Lifestyle patterns influence the composition of the gut microbiome in a healthy Chinese population

High-throughput sequencing allows for the comprehensive analysis of the human intestinal microbiota. However, extensive association analyses between the microbiome and lifestyle differences in the Chinese population are limited. Here, we carried out an independent cohort study—the Chinese Healthy Gut Project (n = 483)—where correlations between the gut microbiota and dietary and lifestyle variables in a healthy Chinese population are defined. We collected both questionnaire data, including basic information and lifestyle and dietary variables, and fecal stools from the enrolled volunteers. We then performed 16S rRNA sequencing on the microbial DNA isolated from the stools to assess the composition of the intestinal microbiota. We found that Prevotella and Bacteroides were the most abundant genera in the healthy Chinese gut microbiome. Additionally, 9 out of 29 clinical and questionnaire-based phenotype covariates were found to be associated with the variation in the composition of the gut microbiota. Among these lifestyle phenotypes, sleep procrastination, negative mood, and drinking habits had the largest effect size. Additionally, an appreciable effect of urbanization was observed, resulting in decreased intra-individual diversity, increased inter-individual diversity, and an increased abundance of the Bacteroides enterotype. The results of this study provide a foundation for assessing the healthy Chinese gut microbiota community structure at baseline in a healthy Chinese population. Furthermore, this study also provides insights into understanding how distinctive living habits influence the relationships between the Chinese gut microbiome and systemic health state.

www.nature.com/scientificreports/ in terms of the diversity, structure, and composition of the gut microbiota [16][17][18][19][20][21][22] .Taken together, these studies suggest that the human gut microbiome is influenced by individual lifestyle variables, and understanding the relationships between the gut microbiome and various lifestyles and dietary patterns prior to the onset of disease may help guide disease treatment.
The identification of the distribution of bacterial taxa in populations with different genetic backgrounds and lifestyle patterns may be useful in understanding mechanisms linking lifestyle patterns with overall health or the risk of disease [23][24][25] .Reproducible patterns of microbial variation-bacterial taxa that can be separated into clusters termed "enterotypes" 26 -have been observed in the adult human gut [26][27][28][29][30][31][32] .To date, most populationlevel studies have described differences and associations in the gut microbiome from individuals with specific diseases or healthy control cohorts in distinct geographic locations, such as Asia, Europe, and the United States of America 18,20,21,26,[33][34][35][36][37] .Although a few studies have investigated the gut microbiota characteristics of China, they do not adequately reflect the gut microbiota of China as a whole, as these studies had limited participants, focused on specific regions, or lacked sufficient lifestyle data [38][39][40][41] .Thus, large-scale phenotyping studies that integrate gut microbiome profiles with comprehensive lifestyle phenotypes in the Chinese population remain scarce and are of great significance for a detailed understanding of the characteristics of the Chinese gut microbiome under different sub-health states.
The aim of this study was to compare the gut microbiota in healthy (no apparent diseases) Chinese volunteers (n = 483) and to correlate differences in the gut microbiota with various lifestyle variables.We performed 16S rRNA sequencing on stools collected from the enrolled volunteers and correlated the results with the demographic, diet, and lifestyle information provided by the volunteers via a questionnaire.The results of this study provide insights into the intricate interplay between dietary and lifestyle variables and the gut microbiota in a healthy Chinese population.

