Skip to main content
NCBI home page
Food Chemistry: X logoLink to Food Chemistry: X
. 2024 Mar 19;22:101311. doi: 10.1016/j.fochx.2024.101311

Characterization and correlation analysis of microbial flora and flavor profile of stinky acid, a Chinese traditional fermented condiment

Zhendong Zhang a, Huijun Zhao b,c, Renzhi Zhu a, Shaojing Cheng a, Yuanqi Yu a, Lan Xiang a, Zhipan Xiang a, Zhuang Guo b,c, Yurong Wang b,c,⁎
PMCID: PMC10978482  PMID: 38559445

Highlights

  • •

    Microbial diversity and sensory properties of stinky acid was investigated.

  • •

    Microbial loads exhibit negative correlations with the total acidity in stinky acid.

  • •

    Debaryomyces yeast was found to be positively related to the sour taste of stinky acid.

  • •

    Rhizopus, Thermoascus and Companilactobacillus were found to be implicated in the promotion of aromatic constituents.

Keywords: Stinky acid, High-throughput sequencing, Microbial diversity, Sensory quality, Correlation analysis

Abstract

To explore the microbial diversity and flavor profiles of stinky acid, we utilized high-throughput sequencing, culture-based techniques, and bionic E-sensory technologies. The results revealed a significant correlation between the acidity levels of stinky acid and the richness of its microbial community. Ten core bacterial genera and three core fungal genera exhibited ubiquity across all stinky acid samples. Through E-nose analysis, it was found that sulfides constituted the principal odor compounds responsible for stinky acid's distinct aroma. Further insights arose from the correlation analysis, indicating the potential contribution of Debaryomyces yeast to the sour taste profile. Meanwhile, three genera—Rhizopus and Thermoascus and Companilactobacillus—were identified as contributors to aromatic constituents. Interestingly, the findings indicated that Rhizopus and Thermoascus could reduce the intensity of the pungent odor of stinky acid. In summary, this investigation's outcomes offer new insights into the complex bacterial diversity of stinky acid.

1. Introduction

China boasts a diverse array of traditional fermented foods, many characterized by their distinctive stinky aromas. Notable examples include stinky tofu (also known as gray sufu), Yan-dong-gua (naturally fermented wax gourd), and Chouguiyu (fermented mandarin fish with a pungent odor) (Song et al., 2021, Song et al., 2021, Wang et al., 2021, Wu et al., 2016). Particularly in Guizhou Province, located in China's southern region, the well-known fermented food product “stinky acid” (Chousuan) is notably prominent. This fermented stinky acid is distinguished as one of the trios of renowned fermented sour foods in Dushan, including shrimp-acid, stinky-acid, and yan-acid (Xu et al., 2020). Stinky acid exhibits a unique and discernible malodorous characteristic, distinguishing it from other fermented foods. Despite its odor often being met with aversion, it possesses an exquisite taste relished by the local populace. Served as a condiment, stinky acid is used in preparing meat and tofu (bean curd) dishes, as well as hot pot. Predominantly popular in the south of Guizhou Province, stinky acid has garnered favor among locals (Wei et al., 2022).

Stinky acid, a solid fermented delicacy, is crafted using local traditional production methods (Xu et al., 2020). These methods involve utilizing raw materials, including impatiens (Impatiens balsamina L.), primary green vegetables (Brassica juncea L.), indigenous pungent litse fruit (Litsea pungens Hemsl.), bamboo shoots, and fried wheat. The initial steps involve washing, draining, and segmenting impatiens and green vegetables. Subsequently, the segmented raw materials and other ingredients, including chopped bamboo shoots, fried wheat, and pungent litse fruit, are meticulously arranged in specially designed pottery jars resembling pickle jars. These contents are then compacted in the jars. Finally, the pottery jars are hermetically sealed and placed in the shade to initiate a natural fermentation process lasting approximately 1 month (Xu et al., 2020). It is noteworthy that this fermentation occurs without the introduction of a starter culture. The final product exhibits a verdant hue and a consistency reminiscent of sauce. The specific ratio of raw materials may vary slightly depending on the maker's preferences. Thus, stinky acid distinctly differs from other stinky foods in both its production process and the use of raw materials (Song et al., 2021, Song et al., 2021, Wang et al., 2021, Wu et al., 2016). Furthermore, a natural fermentation approach that omits starter cultures and the unsterilized raw materials and pottery jars in the fermentation process prevent precise control over the final product's microbial composition, made the microbial composition in traditional food extremely complicated (Bassi et al., 2015; K. Zhang et al., 2023, Zhang et al., 2023). Overall, stinky acid significantly differs from other stinky foods in terms of production process, environmental factors, and raw materials used. This may lead to a unique microbial flora in stinky acid. In recent years, despite growing interest in stinky acid's unique flavor, there have been few studies focusing on its microbial diversity and the central microbial community.

Flavor is indeed one of the most significant characteristics of food, influencing consumer decision-making, and the distinctive flavor profiles are largely contributed to by microbial consortia (Chen et al., 2021, Sequino et al., 2022). The relationship between microorganisms and food flavor characteristics in certain foods has been established. Wang et al. demonstrated that Arcobacter, Psychrilyobacter, and Shewanella are pivotal in conferring characteristic flavors to Chouguiyu (Wang et al., 2021). Similarly, Song et al. used high-throughput sequencing and HPLC-based metabolite analysis to show that Fusobacterium, Providencia, Lactobacillus, and Bacillus could reduce propionic acid content in stinky sufu (Song et al., 2021, Song et al., 2021). Moreover, microbial metabolic network analysis identified Lactobacillus and Dekkera as significant flavor-producing microbes (Xia et al., 2021). Stinky acid, a fermented food from the south of Guizhou Province, China, is not widely accepted outside this area due to its unique odor. Despite the significant role of microbial communities in the fermentation process, the flavor characteristics of stinky acid and their underlying microbial interactions remain largely unexplored. This understanding would offer insights into the fermentation process and provides a foundation for harnessing these microorganisms to optimize the aroma of stinky acid.

Thus, this study aimed to analyze the flavor characteristics and microbial community profile of stinky acid. Furthermore, it aimed to identify the correlation between the flavor characteristics and the dominant microbial genera. Specifically, 42 representative samples of stinky acid were collected. Initially, the counts of bacteria, lactic acid bacteria (LAB), yeast, and molds were detected using culture-based technology, alongside the determination of pH and total acidity. Subsequently, microbial community profiles, including bacterial and fungal flora, were analyzed using high-throughput sequencing. Finally, flavor profiles were evaluated with E-tongue and E-nose sensors. The results of this study are expected to significantly enhance our understanding of microbial diversity, the relationship between functional species and pathways, and provide substantial data support for the production and quality improvement of stinky acid.

2. Materials and methods

2.1. Sampling

In February 2023, a study on the microbial diversity of stinky acid was conducted in Dushan County (107°41′-107°55′ E, 25°04′-25°31′ N), located in the Qiannan Buyi and Miao Autonomous Prefecture, Guizhou Province, China. A total of 42 stinky acid samples were collected from local agricultural markets, labeled CS01 to CS42. All samples were prepared using natural fermentation techniques, with similar raw materials (impatiens, green vegetables, bamboo shoots, pungent litse fruit, and fried wheat), and comparable processing steps. Approximately 150 g of each sample was preserved in duplicate in sterile containers with screw caps at 4℃. Upon arrival at the laboratory, one set of duplicates was immediately stored at −80℃ for amplicon sequencing. The other set was immediately allocated for other analyses.

