{"paper_id":"13c1d02b-86ef-4f63-9882-45f9fb80f2ba","body_text":"R E S E A R C H A R T I C L E Open Access\nHigh-throughput sequencing study of the\neffect of transabdominal hysterectomy on\nintestinal flora in patients with uterine\nfibroids\nWantong Wang, Yibing Li, Qijun Wu, Xin Pan, Xinhui He and Xiaoxin Ma *\nAbstract\nBackground: To investigate the effect of transabdominal hysterectomy on the diversity of the intestinal flora in\npatients with uterine fibroids. Patients with uterine fibroids were selected from September 2018 to December 2018,\nin the Department of Obstetrics and Gynecology, Shengjing Hospital of China Medical University, and stool\nspecimens were collected from patients before and after surgery.\nResults: High-throughput sequencing of the 16S rRNA gene was used to detect the changes in microbial\ncommunity structure and diversity, and the effects of total hysterectomy on the intestinal flora were further\nanalyzed. Estrogen levels decreased after trans-abdominal hysterectomy. High-throughput sequencing showed that\nafter abdominal hysterectomy, the abundance and diversity of the intestinal flora decreased. The abundance\nchanges were mainly due to Proteobacteria, where their abundance increased.\nConclusions: Trans-abdominal hysterectomy changes the intestinal flora of the body by lowering the level of\nestrogen in the body, which reduces the diversity and abundance of the intestinal flora.\nKeywords: Intestinal flora, Uterine fibroids, Hysterectomy, Estrogen, High-throughput sequencing, 16sRNA\nBackground\nUterine fibroids are common benign tumors that mani-\nfest in 30 –50 year-old women. According to autopsy sta-\ntistics, approximately 20% of women over the age of 30\nhave uterine fibroids. The growth and persistence of\nuterine fibroids depends on the woman ’s estrogen and\nprogesterone levels, and the synthesis of estrogen is af-\nfected by many factors in vitro and in vivo [ 1]. The\nchange in estrogen levels is an independent contributor\nto the onset of uterine fibroids in women. The synthesis\nand secretion of estrogen are affected by vitamin D and\nE, and trace elements, such as iodine and selenium.\nWomen with uterine fibroids and trace elements in tu-\nmors have statistically significant differences, when com-\npared to normal women, with zinc and copper as the\nmain features [ 2, 3]. Related studies have also confirmed\nthat in addition to the effects of estrogen and progester-\none on uterine fibroids, trace elements, growth factors\nand immune cells in the body are also closely correlated\nto uterine fibroids. The TNF- α (tumor necrosis factor- α)\nproduced by phagocytic cells can cause proliferative\nchanges in injured smooth muscle cells. TNF- α is corre-\nlated to the occurrence of uterine fibroids, which may be\ndue to the suppression of the immune state of the body\n[4]. Growth factors in the body are correlated to the oc-\ncurrence of uterine fibroids. The most closely correlated\nare EGF (epidermal growth factor), VGEF (vascular\n© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,\nwhich permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give\nappropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if\nchanges were made. The images or other third party material in this article are included in the article's Creative Commons\nlicence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons\nlicence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain\npermission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.\nThe Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/) applies to the\ndata made available in this article, unless otherwise stated in a credit line to the data.\n* Correspondence: maxiaoxin666@aliyun.com\nDepartment of Obstetrics and Gynecology, Shengjing Hospital of China\nMedical University, Shenyang 110004, China\nWang et al. BMC Microbiology           (2020) 20:98 \nhttps://doi.org/10.1186/s12866-020-01779-7\n\nendothelial growth factor), and TGF (transforming\ngrowth factor). Patients with uterine fibroids have higher\nreceptors for related growth factors, when compared to\nnormal uterine myometrium. Growth factors are affected\nby estrogen, and increase the proliferation and division\nof uterine fibroid cells [ 5–7].