Results
Characteristics and distribution of intestinal flora in healthy people.The data presented in this study were collected from 483 healthy Chinese people.All participants completed a questionnaire regarding basic demographic and lifestyle data.The information collected from the questionnaire is described in the Materials and Methods section and is presented in Table 1.Feces were collected from 483 participants, and the gut microbiomes were assessed using 16S rRNA sequencing.In total, 483 sequencing samples were obtained.The participants spanned 11 ethnic groups, had an average age of 36.96years, an average body mass index (BMI) of 22.36, were 65.42% female, and were from 62 residential areas (Table 2).
We then grouped the genera based on their detection rate, defined as the sample size of a certain bacterium/ total sample size.We divided 171 bacterial genera into groups based on their abundance frequency-low-frequency flora (detection rate < 10%), medium-frequency flora (detection rate 10-70%), and high-frequency flora (detection rate > 70%).After grouping, we identified a total of 38 high-frequency genera, 66 medium-frequency genera, and 67 low-frequency genera.The 16 genera of high-frequency bacteria were detected in over 90% of the samples, indicating their status as fundamental intestinal microorganisms within the Chinese population (Fig. 1B).Notably, eight of these genera were also found in the core gut microbiota of Guangdong province 93 , six were among the top 9 most abundant fecal genera in another Chinese cohort 94 , five overlapped with the top 20 fecal genera discovered in the Human Microbiome Project 95 , and eight were part of the core microbiota in a Chinese cohort comprising 2678 healthy individuals from 28 provinces 96 .And Bacteroides, Blautia and Faecalibacterium were overlapped among our study and the studies above.The 38 high-frequency genera in 483 samples are shown in Fig. 1C.

Analysis of enterotypes in a healthy Chinese population.
Previous studies have demonstrated that the gut microbiota of various human populations clusters around three primary driver taxa (enterotypes): Prevotella, Bacteroides, and Ruminococcus 26,27 .In order to analyze the enterotypes of healthy Chinese people, we performed unsupervised clustering on the sequencing results from the 483 stool samples.The clustering results showed that the gut bacteria of healthy Chinese people could be divided into two groups, Prevotella (39.54%, n = 135) and Bacteroides (38.12%, n = 348) with a significant different gut microbiota structure (weighted unifrac distance, Adonis, Pr(> F) = 0.001; Anosim, p = 0.001); a Ruminococcus enterotype was not found (Fig. 2A,B, Supplementary Fig. 1, Supplementary Table 1).This is consistent with a previous study, which demonstrated that only Prevotella and Bacteroides were common enterotypes in Chinese populations 52 .Thus, enterotypes with Prevotella and Bacteroides as the driving taxa are more common in the Chinese population.
Demographic factors and Bacteria.The variables of demographic factors including gender, age, BMI, ethnicity and habitation were associated with the composition and structure of the gut microbiota 89 .In this study, we performed differential analysis of the gut microbiota in five demographic factors variables: (1) gender, (2) age, (3) BMI, (4) ethnicity, and (5) habitation.