2.2. Microbiological and physicochemical analysis

Microbe quantification followed the method described by Zheng et al. (Zheng et al., 2012). Specifically, the number of total aerobic bacteria, LAB, and yeast was evaluated using plate count agar (Hepebio, Qingdao, China), MRS agar (formulated with 1 % calcium carbonate, Hepebio, Qingdao, China) (De Man et al., 1960), and potato dextrose agar (PDA, Hepebio, Qingdao, China), respectively. For total aerobic bacteria and LAB enumeration, culture plates were incubated at 30 °C for 3–5 days. However, yeast quantification required incubation at 25 °C. MRS agar, crucial for LAB cultivation, was incubated in an anaerobic environment in a DG250 workstation (Don Whitley Scientific Ltd., Shipley, Yorkshire, UK). This anaerobic condition was achieved by infusing the workstation with a gas mixture of 85 % nitrogen, 10 % carbon dioxide, and 5 % hydrogen. In contrast, plate count and PDA agars were incubated under aerobic conditions.

The water content of stinky acid was assessed using the direct drying method, according to the GB 5009.3–2016 standard. Total acid content quantification followed the GB 5009.3–2016 guidelines. Additionally, pH was measured according to GB 5009.237–2016 using a PB-10 pH meter (Sartorius, Germany).

2.3. Identification of LAB isolates

Isolation of LAB strains followed the procedures detailed by Zhang et al., 2023, Zhang et al., 2023. Each stinky acid sample underwent 10-fold serial dilution to achieve a 10-6 dilution. Subsequently, dilutions from 10-2 to 10-6 were uniformly spread onto MRS plates supplemented with 1 % calcium carbonate. These plates were incubated at 30 °C in an anaerobic environment for 3–5 days. Colonies exhibiting a transparent halo were selected for purification and preserved at −80 °C. The preservation medium was MRS broth with 25 % glycerol.

Candidate LAB isolates were identified through 16S rDNA sequencing, following the workflow described by Zhang et al. (2019). First, total DNA was extracted using an E.Z.N.A.@Bacterial DNA Kit (Omega Bio-tek Inc., Norcross, GA, USA). The 16S rRNA gene was amplified with universal primers 27F and 1492R, using the extracted DNA as a template, as described by Osborne et al. (2005). The PCR products were then cloned into the pMD18-T vector (TaKaRa, Dalian, China), and used to transform DH5-α Escherichia coli cells. Positive clones were selected and forwarded to Shanghai Sunny Biotechnology Co., Ltd.

The 16S rRNA sequences obtained were used to conduct a Blastn search in the EzBiocloud database (Yoon et al., 2017). The isolates' 16S rRNA sequences were aligned with closely matched sequences from the EzBiocloud database. The sequences were used to construct a neighbor-joining phylogenetic tree with Mega X software for taxonomic classification (Kumar et al., 2018).

2.4. DNA extraction and high-throughput sequencing

To explore stinky acid's microbial diversity, metagenomic DNA was extracted and preserved at −80 °C. The 16S rRNA V3-V4 and ITS regions were selected for PCR amplification, followed by MiSeq sequencing. Metagenomic DNA from stinky acid samples was extracted using an E.Z.N.A.® Food DNA Kit, with DNA quality verified via a Nanodrop ND-2000C microUV spectrophotometer (Thermo Scientific, Wilmington DE, USA), following the manufacturer's guidelines.

For bacterial diversity examination, the 16S rRNA V3-V4 region was amplified with universal primers 338F and 806R, as noted by Wang et al. (2016). For fungal diversity assessment, universal primers ITS3 F and ITS4 R were used, following Kobayashi et al. (2011). PCR products from these primers were used to construct sequencing libraries. Sequencing was performed with the paired-end method on a HiSeq 4000 system at Majorbio Biotech Co., Ltd. (Beijing, China). Ultimately, sequencing was conducted using the paired-end method on the HiSeq 4000 system (Illumina, CA, USA) at Majorbio Biotech Co., Ltd. (Beijing, China). During sequencing, 150-bp reads were generated in both directions.

2.5. Bioinformatic and statistical analysis

To examine microbial diversity, raw sequencing data were processed with QIIME 1.9.0 (Caporaso et al., 2010). Initially, paired-end raw read data were filtered through TrimGalore v0.6.6 (Krueger, 2015). This process yielded high-quality reads, subsequently merged using FLASH (Magoč et al., 2011). Subsequently, potential chimera sequences were identified and removed using ChimeraSlayer on QIIME (Haas et al., 2011). The remaining sequences were clustered into OTUs based on a 97 % sequence identity threshold. For bacterial diversity analysis, taxonomic annotations utilized the RDP database v18 (Wang et al., 2007) based on representative OTU sequences. For fungal diversity, annotations used the UNITE database v9.0 (Nilsson et al., 2018). After annotation, alpha-diversity indices (Chao1, observed species, Shannon, and Simpson) were computed from the OTU table using QIIME scripts. Additionally, beta-diversity was assessed using QIIME's dedicated scripts.

2.6. Flavor evaluation of stinky acid samples

The sensory profile, including aroma and taste of the stinky acid samples, was assessed using a portable Electronic Nose (PEN3, Win Muster Airsense Analytics Inc., Schwerin, Germany) and a commercial Electronic Tongue (Taste-Sensing System SA 402B, Intelligent Sensor Technology Co. Ltd., Atsugi-shi, Kanagawa, Japan), as described by Cai et al. (2020) and Liu et al. (2023).

The E-nose's taste sensory indicators included six basic tastes: acidity, bitterness, astringency, saltiness, umami, and sweetness, along with three aftertastes: A (astringency), B (bitterness), and richness (umami), as identified by Wang et al. (2021).

Additionally, all ten E-nose sensors were utilized: W1C (aromatic organic compounds), W1S (methyl), W1W (sulfides), W2S (alcohols, aldehydes, and ketones), W2W (aromatic and sulfur compounds), W3C (ammonia and aromatics), W3S (long-chain alkanes), W5C (short-chain alkane aromatic components), W5S (broad sensitivity to nitrogen oxides), and W6S (mainly hydrides), as reported by Liu et al. (2023).

2.7. Statistical analysis

Statistical analyses were conducted in RStudio v4.2.1. Linear regression analysis utilized ggtrendline v4.2.3 (https://github.com/PhDMeiwp/ggtrendline) in R. Mantel tests, using the linkET package in R (Huang et al., 2021), evaluated correlations between microbial profiles and sensory indices. Pairwise correlations between microbial genera and sensory indices were assessed with the psych package v2.4.1 (https://cran.r-project.org/web/packages/psych/index.html) in R. Cluster analysis, employing the k-means method, was conducted with factoextra (https://github.com/kassambara/factoextra) in R. Permutational multivariate analysis of variance (PERMANOVA; Adonis, permutation = 999) was carried out using pairwiseAdonis v0.4 in R (Irnawati et al., 2020). Additionally, the Wilcoxon test (Mann-Whitney) determined the statistical significance of differences in dominant taxa (Divine et al., 2013). Graphical representations were created using ggplot2 and ggpubr v0.6.0 in R (Kassambara et al., 2020).

