 ± 0.25) (Table 1). Interestingly, no significant difference was found for fat content of the analyzed yogurt samples. The average fat content of the samples varied from 0.25 to 2.52% (w/w). The Sweet Brand 1 (SwV1) had the highest fat content (2.52 ± 3.68%) while the lowest fat content (0.25 ± 0.05%) was recorded in Sour Brand 3 (SoV2) (Table 1).
Moisture content (MC) is one of the most common properties of food products due to a number of reasons like food quality, microbial stability along with legal and labeling requirements. The MC of the yogurt samples were in the range of 60.72 ± 0.80 to 87.59 ± 0.53% and varied significantly (P < 0.05) across the sample groups (Table 1). The average MC was much higher for all sweet Brands compared to sour Brands except Sweet Brand 4. The Sour Brand 3 had the highest MC (87.59 ± 0.53%) and lowest MC (60.72 ± 0.80%) was recorded in Sweet Brand 4 (SwV4). Likewise, the TS and SNF values differed significantly (P < 0.05) across the sample groups. We found the highest amount of TS (39.29 ± 0.80%) and SNF (38.48 ± 0.74%) in Sweet Brand 4 yogurt which remained lowest (TS; 12.41 ± 0.53% and SNF; 12.16 ± 0.51%) in Sour Brand 3 yogurt (Table 1). However, after removing the organic residues present in samples, Sweet Brand 4 and Sweet Brand 6 yogurt samples yielded the lowest ash content (0.80 ± 0.20%), while the highest amount of ash content (1.26 ± 0.31%) was found in Sour Brand 4 (SoV3) yogurt. No significant difference in ash content was found among the sample varieties.
Analysis of mineral contents
Yogurt is an important source of essential minerals. The contents of minerals including Na, K, Ca, Mg, Fe, Zn, and Cu in yogurt samples of different brands and tastes are given in Table 2. The mean value for Na content ranged from 593.50 ± 65.55 to 1052.65 ± 332.42 mg/kg indicating its highest and lowest content in Sour Brand 4 and Sour Brand 3, respectively. The Brand 4 of sour yogurt had the highest content of K (2744.54 ± 669.79 mg/kg), Ca (2442.27 ± 92.21 mg/kg), and Mg (272.42 ± 27 mg/kg). The lowest concentration of K (719.36 ± 135.50 mg/kg) was found in Sour Brand 5 (SoV4B), while the Sweet Brand 4 had the least mean content of Ca (1115.40 ± 354.57 mg/kg) and Mg (107.72 ± 20.05 mg/kg). Among these trace elements, the value of K varied within a broad range of 719.36 ± 135.5–2744.52 ± 669.79 mg/kg.
The concentration of Fe in yogurt varied from 2.74 ± 0.23 to 8.84 ± 1.20 mg/kg. The content of Fe in different brands of yogurt was statistically significant (P < 0.05) since the sweet yogurt always had higher mean values of Fe content except for Sour Brand 7 in which the Fe content was detected as 21.45 ± 1.23 mg/kg (Table 2). The concentration of Zn ranged from 3.91 ± 0.15 to 8.47 ± 0.32 mg/kg. The highest accumulation of Zn (8.47 ± 0.3 mg/kg) was noted in Sour Brand 7 followed by in sample Sweet Brand 6 (8.45 ± 2.03 mg/kg), while the least concentration (3.91 ± 0.15 mg/kg) was detected in Sweet Brand 4 yogurt. The amount of Cu varied within the range of 1.11 ± 0.75 to 4.87 ± 6.02 mg/kg. The maximum value was detected in Sour Brand 2 (SoV1) with the lowest content of Cu (1.11 ± 0.75 mg/kg) in Sour Brand 3. The extreme SD value with the maximum level of Cu indicates the concentration of Cu differed significantly (P < 0.05) among the brands of yogurt (Table 2). These results highlighted that mineral contents vary among different commercial brands (Brand 1–Brand 7) and taste types (sweet and sour) of yogurt. Although minerals represent a small portion of yogurt, they are fundamental for human health, yogurt quality and its characteristic tastes.