Different living conditions and bacteria.
There are limited studies highlighting the impact of lifestyle variables on the composition and structure of the gut microbiota.In order to explore how various lifestyle variables influence the gut microbiota, we performed differential analysis of the gut microbiota in parallel with six lifestyle variables: (1) the degree of sleep deprivation, (2) state of fatigue, (3) appearance of negative emotions, (4) occurrence of oral ulcers and skin acne, and (5) smoking frequency.
According to the consensus of the American Academy of Sleep Medicine and Sleep Research Society, adults should strive for 7 h of nightly sleep while young adults require 9 h.Insufficient sleep due to occupational or recreational activities is classified as sleep deprivation 101 .One study has demonstrated a correlation between sleep deprivation and the gut dysbiosis 102 .In this study, increased sleep deprivation decreased the alpha diversity of the gut microbiome (Supplementary Table 13)although the differences were not significant, and the gut microbiota structures among the normal sleep (n = 145), more sleep deprivation (n = 131) and less sleep deprivation (n = 207) groups showed a significant difference (Binary-jaccard distance, Adonis, Pr (> F) = 0.029, Fig. 5A) The abundance of Bifidobacterium in the normal sleep group was significantly lower than that in the more sleep deprived (MaAsLin, Coef = 1.367, p = 0.001, q = 0.125) groups, while Parabacteroides showed a significant higher relative abundance in less sleep group than that in normal sleep MaAsLin, Coef = 0.738, p = 0.002, q = 0.153, Fig. 5B, Supplementary Tables 14, 15).
Fatigue is a prevalent symptom frequently encountered in daily life.It can be classified into two types: acute and chronic fatigue.Acute fatigue usually resolves after rest or treatment of the underlying condition, while chronic fatigue is an enduring debilitating process.Therefore, they differ at least in terms of frequency of experiencing fatigue.Our results showed that as the fatigue level and frequency increased (normal, n = 128; sometimes, n = 283, always, n = 72), the level of Sellimonas decreased gradually (t-test, normal vs. always, p = 0.026, q = 0.731; sometimes vs. always, p = 0.030, q = 0.766 (Fig. 5C, Supplementary Table 16).www.nature.com/scientificreports/Previous studies have shown that the intestinal microbiome plays a role in stress response, inflammation, depression and anxiety, but specific changes in microbial composition and structure were not clear [53][54][55] .According to the survey conducted in 2021, the most negative emotions experienced by Chinese individuals were anxiety, anger, fear and sadness 103 .We identified five genera which were significant different in relative abundance, and Negativibacillus (t-test, p = 0.001, q = 0.153; MaAsLin, Coef = -0.687,p < 0.001, q = 0.037) showed a significant higher level after removing other confounding factors (Supplementary Tables 17, 18).Taken together, our results showed that sleep deprivation, the fatigue level, and negative emotions had associations with the gut microbiota community structure.Some unhealthy lifestyle factors may lead to the development of oral ulcers and acne.The underlying mechanism is likely related to proinflammatory response 6,7,56 .Probiotic therapy has been shown to modulate inflammation and improve symptoms of both oral ulcers and acne 2-5, 8, 9, 57-60 .Therefore, it is necessary to characterize the gut microbiome in this population.Our results showed that as the frequency of oral ulcers and acne increased, the abundance of supposedly beneficial bacteria, Dialister significantly decreased gradually although the decrease was not significant after removing the confounding factors (seldom, n = 246, sometimes, n = 173, always, n = 64; t-test, seldom vs. always, p = 0.002, q = 0.199; sometimes vs. always, p = 0.023, q = 0.527, Fig. 5D, Supplementary Table 19).
Different dietary habits and bacteria.Dietary habits have been suggested to be intimately related to the gut microbiome 20,[66][67][68] .In order to explore the impact of different dietary habits on the composition of the intestinal flora, we analyzed the impact of three dietary preferences, (1) starch intake (cereal, rice, flour-based foods, and high-starch foods), (2) protein intake (bean products, dairy, eggs, and meat), and (3) dietary preference (meat-heavy diet, vegetable-heavy diet, or a balanced diet of meat and vegetables), on the structure of the gut microbiota in the healthy Chinese volunteers.
We then grouped the fecal stool samples according to the predominant protein source consumed by the volunteers (bean products, n = 47; dairy, n = 68; eggs, n = 95; meat, n = 273), and found that the level of Sutterella (t-test, meat vs. bean products, p = 0.002, q = 0.116; meat vs. dairy, p < 0.001, q = 0.033; meat vs. eggs, p = 0.024, q = 0.677) and Mitsuokella (t-test, meat vs. bean products, p = 0.011, q = 0.230; meat vs. dairy, p = 0.003, q = 0.193) were significantly higher in people who consumed meat as their primary protein source compared to people to who consumed other primary protein source.The abundance of Fusobacterium was significantly lower in people who consumed dairy as their primary protein source compared to people who derived the majority of their protein from meat or eggs (t-test, dairy vs. meat, p < 0.001, q = 0.033; dairy vs. eggs, p = 0.028, q = 0.750) although the different levels were not significant after removing the confounders (Fig. 6C, Supplementary Table 26).These results demonstrate that the dietary protein source was associated with the community structure of the intestinal flora.
Next, we were interested in determining how general dietary preference was associated with the gut microbiome of a healthy Chinese population.The stool samples were divided into three groups based on the dietary preference of a meat diet (n = 107), vegetable diet (n = 57), or a balanced diet of both meat and vegetables (n = 319).Several bacteria in genus-level were identified between any two groups, but only [Ruminococcus]_gnavus_group showed a significant higher relative abundance in vegetable-diet preference volunteers compared with who preferred balanced diet after removing the confounding factors(t-test, p = 0.046, q = 0.526; MaAsLin, Coef = 0.977, p = 0.001, q = 0.224) (Fig. 6D, Supplementary Table 27).