3. Results

3.1. Physicochemical properties and culture-based microbial counts

Forty-two samples were collected, each exhibiting a green color, sticky texture (Fig. 1a), and distinct stinky smell. To assess microbial richness, microbial counts were conducted (Fig. 1b). Counts of total aerobic bacteria (log cfu/ml) ranged from 4 to 7.78, with LAB and yeasts ranging from 2.32 to 7.24 and 4.18 to 7.74, respectively.

Fig. 1.

Fig. 1

Open in a new tab

Images showing the appearance of the collected stinky acid samples (a), culture-based microbial counts (b), physicochemical properties (c), and the linear regression relationship between the physicochemical properties and microbial counts (d). R2 and p values of the linear regression analyses are shown. The black solid line indicates a significant linear relationship, with p < 0.05; gray areas represent the 95 % confidence intervals. Each black dot indicates a stinky acid sample. *: indicates p < 0.05, **: indicates p < 0.01, and ***: indicates p < 0.001.

Subsequently, pH, total acid content, and water content were measured (Fig. 1c). pH levels spanned from approximately 3.48 to 5.98, and total acid content varied from 2.63 to 36.19 g/L. Water content ranged from 52.08 to 95.17 g/100 g in the samples. These findings highlighted significant variability in physicochemical parameters and culture-based microbial richness among the samples.

Linear regression analysis provided deeper insights into the interplay between physicochemical properties and microbial richness (Fig. 1d). Results revealed a significant pattern: as pH increased, counts of total aerobic bacteria, LAB, and yeasts also increased (p < 0.01). Conversely, counts of total aerobic bacteria, LAB, and yeasts decreased as total acid content increased (p < 0.01). It's important to note pH represents the negative logarithm of hydrogen ion (H+) concentration. This observation highlights contrasting trends in the relationship between pH, total acid, and microbial counts.

3.2. Microbial diversity and components

3.2.1. Bacterial diversity and community structure

Amplicon sequencing of a partial 16S rRNA gene was conducted to explore the bacterial diversity in stinky acid samples. High-throughput sequencing of the 42 stinky acid samples yielded a total of 1,539,932 high-quality sequences, with an average of 38,276 sequences per sample (Fig. S1). These sequences were classified into 32,565 OTUs, averaging 2,989 OTUs per sample (Fig. S1). These OTUs received annotations based on the latest RDP database (Fig. S1). These OTUs were annotated based on the latest RDP database (Fig. S1). In total, the 32,565 OTUs were annotated across 14 phyla, 26 classes, and 240 genera.

Subsequently, we calculated α-diversity indices, as presented in Fig. S1, based on a sequence depth of 21,010 sequences. To evaluate the adequacy of the sequencing depth, we analyzed rarefaction curves for the Shannon index (Fig. S2a) and the observed species index (Fig. S2b). The results confirmed that a sequencing depth of 21,010 was sufficient to capture the bacterial diversity in the 42 stinky acid samples.

Next, relative abundance of the phyla and genera was extracted to evaluate the microbial composition of stinky acid samples. In this study, taxa having relative abundances exceeding 1 % were deemed dominant. At the phylum level, five bacterial phyla showed average relative abundances exceeding 1 % (Fig. 2a), including Firmicutes (49.75 %), Proteobacteria (25.68 %), Bacteroidetes (15.29 %), Actinobacteria (5.95 %), and Campylobacterota (1.93 %). Among these, Firmicutes was identified as the most predominant.

Fig. 2.

Fig. 2

Open in a new tab

Dominant bacterial phyla (a) and genera (b), and the relationships between the pH and total acid levels and the dominant bacterial genera (c). Fig. 2c: R2 and p values of the linear regression analyses are shown; each black dot indicates a stinky acid sample. The black solid line indicates a significant linear relationship, with p < 0.05; gray areas represent the 95 % confidence intervals. Fig. 3a and 3b: Taxa with relative abundances < 1 % are summarized as “other”.

At the genus level, twenty-four genera were found with average relative abundances surpassing 1 % (Fig. 2b), including Lactobacillus (16.57 %), Prevotella (10.08 %), Halomonas (5.99 %), Corynebacterium (5.30 %), Caproicibacter (3.05 %), Ignatzschineria (3.04 %), Pediococcus (3.00 %), Lapidilactobacillus (2.89 %), Companilactobacillus (2.72 %), Levilactobacillus (2.47 %), Pandoraea (1.91 %), Advenella (1.75 %), Halocella (1.56 %), Pseudomonas (1.48 %), Aliarcobacter (1.42 %), Ligilactobacillus (1.37 %), Bacteroides (1.36 %), Clostridium_sensu_stricto (1.24 %), Bacillus (1.18 %), Marinobacter (1.14 %), Halanaerobium (1.13 %), Fermentimonas (1.09 %), Acinetobacter (1.04 %), and Loigolactobacillus (1.03 %) (Fig. 2b). Notably, genera that potentially belonged to the LAB group accounted for an approximate cumulative relative abundance of 30.04 %.

Additionally, ten genera consistently occurred across all samples. These genera were identified as the core dominant genera, including: Lactobacillus, Prevotella, Corynebacterium, Caproicibacter, Levilactobacillus, Pseudomonas, Ligilactobacillus, Bacteroides, Clostridium_sensu_stricto, and Loigolactobacillus.

An analysis was conducted to delineate the relationship between dominant bacterial genera and the pH and total acid content of stinky acid (Fig. 2c). Through linear regression analysis, distinct trends were identified: an increase in the relative abundance of Lapidilactobacillus and Loigolactobacillus was observed with a higher stinky acid pH (p < 0.05). Conversely, a decrease in the relative abundance of Pandoraea was observed as the pH increased (p < 0.05). Furthermore, with an increase in the total acid content of stinky acid, a decrease in the relative abundance of Lapidilactobacillus and Loigolactobacillus was observed, whereas an increase in that of Pandoraea was observed (p < 0.05).

Through linear regression analysis, we recognized distinct trends: the relative abundance of Lapidilactobacillus and Loigolactobacillus increased with increasing stinky acid pH (p < 0.05). Conversely, the relative abundance of Pandoraea decreased as the pH increased (p < 0.05). Furthermore, the relative abundance of Lapidilactobacillus and Loigolactobacillus decreased with an increase in the total acid content of stinky acid, whereas that of Pandoraea increased (p < 0.05).

3.2.2. Fungal diversity and community structure

Based on high-throughput sequencing of the fungal ITS region, a total of 2,082,275 high-quality sequences was obtained, with an average of 49,578 sequences per sample (Fig. S1). These sequences were classified into 2,606 OTUs using a 97 % similarity threshold (Fig. S1). Annotation analysis identified the presence of 4 fungal phyla, 12 fungal classes, and 38 fungal genera across the stinky acid samples (Fig. S1). Although there were fewer fungal than bacterial taxa, the difference was not statistically significant (Fig. S1).

Furthermore, all four α-diversity indices of the fungal community were found to be lower than those of the bacterial α-diversity indices (Fig. S1). The Shannon and observed species index curves, shown in Fig. S3a and S3b, tended to plateau, suggesting that the selected sequencing depth was appropriate for capturing the fungal diversity present in the stinky acid samples.