\nFor women with uterine fibroids without fertility re-\nquirements, total hysterectomy is one of the more classic\nprocedures. Total hysterectomy has an effect on the pa-\ntient’s sex hormone levels. Most women who undergo\ntotal hysterectomy would exhibit varying degrees of sexual\nfunction reduction [ 8]. A study revealed that the effect of\nsex hormone levels on patients with different ranges of\nhysterectomy was not statistically significant. However,\nafter the total hysterectomy, the sex hormone levels were\nsignificantly lower than those before surgery [9].\nA large number of gut microbes constitute the most\ncomplex micro-ecological system in the human body\n[10]. The changes in the composition of the human in-\ntestinal flora affect the body ’s movement, immunity, and\nendocrine and sex hormone levels [ 11–13]. The large\nnumber of human intestinal microbes affect the physio-\nlogical functions of the human body at all times, and has\nprofound effects on the synthesis and secretion of hor-\nmones, trace elements, growth factors and immune sys-\ntem. The intestinal flora and host have a mutually\nbeneficial relationship, and are interdependent. The in-\ntestinal flora plays an important role in regulating meta-\nbolic pathways, synthesizing essential components such\nas vitamins, and promoting the establishment of the im-\nmune system. The flora provides nutrients and a suitable\nliving environment, and the stable balance between the\nhost and intestinal flora plays an indispensable role in\nmaintaining the health of the human body [ 14]. The in-\ntestinal flora plays a vital role in regulating several\nchronic diseases, including obesity, cardiovascular dis-\nease and kidney diseases [ 15–19]. The balance in intes-\ntinal flora composition is the key to maintaining\nintestinal function and systemic homeostasis [ 20]. A\nstudy conducted in 2018 revealed that estrogen could in-\nhibit the overgrowth of Bacteroides fragilis ,Escherichia\ncoli and Fusobacterium nucleatum to maintain the\nhomeostasis of intestine mucosa [ 21]. Estrogen defi-\nciency can lead to a marked reduction in intestinal flora\nbiodiversity and the number of beneficial bacteria with\nimmune regulation, while the number of conditional\npathogens is elevated, thereby triggering a series of in-\nflammatory immune responses that lead to a disease. In\nrecent years, some high-throughput sequencing and\nother cutting-edge technologies have promoted the\nstudy of the intestinal flora. However,there remains little\nknowledge about uterine fibroids and intestinal flora.\nThe present study analyzed the differences in E2 (estra-\ndiol), AMH (anti-Mullerian hormone) and FSH (follicle-\nstimulating hormone) (FSH) before and after transab-\ndominal hysterectomy, and further analyzed the impact\non the diversity of the human intestinal flora.\nResults\nGeneral clinical data\nWe collected fresh fecal samples from 15 patients, a total\nof 30 samples, 15 of which were preoperative specimens\nand 15 postoperative specimens. We compared E2,\nAMH, and FSH in the RS1 group and RS2 group. The\nAMH level of the RS1 group was lower than that of the\nRS2 group, but the difference was not statistically signifi-\ncant. The FSH level was higher in the RS1 group than\nthe RS2 group, but the difference was not statistically\nsignificant. The E2 level was significant higher in the\nRS1 group. Therefore, we believe that ovarian function\nwas greater in the RS1 group compared to the RS2\ngroup (Table 1).\nStudy of the structure of intestinal flora in patients with\nundergoing transabdominal hysterectomy\nA total of 2,083,203 sequences were obtained from 30\nsamples from the 2 groups, with an average of 73,170 se-\nquences per sample. After quality control, 69,440 valid\ndata were obtained, and the quality control efficiency\nwas 94.96%. The mean number of sequences in the RS1\ngroup was 70,122.73 ± 54.71, and the mean number of\nsequences in the RS2 group was 68,755.47 ± 5159.97,\nwhere the difference between the two groups was not\nstatistically significant ( P = 0.314). The sequence was\nclustered into OTUs (operational taxonomic units) with\n97% identity. A total of 6651 OTUs were obtained, with\nan average of 221 OTUs per sample. Among them, the\nRS1 group had 5949 OTUs and the RS2 group had 5290\nOTUs. Comparing the number of unique and common\nOTUs in the two groups according to the Veen chart,\nthe OTU number composition and similar situation of\nthe sample could be compared. The number of OTUs\nshared by the RS1 and RS2 groups was 4588, and the\nnumber of unique OTUs was 1361 in the RS1 group and\n702 in the RS2 group. The Veen diagram showed that\nthe diversity of RS1 was significantly higher than that of\nRS2. The petal plot indicated the number of OTUs con-\ntained in each sample of the RS1 and RS2 groups\n(Fig. 1A, B).