Discussion
Numerous studies have shown that several factors, including diet and lifestyle variables, influence the diversity, structure, and composition of the gut microbiota [16][17][18][19][20][21][22]69 . Delieating the composition of the gut microbiome in populations with various genetic backgrounds and lifestyle patterns may be beneficial in understanding the mechanisms linking lifestyle patterns and overall health and disease risk.Here, we collected questionnaire data and fecal stool samples from 483 healthy Chinese volunteers that spanned 11 ethnic groups and were from 62 residential areas.We then performed 16S rRNA sequencing on the microbial DNA isolated from the stool samples to assess the composition of the gut microbiota in relation to lifestyle and dietary variables.
Our results showed that Bacteroides and Faecalibacterium were the most abundant genera in the healthy Chinese gut microbiome, which is consistent with a previous study that assessed the gut microbiota in healthy Chinese volunteers 35 .This result is consistent with population-level studies conducted in African populations, but it differs from studies conducted on European and American populations, in which Bacteroides and Firmicutes were the primary enterotypes 18,70 .As reported in previous studies 26,69 , Prevotella-and Bacteroides-rich compositions were found to be relatively non-overlapping in energy obtain.People with a Bacteroides-rich enterotype obtain energy from carbohydrates and proteins, while people with a Prevotella-rich enterotype mainly degrade mucin glycoproteins existing in the intestinal mucosal layer.www.nature.com/scientificreports/Bacteroides is an important and abundant member of the gut microbiome and is a core microorganism of common enterotypes 71 .In this study, we found that Bacteroides levels increased in Han ethnic group.People who consumed eggs as their main protein source, and always occurred to oral ulcers and skin acne had an increased relative abundance of Bacteroides, which is consistent with a previous study 75 , although the significance disappeared after the confounder was removed.In previous studies, Bacteroides was reported to degrade glycans and proteins as nutritional sources 43 , and Bacteroides levels have been shown to be closely related to dietary habits 72 .Adjusting Bacteroides levels by modifying the dietary structure may be a way to improve sub-health lifestyle patterns.Additionally, it was reported that the nutrients obtained by Bacteroides via glycan degradation can be used as an energy source by other microorganisms, and the content of Bacteroides in the gut is related to other microorganisms 73 .In addition, Bacteroides has been reported to be related to the immune system 74 , and as such, has been shown to activate CD4 + cells by producing zwitterionic polysaccharide (ZPS), which triggers the immune system.. Prevotella, similar to Bacteroides, is a driving taxon in bacterial enterotypes of the gut microbiome.Hydrolases expressed by Prevotella are essential to the degradation of plant fibers 76 , which supports previous studies that have shown that Prevotella is enriched in populations that consume a non-Western diet and/or fiber-rich diet 18,21,77 .Our results showed that Prevotella abundance was higher in healthy Chinese male volunteers, and the increased level of Prevotella in gut was accompanied by elevated BMI.Additionally, Coprococcus, a butyrate-producing bacterium involved in dopamine-related biological pathways 78 , was found to be significantly lower in volunteers who had higher BMI index, which was similar to Prevotella.
Upon further analysis, we speculated that different living habits might affect the structure and composition of the gut microbiome in a couple of ways.First, some living habits, such as sleep deprivation, can increase stress and gut permeability, which may lead to abdominal distension, stomachache, and inflammation-all of which can reduce the diversity of the gut microbiome [79][80][81] .Second, intestinal inflammation will continue to erode the protective mucosa that protects the intestinal lining from bacterial invasion.Since the stomach wall is thin, microbial by-products, and possibly even entire bacteria, may pass through the stomach wall, triggering an inflammatory reaction of the immune system, thus forming a vicious circle.Through this process, the inflammatory process may induce fatigue 82,83 .From our results, we found that the four staple foods were associated with the growth of different kinds of bacteria.Therefore, it can be speculated that a reasonable mode of carbon and water intake is to eat an appropriate amount of high-starch food and to choose a variety of staple foods.
There are some limitations in our study.First, although we collected demographic, lifestyle, and dietary information from all of the volunteers, our study is limited by the lack of comprehensive physiological indices and detailed lifestyle and dietary information.Such information is imperative for fully interpreting the data.Furthermore, compared with a previous study 73 that examined 150 host phenotypic features, our research is still not detailed enough in regards to phenotypic diversity.Second, while we determined the differential abundance of microorganisms in relation to various phenotypes, we did not investigate differences in metabolic pathways or metabolites.As such, we were not able to reveal any information regarding the metagenome of the volunteers.Nonetheless, the microbial profiles acquired in this study do help elucidate the gut microbiota in the healthy Chinese population at baseline.Third, the correlations established in this study cannot determine causal relationships between the gut microbiota and lifestyle variables.In order to determine causal relationships, animal experiments and in vitro experiments, as well as intervention experiments, need to be performed.Fourth, the data presented in this study are from 483 healthy Chinese volunteers.However, while our total sample number is one of the largest reported in the literature, our sample number per variable is small, and additional samples are needed to verify the statistical analysis.In a follow-up study, we hope to further accumulate relevant data to improve the dimensional phenotypic analyses.There is no doubt that changing the structure and composition of intestinal microorganisms may become an important part of precision medicine in the twenty-first century.
In summary, our study compared the gut microbiota with demographic, lifestyle, and dietary variables in healthy Chinese volunteers.We found that the most abundant genera in the healthy Chinese gut microbiome were Prevotella and Bacteroides.Additionally, nine clinical and questionnaire-based phenotype covariates were found to be associated with the composition of the gut microbiota.The results of this study provide a foundation for elucidating the gut microbiome in the Chinese population at baseline.Moreover, understanding the complex interactions between the gut microbiome and various lifestyle and dietary variables prior to disease onset may help prevent disease or guide disease treatment.