The fungal phyla whose relative abundances exceeded 1 % were Ascomycota (94.58 %), Basidiomycota (3.01 %), and Mucoromycota (2.18 %, Fig. 3a). At the genus level, seventeen genera displayed relative abundances exceeding 1 %: Pichia (16.42 %), Debaryomyces (15.12 %), Dipodascus (11.64 %), Saccharomyces (9.47 %), Yarrowia (6.94 %), Saccharomycopsis (5.14 %), Candida (4.17 %), Dekkera (3.80 %), Wickerhamomyces (2.56 %), Aspergillus (2.53 %), Wallemia (1.87 %), Scedosporium (1.61 %), Penicillium (1.60 %), Rhizopus (1.53 %), Trichosporonoides (1.36 %), Thermoascus (1.20 %), and Mucor (1.09 %, Fig. 3b). Among these genera, three core genera consistently occurred across all samples, namely Pichia, Dipodascus, and Thermoascus. Analysis of the dominant genera revealed significant variability in the fungal community structure of the stinky acid samples.

Fig. 3.

Fig. 3

Open in a new tab

Dominant fungal phyla (a) and genera (b), and relationships between pH and the dominant fungal genera (c). Fig. 3c: R2 and p values of the linear regression analyses are shown; each black dot indicates a stinky acid sample. The black solid line indicates a significant linear relationship, with p < 0.05; gray areas represent the 95 % confidence intervals. Fig. 3a and 3b: Taxa with relative abundances < 1 % are summarized as “other”.

Regression analysis revealed clear trends: an increase in the relative abundance of Dipodascus was observed with a higher pH of the stinky acid, while it decreased with an increase in total acid content (Fig. 3c).

3.3. LAB isolates from stinky acid

To assess the LAB diversity within stinky acid, culture-based techniques were employed to isolate LAB strains, which were then identified through 16S rRNA sequencing analysis. Seventy-three strains were successfully isolated. Of these strains, only three exhibited a coccus morphology, whereas the rest were rod-shaped (Table S1). A hybrid approach involving blastn and phylogenetic tree analysis was employed to achieve a more detailed classification of these strains (Fig. 4).

Fig. 4.

Fig. 4

Open in a new tab

Neighbor-joining (NJ) phylogenetic tree of LAB strains from stinky acid and their closest type material based on 16S rRNA gene sequences. The NJ tree was constructed with MEGA X using 1000 bootstrap replicates. Alignment of the 16S rRNA gene sequences was performed using Mafft v7.475. The accession numbers of 16S rRNA sequences used to construct NJ tree were provided in square brackets.

Seventy-three strains identified in the study were classified into eight bacterial genera, including Loigolactobacillus, Companilactobacillus, Lacticaseibacillus, Lactiplantibacillus, Levilactobacillus, Ligilactobacillus, and Pediococcus. Notably, Loigolactobacillus was the predominant genus, with 34 strains of Loigolactobacillus coryniformis identified, constituting 46.57 %. Included within the genus Levilactobacillus were five species: Levilactobacillus angrenensis (one strain, constituting 1.28 %), Levilactobacillus brevis (three strains, constituting 2.56 %), Levilactobacillus tujiorum (three strains, constituting 4.11 %), Levilactobacillus yiduensis (one strain, constituting 1.28 %), and Levilactobacillus sp. (two strains, constituting 8.22 %). Furthermore, within the genus Ligilactobacillus, one species, Ligilactobacillus acidipiscis, was identified, comprising eight strains and constituting 10.96 %. Within the genus Companilactobacillus, three species were identified, namely Companilactobacillus alimentarius (five strains, constituting 6.41 %), Companilactobacillus nuruki (one strain, constituting 1.28 %), and Companilactobacillus baiquanensis (one strain, constituting 1.28 %). The remaining genera each harbored fewer than five strains within their respective classifications.

Although isolates of two prominent LAB genera, Lactobacillus and Lapidilactobacillus, were identified through high-throughput sequencing, they were not detected via the culture-based approach. Additionally, the culture-based method effectively detected three genera: Lacticaseibacillus, Lactiplantibacillus, and Lentilactobacillus, each with a relative abundance of less than 1 % in high-throughput sequencing.

3.4. Flavor evaluation

3.4.1. Taste evaluation

An electronic tongue (E-tongue) system was utilized to evaluate the taste profile of the stinky acid samples (Fig. 5a). The sensor response values for bitterness, astringency, umami, aftertaste A, aftertaste B, and richness showed relatively minor variations, indicating a consistent pattern. In contrast, response values associated with sourness and saltiness demonstrated consistently greater variability. A widespread distribution of response values from the acidity sensor indicated a highly variable perception of sourness. This variability could likely be attributed to fluctuations in the microbial community structure during the natural fermentation process.

Fig. 5.

Fig. 5

Open in a new tab

Box plot for taste properties (a), and correlation analysis between E-tongue indices and the overall microbial groups (b), dominant bacterial genera (c), and dominant fungal genera (d). *: indicates p < 0.05; **: indicates p < 0.01, ***: indicates p < 0.001. Corr. = correlation coefficient.

Pearson’s correlation analysis was used to explore the relationships between E-tongue sensor response values and the dominant bacterial and fungal genera. The overall dominant bacterial communities showed substantial correlations with sensory attributes, including sourness, astringency, umami, saltiness, and the aftertaste of umami (p < 0.05), as confirmed through the Mantel test. Furthermore, significant correlations were observed between the dominant fungal communities and sensory attributes like sourness, saltiness, and umami (p < 0.05, Fig. 5b).

Notably, the findings revealed specific associations. The undesired flavor dimension of bitterness showed a positive correlation (p < 0.05) with the bacterial genus Companilactobacillus and the fungal genera Thermoascus and Rhizopus. Similarly, aftertaste B, linked to bitterness, demonstrated a positive correlation with Companilactobacillus and Rhizopus (p < 0.05, Figs. 5c and 5d). Another unfavorable flavor component, astringency, exhibited a positive correlation (p < 0.05) with the bacterial genera Halomonas and Marinobacter, and the fungal genus Aspergillus. Furthermore, saltiness showed a negative correlation (p < 0.05) with the LAB genera Loigolactobacillus and Levilactobacillus. Of particular significance, the characteristic sour taste demonstrated a positive correlation (p < 0.05) with Debaryomyces and a negative correlation with Dipodascus.

3.4.2. Odor evaluation

Each stinky acid sample underwent E-nose simulated sensory analysis (Fig. 6a). As shown in Fig. 6a, most sensor response values were minimal. Conversely, sensors W1W and WS (W1S, W2S, W5S) showed notably higher response values, indicating elevated concentrations of aromatic constituents such as sulfides, methyl compounds, alcohols, aldehydes, ketones, and nitrogen oxides in the samples. Among these, sulfides, potentially responsible for the characteristic odor profile, emerged as the most prominent contributor detected by W1W. Dominant bacterial communities showed a highly significant correlation with WS (W5S, W2S) response values (p < 0.01) and significant correlations with WC (W1C, W3C, W5C), W1W, and W6S (p < 0.05). Fungal communities showed significant correlations with W1S and W2S sensor response values (p < 0.05) (Fig. 6b).

Fig. 6.

Fig. 6

Open in a new tab

Box plot for odor properties (a), and correlation analysis between E-nose indices and the overall microbial groups (b), dominant bacterial genera (c), and dominant fungal genera (d). *: indicates p < 0.05; **: indicates p < 0.01, ***: indicates p < 0.001. Corr. = correlation coefficient.