\nTable 1 Comparison of ovarian function between the two\ngroups\nGroup T\nvalue\nP\nvalueRS1(n = 15) RS2(n = 15)\nAMH (nmmol/L) 4.78 ± 1.75 4.01 ± 1.00 1.82 0.09\nE2(pmol/L) 460.71 ± 303.08 229.36 ± 110.19 2.65 0.02\nFSH (IU/L) 11.86 ± 11.77 19.47 ± 28.29 −0.94 0.36\nWang et al. BMC Microbiology           (2020) 20:98 Page 2 of 11\n\nIn this study, the statistical analysis of the sample at\n97% similarity level produced a rarefaction curve, and a\nsignificant plateau appeared on the starting curve of\n7737 sequences, indicating that the sequencing depth\nwas close to saturation, increasing the sequencing depth\nat 97% similarity. No more bacterial species could be\nfound at the top. The combination of the rarefaction\ncurve and the Shannon diversity curve indicated that the\namount of data in this study was reasonable, the sequen-\ncing depth was sufficient; the detection rate of the bac-\nterial species of the sample was close to saturation,\nmeeting the requirements of subsequent bioinformatics\nanalysis (Fig. 1C, D, E). Second, by analyzing Good ’s\ncoverage index, the sequencing coverage of each sample\nwas over 98%. 16S rRNA gene sequencing was effective\nin this study and represented more than 98% of the bac-\nterial species in each sample, and the coverage of the\nbacterial species was good. The rank-abundance curve\nwas steep, indicating that sample distribution was un-\neven, and there could have been a dominant flora; the\ncurve span was large, indicating that the abundance of\nthe species was high. As shown in (Fig. 1F), the RS1\ncurve had a wide and flat span on the horizontal axis, in-\ndicating that species richness and uniformity of the RS1\ngroup were better.\nThe analytical indices included ACE, Chao1, Simpson,\nand Shannon. ACE and Chao1 are indices for evaluating\nthe number of OTUs contained in the sample. Simpson\nand Shannon indices were used to reflect the diversity of\nthe sample population: the larger the Simpson, the lower\nFig. 1 Comparison of the the structure of intestinal flora between patients with uterine fibroids before undergoing transabdominal hysterectomy\n(group RS1) and patients with uterine fibroids after undergoing transabdominal hysterectomy (group RS2). a Veen diagrams of OTUs . b Petal plot:\nEach petal in the petal diagram represents a sample, different colors represent different samples, the core number in the middle represents the\nnumber of OTUs common to all samples, and the number on the petal represents the sample unique OTU number. c Shannon description in the\ntwo groups . d Shannon description among each sample. e Rarefaction Curve.f Rank Abundance\nWang et al. BMC Microbiology           (2020) 20:98 Page 3 of 11\n\nthe diversity of the flora, and the larger the Shannon, the\nhigher the diversity of the flora. The Wilcoxon rank sum\ntest found that there was no significant difference be-\ntween the RS1 group and RS2 group, P = 0.2328: Shan-\nnon index, P = 0.2169; Simpson index, P = 0.2017; ACE,\nP = 0.3669; and Chao1, P = 0.5125. There was no signifi-\ncant difference in diversity between RS1 and RS2\n(Table 2).\nChanges in microbial community composition after\nhysterectomy\nThe relevant bacterial composition of the patient was\nanalyzed from the perspective of phylogeny (domain/\nphylum/class/order/family/genus). In the human micro-\nbiota, the bacterial group is quite conservative, which\ncan directly reflect the heterogeneity of bacterial com-\nmunity structure in different human body parts. There-\nfore, when analyzing the composition of common\nhuman microorganisms, we must first explain the rela-\ntive abundance of bacterial phyla. The law of variation,\nin the bacterial classification unit, the phylum can be\ncalled the highest classification unit. Species annotations\nwere made by comparison with the Silva 132 database,\nand statistics were analyzed at different classification\nlevels: there were 6651 OTUs, of which all could be an-\nnotated to the database (100.00%), and the proportion of\nannotations to the boundary level was 100.00%. The ra-\ntio of the phylum level is 91.35%, the ratio of the class\nlevel is 81.01%, the ratio of the order level is 69.59%, the\nproportion of the family level is 58.39%, the proportion\nof the genus level is 33.82%. According to the results of\nthe species annotation, each species or group is selected\nin the top 10 species of the highest abundance in the\nhorizontal phylum, class, order, family, genus, and the\nrelative abundance of the species is generated. A cylin-\ndrical cumulative graph to visually view the species and\ntheir proportions of the relatively abundant abundance\nof each sample at different classification levels.