Materials and methods
Study population and research data.After excluding the volunteers who were diagnosed with any illness, we recruited 483 healthy volunteers ranging in age from 5 to 80 years old.Information was collected from each recruited volunteer via an online questionnaire based on their lifestyles over the past week.The questionnaire contained 28 questions grounded on contemporary research that elucidates the factors associated with the gut microbiome-7 of which regarded basic information, such as age and gender, and the remaining 21 were questions relating to dietary and lifestyle variables divided into three categories: 1) bowel habits (four questions), dietary habits (five questions), allergens and health (five questions), and other lifestyle habits (seven questions).Detailed information regarding the questionnaire questions is listed in Table 1.
To gain a better understanding of the gut microbiome features among individuals with different lifestyles in China, the initial step involves characterizing distinct groups.According to WHO guidelines, adults should engage in 150-300 min of exercise per week 104 .Participants were classified into three groups based on selfreported exercise frequency: "No" indicated minimal physical activity, "1-2 times a week" indicated insufficient exercise, and "3 or more times" indicated compliance with guidelines.Fatigue was categorized into three groups based on responses to the question "Have you experienced recent feelings of tiredness?":"No" indicated sustained where p a and p b are the abundance distributions of samples a and b, and KLD x, y is the Kullback-Leibler divergence between x and y, defined as A pseudocount of 0.000001 was applied to the abundance distributions to avoid zero in the equation.Then, using the JSD, we assessed the Calinski-Harabasz (CH) index to evaluate the optimum number of clusters 9 ; it is defined as where B k is the squared distance between all points i and j, for which i and j are not in the same cluster, and W k is the squared distance between all points i and j, for which i and j are in the same cluster.The result indicated that the CH index reached the maximum when divided into two clusters, so the number of clusters was set to k = 2.
Contingency tables were used for checking the independence between enterotypes and the other phenotypes, and the results showed that the distribution of enterotypes was not affected by demographic characteristics and phenotypic information (for each phenotype p > 0.05).
Here, we chose the top 35 bacteria in genus-level in abundance to cluster.Last, the clustering quality was assessed using the silhouette validation technique.Values derived from the silhouette are located between − 1 and + 1, and the value in this research was 0.262.In theory, the smaller the gap between the acquired value and the value derived from the silhouette, the more accurate the clustering technique.