Subsequently, focus was on individual dominant genera. Pediococcus, along with Rhizopus and Thermoascus, showed a significant negative correlation with WS (W1S, W2S, W5S) sensor response values (p < 0.05, Fig. 6c and 6d). Ignatzschineria and Ligilactobacillus demonstrated a positive, significant correlation with the W5S sensor response value (p < 0.05). Notably, only Rhizopus and Thermoascus had significant correlations with the W1W sensor response values (p < 0.05). In contrast, no dominant bacterial genera showed significant correlations with the W1W sensor response value (p > 0.05).

3.5. Differences between types of stinky acid

The E-sensory profile of stinky acid showed significant variability in response values for attributes including sourness, saltiness, W1W response, and W5S response, among others. Using the k-means method based on Bray distances of E-sensory indices, the 42 stinky acid samples were successfully clustered into three distinct types (Fig. 7a and 7b). Bray–Curtis distances confirmed highly significant differences among the three types of stinky acid samples (p < 0.001). Random forest analysis and the Wilcoxon test were employed to determine the characteristics of the three types of stinky acid (Fig. 7c, 7d, and 7e). Type 1 stinky acid showed the highest response values for W5S, W1W, W1S, and W2S sensors, and the lowest for WC sensors (W1C, W3C, and W5C) (p < 0.01). Conversely, type 3 showed the highest response values for WC sensors (W1C, W3C, and W5C). Type 2 exhibited the highest relative abundance of Pandoraea and Wickerhamomyces (p < 0.01).

Fig. 7.

Fig. 7

Open in a new tab

Cluster analysis of stinky acid, and biomarkers for the pairwise discrimination of the three types of stinky acid. a, b Cluster analysis of stinky acid based on E-sensory indices using the k-means method. c Classification performance of the top 10 discriminant E-sensory indices using a random forest model and heatmap based on the response values of these indices. d Classification performance of the 10 most discriminant bacterial genera using a random forest model and heatmap based on the relative abundance of the bacterial genera. e Classification performance of the 10 most discriminant fungal genera using a random forest model and heatmap based on the relative abundance of these fungal genera. f Receiver operating characteristic (ROC) curve analysis for the E-nose marker W5S identified based on random forest analysis. g ROC analysis for the bacterial genus marker Pandoraea identified based on random forest analysis. h ROC analysis for the fungal genus marker Wickerhamomyces identified based on random forest analysis. Rel. Abu. = Relative abundance; Res. Val. = Response values.

ROC analysis and leave-one-out cross-validation demonstrated W1W's sufficient discriminatory power for stinky acid samples. However, in the absence of sensory indices, a single microbial genus proved insufficient to differentiate among the three types of stinky acid (Fig. 7f, 7g, and 7h). Pandoraea and Bacteroides were identified as key markers, offering enhanced discriminatory power for pairwise differentiation among type 1, type 2, and type 3 stinky acids.

4. Discussion

The traditional fermented condiment stinky acid originates primarily from the southern region of Guizhou Province. Its unique microbial diversity is likely shaped by the specific environmental conditions and raw materials of this area, distinguishing it from other fermented foods. To investigate the microbial profile and flavor properties of stinky acid comprehensively, a combination of high-throughput sequencing, culture-based methods, and bionic E-tongue and E-nose analyses was employed.

Numerous traditional fermented foods from specific regions possess unique odors, including grey sufu, Chouguiyu, and Yan-dong-gua (fermented wax gourd). Acinetobacter and Enterococcus, consistently detected throughout grey sufu's fermentation process, are considered resident bacteria. Furthermore, the dominant bacterial genus Pseudomonas has been linked to the characteristic stinky aroma of grey sufu (Ding et al., 2023). Chouguiyu is characterized by the dominance of Vibrio, Psychrobacter, Psychrilyobacter, and Fusobacterium, with Psychrilyobacter notably influencing flavor formation (Yang et al., 2021). While grey sufu and Chouguiyu fermentation transform proteins into amino acids for a unique flavor, stinky acid distinguishes itself by using impatiens as a primary raw material. This study's high-throughput sequencing revealed 24 dominant bacterial and 14 dominant fungal genera in diverse stinky acid samples. A comparison with other fermented foods revealed commonalities and differences. Compared to grey sufu, Chouguiyu, and Yan-dong-gua, Advenella, Aliarcobacter, Caproicibacter, and Halocella are uniquely dominant only in stinky acid, highlighting its unique microbial community (Ding et al., 2023, Lan et al., 2009, Yang et al., 2021). Stinky acid and grey sufu share dominant genera including Pseudomonas, Pediococcus, Bacillus, and Lactobacillus, albeit with varying levels of abundance (Ding et al., 2023), whereas, stinky acid, Yan-dong-gua and Chouguiyu only exhibit a dominant abundance of Lactobacillus (Lan et al., 2009, Yang et al., 2021). The dominance of the LAB genus Lactobacillus in all four stinky fermented foods suggests a common role in enhancing sourness. In addition, the LAB group accounted for an approximate cumulative relative abundance of 30.04 %. LAB is also found in pickled vegetables, zha-chili, and sour soup (Guo et al., 2021, Cai et al., 2021, Zhang et al., 2023, Zhang et al., 2023, Lin et al., 2022). This group of bacteria, characterized by unique fermentative behavior yielding lactic acid as the primary end product, influences both sour taste and aroma development. In this study, correlation analysis was used to speculate on the influence of certain lactic acid bacteria, including Lactobacillus, on the flavor of stinky acid.

Factors contributing to microbial diversity include pH, salt content, and temperature (Shu et al., 2022). Our study measured pH and total acid across 42 samples, prompted by stinky acid's characteristic sour taste. Simultaneously, culture-based microbial counts were conducted, including total aerobic bacteria, LAB, and yeast. Regression analysis showed a positive correlation between pH and microbial richness (p < 0.05), and a negative correlation between total acid content and microbial richness. This highlighted the role of acidity as a potential constraint on microbial counts. Interestingly, regression analysis identified significant correlations between acidity and specific taxa: increased acidity was associated with decreased relative abundance of Lapidilactobacillus, Loigolactobacillus, and Dipodascus (p < 0.05). Notably, the dominant bacterial genus Pandoraea showed a positive correlation with acidity (p < 0.05), indicating notable resistance to high acidity. This is significant as Pandoraea spp. typically thrive in liquid mediums at pH levels as low as 4. This suggests Pandoraea spp.'s potential adaptation in stinky acid, highlighting their resilience to low pH (Peeters et al., 2019). Through LAB isolation and identification, stinky acid-dominant LAB Loigolactobacillus was identified as Loigolactobacillus coryniformis, listed in the 2019 QPS update, underscoring its potential food production applications.

Although stinky acid is known for its distinctive smell, it notably offers an appealing taste as a condiment. Despite sourness being a characteristic flavor of stinky acid, considerable variability in sourness exists among samples produced through natural fermentation. This study's correlation analysis revealed important links between specific microbes and stinky acid's sensory attributes. Specifically, the dominant yeast Debaryomyces was positively correlated with sourness, aligning with findings that Debaryomyces spp. contribute to acid production (Santos de Almeida et al., 2023). This is likely due to their ability to convert sugars into organic acids via glycolysis, the TCA cycle, and the pentose phosphate pathway. Additionally, Debaryomyces spp. have been shown to enhance the flavor of cheese and fermented sausages (Gori et al., 2012, Santos de Almeida et al., 2023). Consequently, Debaryomyces spp. could be utilized to enhance stinky acid's sour taste. Interestingly, the dominant LAB Loigolactobacillus showed a negative correlation with saltiness, suggesting its utility in reducing stinky acid's salty taste.