Microbiome composition and diversity
Since the microbiota composition is a key factor affecting the production and quality of fermented milk products, it is important and necessary to explore desirable microbes in order to manufacture high-quality yogurt1,2,3. Yogurt microbiomes of 30 samples (sweet = 15 and sour = 15) belonging to seven different commercial brands (Brand 1–Brand 7) were analyzed through high-throughput amplicon sequencing. The sweet yogurt samples included only cow samples (n = 15), and the sour yogurt included both cow (n = 12) and buffalo (n = 3) samples. During this study, the targeted sequencing approach generated a total of 3.10 million high quality reads (with an average of 0.104 million reads per sample), of which 1.4 and 1.7 million reads were assigned into 306 bacterial and 3144 fungal OTUs (Data S1). Among the reads, 44.86% and 55.14% reads were assigned to bacterial and fungal taxa, respectively. There is a clear difference between phylogenetic profiles and microbiota quantitation obtained using 16S rRNA (V3-V4) and ITS primers. An average good’s coverage index of 0.995 for bacteria and 0.996 for fungi indicated that sequencing depth was sufficient enough to capture most of the microbial community at different taxa levels (phylum, order, genus etc.) (Data S1).
The alpha–beta diversity of microbiomes was analyzed to observe the differences in microbial composition and diversity in yogurt. Figure 2 shows the alpha–beta diversity of bacterial and fungal communities in yogurt samples of different brands and tastes types (sweet and sour). Both alpha (within sample) and beta (between sample) diversities in yogurt samples of seven different brands (Brand 1–7) and two different tastes (sweet and sour) were estimated using different diversity indices (Fig. 2). Significant differences (P < 0.001; Kruskal–Wallis test) were found in α-diversity (observed richness, Shannon and Simpson estimated) in bacterial components of the microbiomes across the samples of both sweet and sour yogurts from cow and buffalo, and different brands (Fig. 2A,B). This diversity difference was evidenced significantly (P < 0.001; Kruskal–Wallis test) by higher taxonomic resolution in sour yogurt especially in Brand 5 (Fig. 2A). Though the observed species and Shannon estimated diversity of the fungal fraction of the yogurt microbiomes varied (P < 0.001; Kruskal–Wallis test) among different brands, it was less diverse between the taste types (sour and sweet) of yogurt (Fig. 2C,D). Moreover, the α-diversity of the fungal component of the yogurt microbiomes was found more diverse in sour yogurt compared to sweet yogurt samples (Fig. 2D).
Differences in microbiome diversity and community structure in seven different brands (Brand 1–7) and two different tastes (sweet and sour) of yogurt samples. (A–D) Alpha diversity was measured through the observed species richness, Simpson and Shannon estimated diversity indices. (A,B) α-diversity of bacterial communities and (C,D) α-diversity of fungal fraction of yogurt microbiomes according to different brands (A,C) and tastes (B,D). (E–H) Beta-diversity was estimated through PCoA and NMDS plots. (E, F) Principal coordinate analysis (PCoA) to measure the β-diversity in bacterial component of the yogurt microbiomes segregated yogurt samples according to (E) seven different brands and (F) two different tastes (sweet and sour); (G,H) Beta ordination (NMDS) plots representing the fungal component of the microbiomes separated yogurt samples according to (G) different brands and (H) two different tastes (sweet and sour) categories. Statistical analysis using Kruskal–Wallis tests showed significant microbial diversity variation (P < 0.01 = *, P < 0.001 = ***). The figures were generated using phyloseq (version 1.34.0)52 and Vegan (version 2.5-7)53 packages of R52.