\nAnalysis at the phylum level\nAt the phylum level, the top ten strains of the RS1\nand RS2 groups ranked as the most abundant. In the\nRS1 group, they were: Bacteroidetes , Proteobacteria ,\nFirmicutes , Acidobacteria , Actinobacteria , Gemmati-\nmonadetes, Planctomycetes , Chloroflexi , Verrucomicro-\nbia, Tenericutes , and bacteria that could not be\nclassified (Fig. 2A). In the RS2 group, they were: Bac-\nteroidetes , Proteobacteria , Firmicutes , Melainabacteria ,\nAcidobacteria , Cyanobacteria , Gemmatimonadetes ,\nActinobacteria , Verrucomicrobia , Planctomycetes ,a n d\nbacteria that could not be classified (Fig. 2B). At the\nphylum level,the dominant bacteria were basically\nthe same, and the three dominant bacteria, Bacteroi-\ndetes , Proteobacteria and Firmicutes , accounted for\nmore than 75% of the intestinal flora (Fig. 2C). Sig-\nnificant individual differences occurred between the\nsamples. The proportion of Bacteroidetes in each\nsample ranged from 2.85 to 78.77%, and the propor-\ntion of Proteobacteria in each sample ranged from\n3.15 to 92.13%. Firmicutes were present in each sam-\nple at a proportion of 2.31 to 66.03%, and the rela-\ntive abundance of Bacteroidetes was higher in the\nRS1 group than RS2 group, P = 0.003. After surgery,\nBacteroidetes decreased significantly, and the relative\nabundance of Proteobacteria was significantly lower\nin the RS1 group, P = 0.016, and thus increased after\nsurgery. The relative abundance of Firmicutes was\nlower in the RS1 group but not significantly, P =\n0.926 (Table 3); combined with UPGMA (Un-\nweighted Pair-group Method with Arithmetic Means)\nclustering tree, in environmental biology, UPGMA is\na commonly used clustering analysis method. It is\nthe earliest method used to solve a classification\nproblem. The UPGMA clustering analysis was per-\nformed with the Weighted UniFrac distance matrix\nand the Unweighted UniFrac distance matrix, and\nthe clustering results were integrated with the rela-\ntive abundance of the species at the phylum level\n(Fig. 2D, E, suggesting that the structure of the two\ngroups of bacteria is not significantly different.\nAnalysis at the class level\nAt the class level, the top ten strains in the RS1 and RS2\ngroups were selected. The RS1 group included: Bacteroi-\ndia, Gammaproteobacteria, Clostridia, Bacilli, Alphapro-\nteobacteria, unidentified_Actinobacteria, Negativicutes,\nunidentified_Acidobacteria, Deltaproteobacteria, uniden-\ntified_Gemmatimonadetes, and bacteria that could not\nbe classified (Fig. 3A). The RS2 group included: Gam-\nmaproteobacteria, Clostridia, Bacteroidia, unidentified_\nMelainabacteria, Bacilli, Negativicutes, unidentified_\nCyanobacteria, Alphaproteobacteria, unidentified_Acido-\nbacteria, Deltaproteobacteria, and bacteria that could\nnot be classified (Fig. 3B). The dominant species at the\nclass level were Gammaproteobacteria, Bacteroidia and\nClostridia, which accounted for more than 50% of the\nintestinal flora (Fig. 3C). Significant individual\nTable 2 Sequencing results and diversity index of two groups\nof samples. *The operational taxonomic units (OTUs) were\ndefined at the 97% similarity level\nGroup P value\nRS1 RS2\nACE 2032.94 ± 1342.18 1468.45 ± 1207.79 0.37\nChao1 2200.16 ± 1437.12 1664.03 ± 1300.57 0.51\nSimpson 0.91 ± 0.12 0.86 ± 0.13 0.20\nShannon 6.87 ± 2.73 5.28 ± 2.44 0.22\nWang et al. BMC Microbiology           (2020) 20:98 Page 4 of 11\n\ndifferences occurred between the samples. Gammapro-\nteobacteria accounted for 2.04 –91.43% of each sample,\nand Bacteroidia accounted for 2.85 –78.77% of each\nsample, while Clostridia accounted for 0.37 –64.80% of\neach sample.