Defective flavors negatively impact food quality. Correlation analysis demonstrated that Companilactobacillus, Thermoascus, and Rhizopus were positively associated with bitterness and its aftertaste. Furthermore, Halomonas, Marinobacter, and Aspergillus were positively correlated with the undesirable taste of astringency. Therefore, these microorganisms may contribute to the development of unfavorable tastes in stinky acid. Despite the identification of numerous taxa potentially promoting off-flavors, it is noteworthy that, minimal variations were observed in terms of bitterness, aftertaste B, and astringency across different samples. Remarkably, local consumers scarcely perceive the bitterness and astringency of stinky acid, suggesting that these off-flavors may not be the primary determinants affecting product quality.

Stinky acid's prominent characteristic is its distinct odor, largely influenced by microbial fermentation and various metabolic pathways. Research suggests sulfides are the characteristic volatile compounds responsible for the distinctive odors. Liu et al. illustrated that fermented stinky tofu contains sulfide variants and indole, contributing to its unfavorable aroma (Liu et al., 2012). Additionally, dimethyl disulfide and methyl thiopropionaldehyde were detected in Yan-gui-yu (Wu et al., 2019, Zhou et al., 2019). Sulfide compounds, with low odor thresholds and strong flavor profiles, are key contributors to the aromas of products like cheeses, beer, and baijiu. Our study used E-nose sensors to assess stinky acid's flavor characteristics, showing the highest response for W1W, indicative of sulfides, confirming their role as distinctive volatile flavor constituents in stinky acid. Interestingly, correlation analysis showed a negative association between the fungal genera Rhizopus and Thermoascus and the W1W response value (p < 0.001), suggesting these fungi may reduce sulfide content, thus diminishing stinky acid's odor. Furthermore, Pediococcus, Rhizopus, and Thermoascus were found to enhance aromatic constituents detectable by WC sensors. Additionally, Companilactobacillus bacteria were associated with enhancing aromatic constituents. Reclassified from Lactobacillus in 2021, Companilactobacillus has been detected in fermented foods like zha-chili (Song et al., 2021, Song et al., 2021) and pickles (Shang et al., 2022), and the human gut (Zhang et al., 2021). In this study, Companilactobacillus primarily consisted of Companilactobacillus alimentarius. Therefore, beyond influencing off-flavors, Companilactobacillus, Rhizopus, and Thermoascus could enhance stinky acid's aroma and reduce its characteristic odor, improving palatability for more consumers.

Microorganisms play a pivotal role in influencing the flavor of fermented foods. This study observed considerable variability in stinky acid's microbial communities, including both bacterial and fungal. The variability can be attributed to using unsterilized raw materials, pottery jars, and the absence of starter cultures. Consequently, significant differences in the E-sensory indices of stinky acid samples were observed. Using the E-sensory indices, the collected stinky acid samples were classified into three distinct types through the k-means method. Sensory indicators like W1W, capable of detecting sulfides and nitrogen oxides, effectively discriminated these three stinky acid types. Although individual microbial genera may lack sufficient discriminatory capability, combining Pandoraea and Bacteroides improves discriminatory power for differentiating among stinky acid types 1, 2, and 3.

5. Conclusion

This study delivers a thorough examination of the microbial diversity and flavor characteristics of stinky acid, a unique fermented condiment from Guizhou Province, shedding light on the complex interplay between these aspects. The research demonstrates that the microbial loads of lactic acid bacteria (LAB), total aerobic bacteria, and yeast exhibit negative correlations with the acidity in stinky acid. Despite considerable variability in the microbial communities of stinky acid, ten core dominant bacterial genera, including Lactobacillus, Prevotella, Corynebacterium, among others, and core fungal genera such as Pichia, Dipodascus, and Thermoascus, were identified. The research further establishes a link between these dominant microbes and specific flavor attributes, identifying sulfides as critical volatile compounds shaping the unique odor of stinky acid. Notably, the fungi Rhizopus and Thermoascus were implicated in moderating this distinct aroma, while also, along with Companilactobacillus, augmenting aromatic qualities. Additionally, Debaryomyces spp. emerged as a significant contributor to the sourness profile, underscoring its role in flavor modulation. This investigation not only deepens the understanding of stinky acid's fermentation ecology but also opens avenues for refining its sensory appeal through microbial management.

Data Statement

The high-throughput sequencing data can be downloaded from the NCBI SRA database under accession number SRR26079047-SRR26079088. The 16S rRNA gene sequences for the LAB strain can be obtained from NCBI GenBank using the accession number OR563707-OR563779.

CRediT authorship contribution statement

Zhendong Zhang: Writing – review & editing, Writing – original draft, Visualization, Funding acquisition. Huijun Zhao: Supervision, Investigation. Renzhi Zhu: Software, Resources, Methodology. Shaojing Cheng: Supervision, Resources, Methodology. Yuanqi Yu: Validation, Supervision, Resources. Lan Xiang: Formal analysis, Data curation. Zhipan Xiang: Software, Formal analysis. Zhuang Guo: Visualization, Funding acquisition, Data curation. Yurong Wang: Writing – review & editing, Visualization, Methodology, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This research was supported by the Hubei Provincial Natural Science Foundation Joint Fund for Innovation and Development Project (2022CFD008), the Qiannan Normal University for Nationalities Supports the Special Project of Introducing High-level Talents (qnsyrc2022011) and Hubei University of Arts and Science Cultivation Fund for Teachers' Scientific Research Ability: Technological Innovation Team (2020kypytd009).

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochx.2024.101311.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary data 1
mmc1.pdf (228.2KB, pdf)
Supplementary data 2
mmc2.pdf (121.4KB, pdf)
Supplementary data 3
mmc3.pdf (135.2KB, pdf)
Supplementary data 4
mmc4.xlsx (12.7KB, xlsx)

Data availability

Data will be made available on request.