The principal coordinate analysis (PCoA) plots representing β-diversity showed significant differences (P = 0.001, Kruskal–Wallis test) in bacteriome composition among different brands (Fig. 2E). Significant differences (P = 0.001, Kruskal–Wallis test) were also observed in bacterial diversity between sweet and sour yogurt sample groups (Fig. 2F), representing the prevalence of different sets of bacteria in each sample category. Similarly, NMDS estimated beta-ordination plots depicting the diversity of the fungal fraction of the yogurt microbiomes showed significant variations too in different brands (Ppermanova = 0.001) and tastes (Ppermanova = 0.013) (Fig. 2G,H). For instance, Brand 3 maintained the highest beta-dispersion in relation to Brand 6, Brand 5, Brand 2, and Brand 7, respectively (Table S1). According to taxonomic composition and differential taxonomic analysis, the sequence of the brands was: Brand 3 > Brand 2 > Brand 4 > Brand 1 > Brand 7 > Brand 5 > Brand 6. Least significant clustering differences were detected between Brand 1 and 6, and Brand 1 and 2 (Fig. 2G, Table S1).
Multivariate analysis and downstream bioinformatics
During data analyses, the detected OTUs were assigned into 11 phyla and 76 genera for bacteria, and 5 phyla and 70 genera for fungi. Among the bacterial phyla, 10 were detected in sweet yogurt of cow milk, 10 and 9 were in sour yogurt samples of cow and buffalo milk, respectively (Table S2). Firmicutes was the most abundant phylum with a relative abundance of 92.89% followed by Proteobacteria (7.03%) (Fig. S1A). By comparing the relative abundances of these predominant phyla, Firmicutes was found to be the most abundant phylum in both sweet (99.84%) and sour (86.36%) yogurt samples of cow milk. Conversely, Proteobacteria was the most abundant phylum in sour yogurt (61.46%) of buffalo milk followed by Firmicutes (38.12%) (Fig. S1A, Data S1). In addition, a total of 30 orders of bacteria were identified across these sample groups, of which 24, 25 and 21 orders were detected in sweet yogurt of cow, sour yogurt of cow and sour yogurt of buffalo, respectively (Table S2). Among the bacterial orders, Lactobacillales was found as the predominating order in sweet yogurt of cow (99.48%) and sour yogurt of cow (86.20%), while the sour yogurt of buffalo milk had higher abundance of Enterobacteriales (52.82%), Lactobacillales (37.48%) and Aeromonadales (5.52%) (Data S1).
The taxonomic classification and comparison of microbiomes at genus-level in yogurt samples of different tastes and milk sources are illustrated in Fig. 3. In this study, the yogurt samples collectively harbored 76 bacterial genera (Fig. 3A) and of them, Streptococcus (50.82%), Lactobacillus (39.92%), Enterobacter (4.85%), Lactococcus (2.84%) and Aeromonas (0.65%) were the top abundant genera (Data S1). In addition, notable differences were demonstrated in the diversity and composition of the identified bacterial genera both in sweet and sour yogurt samples of different brands. Among the detected bacterial genera, 36.84% genera were found to be shared in sweet yogurt of cow, sour yogurt of cow and sour yogurt of buffalo metagenomes (Table S2). The sweet and sour yogurt of cow had a sole association of 7 (9.2%) and 4 (5.26%) genera, respectively (Fig. 3A, Data S1). Likewise, 70 fungal genera were detected in the study samples, of which 22.86% genera were shared in the metagenomes of sweet yogurt cow, sour yogurt cow and sour yogurt buffalo (Fig. 3B; Table S2). Of the detected fungal genera in both sweet and sour yogurt samples, Kluyveromyces (65.75%), Trichosporon (8.21%), Clavispora (7.19%), Candida (6.71%), Iodophanus (2.22%), Apiotrichum (1.94%), and Issatchenkia (1.35%) were the most abundant genera (Data S1). The sweet and sour yogurt samples of cow had a sole association of 16 (22.86%) and 18 (25.71%) fungal genera. However, the sour yogurt samples of buffalo were observed to have no unique bacterial and fungal genera among the study metagenomes (Fig. 3B, Data S1).