Figure 1 .
Figure 1.The gut microbiota composition and the high-frequency bacteria of 483 healthy Chinese volunteers.(A) GraPhlAn diagram of bacterial genera based on amplicon sequence variants (ASVs annotations).According to these annotations, two phyla, Bacteriodota and Firmicutes, accounted for the majority of the microbial composition in healthy Chinese volunteers.(B) The 16 genera of bacteria were detected in over 90 percent of the samples, with Bacteriodes and Faecalibacterium exhibiting the highest relative abundance.(C) Relative abundances of 38 high-frequency genera in 483 stool samples.Similarly, Bacteroides (shown in blue), Prevotella (red), and Faecalibacterium (green) combined accounted for nearly 80% of the feces of the healthy Chinese people.

Figure 2 .
Figure 2. The major enterotypes found in the stool samples from the healthy Chinese population.(A) Nonmetric multidimensional scaling analysis (NMDS) of the sequencing results from 483 stool samples showed that the intestinal flora in the healthy population of Chinese could be divided into two groups.(B) Heatmap of the intestinal flora with significant difference between the two groups in a healthy Chinese population.As shown, Bacteroides (38.12%) and Prevotella (39.54%) were the driving taxa in Group 1 and Group 2, respectively.

Figure 3 .
Figure 3. Differences in the microbiota alpha diversity, composition, and bacteria in genus-level between male (n = 167) and female (n = 316).(A) Female had higher richness and evenness of gut microiota than male.(B) Non-metric multidimensional scaling analysis (NMDS) based on unweighted Unifrac distance matric showed that the two gender groups had separated microbita composition.(C) Nine bacteria in genus-level were significantly different between the two gender groups.

Figure 4 .
Figure 4. Differences in the microbiota alpha diversity, composition, and bacteria in genus-level between Han ethnic group (n = 418) and other ethnic groups (n = 65).(A) Chinese Han had lower richness of gut microiota than other ethnic groups.(B) Non-metric multidimensional scaling analysis (NMDS) based on weighted Unifrac distance matric showed the separated microbita composition between Chinese Han and others.(C) Parabacteroides and Bacteroides showed a higher level in Chinese Han while Dorea showed a lower relative abundance.

Figure 5 .
Figure 5. Differences in the microbiota composition and bacteria in genus-level between people different lifestyle conditions.(A) Gut microbiota structures among the normal sleep (n = 145), more sleep deprivation (n = 131) and less sleep deprivation (n = 207) groups showed a significant difference.(B) Two bacteria in genus-level were identified significantly different among three sleep deprivation groups.(C) T-test revealed the difference in relative abundance of Sellimonas in healthy Chinese volunteers with varying degrees of fatigue (normal, n = 128; sometimes, n = 283, always, n = 72).(D) T-test revealed the difference in relative abundance of Dialister in healthy Chinese volunteers with varying degrees of oral ulcers and acne (seldom, n = 246, sometimes, n = 173, always, n = 64). https://doi.org/10.1038/s41598-023-41532-4www.nature.com/scientificreports/

Figure 6 .
Figure 6.Differences in the microbiota alpha diversity, composition, and bacteria in genus-level between people with different dietary habits.(A) Gut microbiota structures among the different starch intake groups (cereal, n = 30; rice, n = 390; flour-based, n = 57; high-starch, n = 6) showed a significant difference.(B) Healthy Chinese who took cereal as their staple food had higher level of gut microbita diversity than who consumed flour-based foods or rice.(C) T-test revealed the difference in relative abundance of Sutterellas in healthy Chinese volunteers with different types of protein intake (bean products, n = 47; dairy, n = 68; eggs, n = 95; meat, n = 273).(D) Healthy Chinese who preferred vegetable diet (n = 57) had higher level of [Ruminococcus]_gnavus_ group than who preferred balanced (n = 319) or meat (n = 107) diet.

Table 1 .
Lifestyle patterns collected from the questionnaires of 483 healthy Chinese people (n = 483).

Table 2 .
The characteristics of the study participants (n = 483).BMI body mass index