References

  1. Bassi D., Puglisi E., Cocconcelli P.S. Comparing natural and selected starter cultures in meat and cheese fermentations. Current Opinion in Food Science. 2015;2:118–122. doi: 10.1016/j.cofs.2015.03.002. [DOI] [Google Scholar]
  2. Cai W., Tang F., Guo Z., Guo X., Zhang Q., Zhao X., Ning M., Shan C. Effects of pretreatment methods and leaching methods on jujube wine quality detected by electronic senses and HS-SPME–GC–MS. Food Chemistry. 2020;330 doi: 10.1016/j.foodchem.2020.127330. [DOI] [PubMed] [Google Scholar]
  3. Cai W., Wang Y., Hou Q., Zhang Z., Tang F., Shan C., Yang X., Guo Z. Rice varieties affect bacterial diversity, flavor, and metabolites of zha-chili. Food Research International. 2021;147 doi: 10.1016/j.foodres.2021.110556. [DOI] [PubMed] [Google Scholar]
  4. Caporaso J.G., Kuczynski J., Stombaugh J., Bittinger K., Bushman F.D., Costello E.K., Fierer N., Peña A.G., Goodrich J.K., Gordon J.I., Huttley G.A., Kelley S.T., Knights D., Koenig J.E., Ley R.E., Lozupone C.A., McDonald D., Muegge B.D., Pirrung M., Reeder J., Sevinsky J.R., Turnbaugh P.J., Walters W.A., Widmann J., Yatsunenko T., Zaneveld J., Knight R. QIIME allows analysis of high-throughput community sequencing data. Nature Methods. 2010;7(5):335–336. doi: 10.1038/nmeth.f.303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chen G., Huang Z., Wu L., Wu Q., Guo W., Zhao W., Liu B., Zhang W., Rao P., Lv X., Ni L., Sun J., Sun B. Microbial diversity and flavor of chinese rice wine (huangjiu): An overview of current research and future prospects. Current Opinion in Food Science. 2021;42:37–50. doi: 10.1016/j.cofs.2021.02.017. [DOI] [Google Scholar]
  6. Ding S., Tian M., Yang L., Pan Y., Suo L., Zhu X., Ren D., Yu H. Diversity and dynamics of microbial population during fermentation of gray sufu and their correlation with quality characteristics. LWT. 2023;180 doi: 10.1016/j.lwt.2023.114711. [DOI] [Google Scholar]
  7. Divine G., Norton H.J., Hunt R., Dienemann J. A review of analysis and sample size calculation considerations for Wilcoxon tests. Anesthesia & Analgesia. 2013;117(3):699–710. doi: 10.1213/ANE.0b013e31827f53d7. [DOI] [PubMed] [Google Scholar]
  8. De Man J.C., Rogosa D., Sharpe M.E. A medium for the cultivation of lactobacilli. Journal of applied microbiology. 1960;23(1):130–135. doi: 10.1111/j.1365-2672.1960.tb00188.x. [DOI] [Google Scholar]
  9. Gori K., Sørensen L.M., Petersen M.A., Jespersen L., Arneborg N. Debaryomyces hansenii strains differ in their production of flavor compounds in a cheese-surface model. MicrobiologyOpen. 2012;1(2):161–168. doi: 10.1002/mbo3.11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Guo Z., Wang Y., Xiang F., Hou Q., Zhang Z. Bacterial diversity in pickled cowpea (Vigna unguiculata [Linn.] Walp) as determined by illumina MiSeq sequencing and culture-dependent methods. Current Microbiology. 2021;78:1286–1297. doi: 10.1007/s00284-021-02382-3. [DOI] [PubMed] [Google Scholar]
  11. Haas B.J., Gevers D., Earl A.M., Feldgarden M., Ward D.V., Giannoukos G., Ciulla D., Tabbaa D., Highlander S.K., Sodergren E. Chimeric 16S rRNA sequence formation and detection in sanger and 454-pyrosequenced PCR amplicons. Genome research. 2011;21(3):494–504. doi: 10.1101/gr.112730.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Huang, H. (2021) LinkET: everything is linkable. R package version 0.0.2. https://github.com/Hy4m/linkET.
  13. Irnawati I., Riswanto F.D.O., Riyanto S., Martono S., Rohman A. The use of software packages of R factoextra and FactoMineR and their application in principal component analysis for authentication of oils. Indonesian Journal of Chemometrics and Pharmaceutical Analysis. 2020;1:1–10. doi: 10.22146/ijcpa.482. [DOI] [Google Scholar]
  14. Kassambara, A. (2020). ggpubr: “ggplot2” based publication ready plots (R package version 0.4.0). https://cran.r-project.org/web/packages/ggpubr/.
  15. Kobayashi N., Kumagai M., Minegishi D., Tamura K., Aotsuka T., Katakura H. Molecular population genetics of a host-associated sibling species complex of phytophagous ladybird beetles (coleoptera: Coccinellidae: Epilachninae) Journal of Zoological Systematics and Evolutionary Research. 2011;49(1):16–24. doi: 10.1111/j.1439-0469.2010.00581.x. [DOI] [Google Scholar]
  16. Krueger F. TrimGalore: A wrapper around cutadapt and FastQC to consistently apply adapter and quality trimming to FastQ files, with extra functionality for RRBS data. Babraham Institute. 2015 https://github.com/FelixKrueger/TrimGalore [Google Scholar]
  17. Kumar S., Stecher G., Li M., Knyaz C., Tamura K. MEGA X: Molecular evolutionary genetics analysis across computing platforms. Molecular biology and evolution. 2018;35(6):1547–1549. doi: 10.1099/ijsem.0.003362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lan W.-T., Chen Y.-S., Yanagida F. Isolation and characterization of lactic acid bacteria from yan-dong-gua (fermented wax gourd), a traditional fermented food in Taiwan. Journal of Bioscience and Bioengineering. 2009;108(6):484–487. doi: 10.1016/j.jbiosc.2009.06.009. [DOI] [PubMed] [Google Scholar]
  19. Lin L.-J., Zeng J., Tian Q.-M., Ding X.-Q., Zhang X.-Y., Gao X.-Y. Effect of the bacterial community on the volatile flavour profile of a chinese fermented condiment – red sour soup – during fermentation. Food Research International. 2022;155 doi: 10.1016/j.foodres.2022.111059. [DOI] [PubMed] [Google Scholar]
  20. Liu X., Wang X., Cheng Y., Wu Y., Yan Y., Li Z. Variations in volatile organic compounds in Zhenyuan Daocai samples at different storage durations evaluated using E-nose, E-tongue, gas chromatography, and spectrometry. LWT. 2023;173 doi: 10.1016/j.lwt.2022.114186. [DOI] [Google Scholar]
  21. Liu Y., Miao Z., Guan W., Sun B. Analysis of organic volatile flavor compounds in fermented stinky tofu using SPME with different fiber coatings. Molecules. 2012;17(4):3708–3722. doi: 10.3390/molecules17043708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Magoč T., Salzberg S.L. FLASH: Fast length adjustment of short reads to improve genome assemblies. Bioinformatics. 2011;27(21):2957–2963. doi: 10.1093/bioinformatics/btr507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Nilsson R.H., Larsson K.-H., Taylor A.F.S., Bengtsson-Palme J., Jeppesen T.S., Schigel D., Kennedy P., Picard K., Glöckner F.O., Tedersoo L., Saar I., Kõljalg U., Abarenkov K. The UNITE database for molecular identification of fungi: Handling dark taxa and parallel taxonomic classifications. Nucleic Acids Research. 2018;47(D1):D259–D264. doi: 10.1093/nar/gky1022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Osborne C.A., Galic M., Sangwan P., Janssen P.H. PCR-generated artefact from 16S rRNA gene-specific primers. FEMS Microbiology Letters. 2005;248(2):183–187. doi: 10.1016/j.femsle.2005.05.043. [DOI] [PubMed] [Google Scholar]