Taxonomic composition of microbiomes in two different tastes (sweet and sour), and milk sources (cow and buffalo) of yogurt samples. (A) Venn diagram showing unique and shared bacterial genera in sweet (cow) and sour (cow and buffalo) yogurt. Out of 76 detected genera, only 28 (36.84%) genera (highlighted in red circle) were found to be shared in sweet yogurt cow, sour yogurt cow and sour yogurt buffalo metagenomes. (B) Venn diagram comparison of unique and shared fungal genera identified in the study metagenomes. Of the detected fungal genera (n = 70), only 16 (22.86%) genera (highlighted in red circle) were found to be shared in the metagenomes of sweet yogurt cow, sour yogurt cow and sour yogurt buffalo samples. More information on the taxonomic composition and relative abundances is available in Data S1. Venn diagrams were generated through an online tool used for Bioinformatics & Evolutionary Genomics (http://bioinformatics.psb.ugent.be/webtools/Venn/).
The taxonomic classification of bacteria and fungi was also analyzed at genus level in yogurt of various brands, tastes and milk source varieties (Fig. 4). Among the bacterial genera, Streptococcus (71.49%) and Lactobacillus (27.93%) had higher relative abundances in cow originated sweet yogurt samples (Fig. 4A; Data S1). On the other hand, sour yogurt samples of cow had higher relative abundances of Lactobacillus (50.34%), Streptococcus (30.32%), Enterobacter (9.28%), Lactococcus (5.41%) and Aeromonas (1.22%). Simultaneously, Streptococcus (42.43%), Macrococcus (24.76%), Enterobacter (11.88%), Empedobacter (9.76%), Aeromonas (5.54%) and Enterococcus (2.91%) were detected as the top abundant genera in sour yogurt of buffalo. The rest of the genera had relatively lower abundances (< 1.0%) in these metagenomes. Further investigation was performed to observe whether genus level relative abundances of the fungi differed between sweet and sour yogurt of different hosts (cow and buffalo) and brands (Fig. 4B). Ascomycota represented more than 98.0% of the total reads at genus level, except for Brand 5 and 6 that consisted of 59.8% of Basidiomycota (Fig. S1B). The sweet yogurt metagenome of cow origin had higher relative abundance of Kluyveromyces (78.45%), Clavispora (12.23%), and Candida (5.37%) while Kluyveromyces (47.62%), Trichosporon (19.74%), Candida (8.63%), Iodophanus (5.37%), Apiotrichum (4.70%), and Issatchenkia (2.71%) were the most abundant fungal genera in the sour yogurt metagenome of cow samples (Fig. 4B). Conversely, the sour yogurt sample of buffalo showed higher relative abundances of Trichosporon (70.14%), Iodophanus (12.07%), Apiotrichum (11.94%), and Neoascochyta (1.08%), and rest of the fungal genera detected in both sweet and sour yogurt metagenomes had lower relative abundances (< 1.0%) (Fig. 4B, Data S1).
The genus level taxonomic profile of bacteria and fungi in sweet yogurt of cow, sour yogurt of both cow and buffalo samples. The relative abundance of (A) 20 most abundant bacterial and (B) 25 most abundant fungal genera are sorted from bottom to top by their deceasing proportion with the remaining genera keeping as ‘other genera’. Each stacked bar plot represents the abundance of bacterial and fungal genera in each sample of the corresponding category. Notable differences in bacterial and fungal populations are those where the taxon is abundant in sweet yogurt samples, and effectively undetected in the sour yogurt samples. The distribution and relative abundance of the bacterial and fungal genera in the study metagenomes are also available in Data S1. The bar plots are generated using Microsoft Excel program.