  25. Peeters C., De Canck E., Cnockaert M., De Brandt E., Snauwaert C., Verheyde B., Depoorter E., Spilker T., LiPuma J.J., Vandamme P. Comparative genomics of pandoraea, a genus enriched in xenobiotic biodegradation and metabolism. Frontiers in Microbiology. 2019;10:2556. doi: 10.3389/fmicb.2019.02556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Santos de Almeida T., Alves dos Santos B., Stefanello A., Duarte dos Santos I., Copetti Fracari J., Silva M., Giongo C., Wagner R., Silveira Nalério E., Venturini Copetti M. Spontaneously growing fungi on the surface and processing areas of matured sheep ham and volatile compounds produced. Food Research International. 2023;173 doi: 10.1016/j.foodres.2023.113287. [DOI] [PubMed] [Google Scholar]
  27. Sequino G., Valentino V., Villani F., De Filippis F. Omics-based monitoring of microbial dynamics across the food chain for the improvement of food safety and quality. Food Research International. 2022;157 doi: 10.1016/j.foodres.2022.111242. [DOI] [PubMed] [Google Scholar]
  28. Shang Z., Ye Z., Li M., Ren H., Cai S., Hu X., Yi J. Dynamics of microbial communities, flavor, and physicochemical properties of pickled chayote during an industrial-scale natural fermentation: Correlation between microorganisms and metabolites. Food Chemistry. 2022;377 doi: 10.1016/j.foodchem.2021.132004. [DOI] [PubMed] [Google Scholar]
  29. Shu W.S., Huang L.N. Microbial diversity in extreme environments. Nature Reviews Microbiology. 2022;20(4):219–235. doi: 10.1038/s41579-021-00648-y. [DOI] [PubMed] [Google Scholar]
  30. Song J., Peng S., Yang J., Zhou F., Suo H. Isolation and identification of novel antibacterial peptides produced by Lactobacillus fermentum SHY10 in chinese pickles. Food Chemistry. 2021;348 doi: 10.1016/j.foodchem.2021.129097. [DOI] [PubMed] [Google Scholar]
  31. Song Z., Hu Y., Chen X., Li G., Zhong Q., He X., Xu W. Correlation between bacterial community succession and propionic acid during gray sufu fermentation. Food Chemistry. 2021;353 doi: 10.1016/j.foodchem.2021.129447. [DOI] [PubMed] [Google Scholar]
  32. Wang Q., Garrity G.M., Tiedje J.M., Cole J.R. Naïve bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Applied and Environmental Microbiology. 2007;73:5261–5267. doi: 10.1128/AEM.00062-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wang Y., Shen Y., Wu Y., Li C., Li L., Zhao Y., Hu X., Wei Y., Huang H. Comparison of the microbial community and flavor compounds in fermented mandarin fish (Siniperca chuatsi): Three typical types of Chinese fermented mandarin fish products. Food Research International. 2021;144 doi: 10.1016/j.foodres.2021.110365. [DOI] [PubMed] [Google Scholar]
  34. Wang Z., Liu S., Xu W., Hu Y., Hu Y., Zhang Y. The microbiome and functions of black soils are altered by dibutyl phthalate contamination. Applied Soil Ecology. 2016;99:51–61. doi: 10.1016/j.apsoil.2015.11.024. [DOI] [Google Scholar]
  35. Wei W., Wang L. Region, society emotion — an anthropological survey of sour food in multi – ethnic aeras in southern Guizhou Province. Journal of Qinghai University(Natural. 2022;Science),48 (04):64–72. [Google Scholar]
  36. Wu Y., Yu C., Wang T., Wu M., Sun H. Research progress on traditional fermention of stinky mandarin fish. Food and Fermentation Industries. 2019;45(19):299–306. doi: 10.13995/j.cnki.11-1802/ts.021280. [DOI] [Google Scholar]
  37. Wu Z., Zhuang B., Weng P., Zhang X. Fermentation quality characteristics and flavor formation changes during the process of pickled wax gourd in Eastern Zhejiang. International Journal of Food Properties. 2016;19(2):409–419. doi: 10.1080/10942912.2015.1027775. [DOI] [Google Scholar]
  38. Xia Y., Yu J., Liu H., Feng C., Shuang Q. Novel insight into physicochemical and flavor formation in koumiss based on microbial metabolic network. Food Research International. 2021;149 doi: 10.1016/j.foodres.2021.110659. [DOI] [PubMed] [Google Scholar]
  39. Xu C., Kong C., Li C., Wang J., Hu R., Wei G. Research process on the three kinds of sour foods in Dushan region of Guizhou Province. China Condiment. 2020;45(06):120–122. [Google Scholar]
  40. Yang Z., Wu R., Wei X., Zhang Z., Wang W., Liu A., Yang J., Ji C., Liang H., Zhang S., Lin X. Moderate fermentation contributes to the formation of typical aroma and good organoleptic properties: A study based on different brands of Chouguiyu. LWT. 2021;152 doi: 10.1016/j.lwt.2021.112325. [DOI] [Google Scholar]
  41. Yoon S.H., Ha S.M., Kwon S., Lim J., Kim Y., Seo H., Chun J. Introducing EzBioCloud: A taxonomically united database of 16S rRNA gene sequences and whole-genome assemblies. International Journal of Systematic and Evolutionary Microbiology. 2017;67(5):1613–1617. doi: 10.1099/ijsem.0.001755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Zhang K., Zhang T.T., Guo R.R., Ye Q., Zhao H.L., Huang X.H. The regulation of key flavor of traditional fermented food by microbial metabolism: A review. Food Chemistry: X. 2023;19(30) doi: 10.1016/j.fochx.2023.100871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zhang Z., Hou Q., Wang Y., Li W., Zhao H., Sun Z., Guo Z. Lactobacillus zhachilii sp. nov., a lactic acid bacterium isolated from Zha-Chili. International Journal of Systematic and Evolutionary Microbiology. 2019;69(8):2196–2201. doi: 10.1099/ijsem.0.003362. [DOI] [PubMed] [Google Scholar]
  44. Zhang Z., Wang Y., Xiang F., Hou Q., Li B., Guo Z. Isolation and identification of lactic acid bacteria from the elderly gut and genetic evolution of isolated Lactobacillus fermentum. Food Science. 2021;42(22):105–122. doi: 10.7506/spkx1002-6630-20200718-245. [DOI] [Google Scholar]
  45. Zhang Z., Zhao H., Deng Y., Luo W., Luo X., Wang C., Quan C., Guo Z., Wang Y. Bacterial diversity and its correlation with sensory quality of two types of zha-chili from Shennongjia region, China. Food Research International. 2023;168 doi: 10.1016/j.foodres.2023.112789. [DOI] [PubMed] [Google Scholar]
  46. Zheng X.-W., Yan Z., Han B.-Z., Zwietering M.H., Samson R.A., Boekhout T., Robert Nout M.J. Complex microbiota of a Chinese “Fen” liquor fermentation starter (Fen-Daqu), revealed by culture-dependent and culture-independent methods. Food Microbiology. 2012;31(2):293–300. doi: 10.1016/j.fm.2012.03.008. [DOI] [PubMed] [Google Scholar]
  47. Zhou Y., Yan Y., Zhou J., Huang J., Zhang F., Cui K., Qian K., Jiang J., Xie N. Effect of fermentation methods on microflora and volatile compound in Huangshan smelly mandarin fish. Meat Research. 2019;33(10):36–43. doi: 10.7506/rlyj1001-8123-20190626-148. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary data 1
mmc1.pdf (228.2KB, pdf)
Supplementary data 2
mmc2.pdf (121.4KB, pdf)
Supplementary data 3
mmc3.pdf (135.2KB, pdf)
Supplementary data 4
mmc4.xlsx (12.7KB, xlsx)

Data Availability Statement

Data will be made available on request.

Articles from Food Chemistry: X are provided here courtesy of Elsevier

RESOURCES