The heatmap representation (Fig. 5) of the yogurt microbiomes shows distinct separation among the detected bacterial genera into two major clusters according to their host origin i. e., cow and buffalo rather than tastes of the yogurt. The heatmap of relative abundance of bacteria and fungi at genus level across the tastes and source varieties of yogurt is also presented in Fig. 5. The cow originated bacterial genera of the yogurt microbiomes were further sub-clustered according to tastes (sweet and sour) (Fig. 5A, Data S1). This clustering was consistent with the bar plot (Fig. 4A) representation of bacteria since Streptococcus and Lactobacillus were the top abundant bacterial genera in cow originated sweet yogurt samples (Fig. 5A). Likewise, Lactobacillus, Streptococcus, Enterobacter, Lactococcus and Aeromonas were the predominating genera in sour yogurt samples of cow origin, and Macrococcus, Enterobacter, Empedobacter, Aeromonas, Nitromonous, Stenotrophomonous, Kurthia Escherichia, Acinetobacter, Klebsiella, Shigella and Enterococcus were predominantly abundant in buffalo originated sour yogurt (Fig. 5A). In addition, similar to bacterial genera, the detected fungal genera showed two distinct clusters in the heatmap representation (Fig. 5B) according to their host origin (i. e. cow and buffalo) rather than tastes of the yogurt. The cow originated fungal genera of the yogurt microbiomes were further sub-clustered according to tastes (sweet and sour) (Fig. 5B). The taxonomic distinction and clustering of fungal genera (Fig. 5B) according to hosts and tastes also corroborated with the results of bar plot (Fig. 4B) representation of fungal taxa. Differential abundance of bacteria and fungi in yogurt samples of different brands and tastes is listed in Table 3. Differential abundance analysis showed that Aeromonas, Enterobacter, Lactococcus, and Acinetobacter had significantly (Bonferroni corrected PBC = 0.033–0.044, Bonferroni test) higher number of reads assigned for these genera in Brand 5 compared to other Brands (Table 3A). In addition, significantly (PBC = 0.043, 0.024) higher number of reads assigned for Streptococcus and Lactobacillus were found in Brand 6 and 7, respectively. Significantly (PBC = 0.028) greater number of Streptococcus associated reads (n = 4494.5) were observed in Brand 1 and this might be linked to the sweetness of yogurt brands (Table 3B).
Genus level taxonomic differentiation of bacteria and fungi in sweet yogurt (from cow milk) and sour yogurt of cow and buffalo milk. Heatmap showing 40 top abundant bacterial genera across the sample categories. The color bars (column Z score) at the top represent the relative abundance of each (A) bacterial genus and (B) fungal genus in the corresponding group. Noteworthy differences in bacterial and fungal populations are those where the genus is abundant in either sweet or sour yogurt samples, and effectively not detected in other metagenomes. The color codes indicate the presence and completeness of each genus in the corresponding sample group, expressed as a value between − 3 (lowest abundance) and 3 (highest abundance). The red color indicates the more abundant patterns, while green cells account for less abundant putative genes in that particular metagenome. The Heatmap is built through a stand-alone software tool; FunRich (http://www.funrich.org/).
A differential abundance analysis of the fungal genera showed that higher number of reads (n = 945.3) mapped to Apiotrichum in Brand 7 (Table 3A). Similarly, higher abundance of reads assigned for Clavispora in sweet brands and Apiotrichum in sour brands might reflect the differences in the taste of commercial yogurt (Table 3B). To investigate the relationships between relative abundance of microbial genera and presences of minerals in both sweet and sour yogurt, we used non-parametric Kruskal–Wallis test which revealed significant differences (P < 0.05) between the given conditions. The higher level of Na, K, Ca and Mg in the Sour Brand 4 whereas Cu in the Sour Brand 2 was observed in this research. Both of these sour Brands (Brand 2 and 4) were mostly dominated by a single genus of bacteria (Lactobacillus; > 97.0%) and fungus (Kluyveromyces; 86.23%) (Data S1). By contrast, Fe rich Sweet Brand 6 was mainly dominated by Streptococcus (89.54%) and Clavispora (71.0%). The Zn enriched Sweet Brand 6 and Sour Brand 7 were mostly dominated by both of these bacterial (Lactobacillus and Streptococcus), and fungal (Clavispora) genera (Data S1).