\nAnalysis at the order level\nAt the order level, the top ten strains in the RS1 and\nRS2 groups were selected. The RS1 group included:\nBacteroidales , Enterobacteriales , Xanthomonadales ,\nClostridiales , Lactobacillales , Flavobacteriales , Bifido-\nbacteriales , unidentified_Gammaproteobacteria , Sele-\nnomonadales , unidentified_Acidobacteria ,a n do t h e r\nbacteria that could not be classified (Fig. 3D). The\nRS2 group included:Enterobacteriales, Clostridiales, Bacter-\noidales, unidentified_Melainabacteria, Lactobacillales,\nunidentified_Gammaproteobacteria, Selenomonadales, un-\nidentified_Cyanobacteria, unidentified_Acidobacteria, Aero-\nmonadales, and other bacteria that could not be classified\n(Fig. 3E). The dominant species at the order level were\nEnterobacteriales, Bacteroidalesand Clostridiales(Fig. 3F).\nFig. 2 Changes in microbial community composition after hysterectomy at the level of phylum. a The top ten strains of the RS1 group ranked as\nthe most abundant. b The top ten strains of the RS2 group ranked as the most abundant. c At the phylum level, the three dominant bacteria,\nBacteroidetes, Proteobacteria and Firmicutes, accounted for more than 75% of the intestinal flora . d UPGMA clustering tree based on the Weighted\nUnifrac distance.e UPGMA clustering tree based on the Unweighted UniFrac distance\nTable 3 Comparison of relative abundance of two groups of\nTOP3 strains at the phylum level\nGroup P value\nRS1 RS2\nBacteroidetes(%) 24.54 11.43 0.003\nProteobacteria(%) 34.36 54.04 0.016\nFirmicutes(%) 0.003 17.26 0.926\nWang et al. BMC Microbiology           (2020) 20:98 Page 5 of 11\n\nAnalysis of species differences and differences between\nspecies\nThe weighted UniFrac distance and the unweighted\nUniFrac distance were used to measure the difference\ncoefficient between the two samples (Fig. 4A). The un-\nweighted UniFrac distance was tested by the Wilcoxon\nrank sum test, P = 0.4646, and the weighted UniFrac dis-\ntance was also tested by the Wilcoxon rank sum test,\nP = 0.1083, indicating that there was no significant differ-\nence in species diversity between the two groups.\nNMDS (Non-Metric Multi-Dimensional Scaling)\nstatistics is a sorting method suitable for ecological\nresearch. The smaller the stress (< 0.2), the more accurately.\nStress = 0.156 indicated that NMDS accurately reflected the\ndegree of difference between samples (Fig.4B).\nMRPP (Multi Response Permutation Procedure) ana-\nlysis was used to analyze whether the differences in mi-\ncrobial community structure between groups were\nsignificant. A value of less than 0.05 indicates a signifi-\ncant difference. Table 4 showed that the differences be-\ntween the two groups were significant.\nTo find the differential species between the groups at\neach classification level (phylum, class, order), a t-test\ntest between the groups was performed to determine the\nspecies with significant differences ( P < 0.05).\nAt the phylum level, the species difference analysis be-\ntween the t-test groups was obtained. Proteobacteria\nshowed a significant difference between the two groups.\nThe average abundance of the RS1 group was 34.36%,\nand the average abundance of the RS2 group was\n54.04%, P < 0.05 (Fig. 4C).\nAt the class level, the difference in species between the\nt-test groups was obtained. Gammaproteobacteria (Pro-\nteobacteria) showed a significant difference between the\ntwo groups; the average abundance of RS1 was 22.74%,\nand the average abundance of RS2 was 48.89%, P < 0.05.\nFig. 3 Changes in microbial community composition after hysterectomy at the level of calss and order. a The top ten strains of the RS1 group\nranked as the most abundant at the level of class. b The top ten strains of the RS2 group ranked as the most abundant at the level of class. c At\nthe class level, the three dominant bacteria, Gammaproteobacteria, Bacteroidia and Clostridia, accounted for more than 50% of the intestinal\nflora.d The top ten strains of the RS1 group ranked as the most abundant at the level of order. e The top ten strains of the RS2 group ranked as\nthe most abundant at the level of order. f The dominant species at the order level were Enterobacteriales, Bacteroidales and Clostridiales\nWang et al. BMC Microbiology           (2020) 20:98 Page 6 of 11\n\nThere was a significant difference between the groups\nfor Alphaproteobacteria (Proteobacteria); the average\nabundance of RS1 was 7.56%, and the average abun-\ndance of RS2 was 3.20%, P < 0.05 (Fig. 4D).\nAt the order level, the difference in species between\nthe t-test groups was determined. For Enterobacteriales\n(p__Proteobacteria; c__Gammaproteobacteria), the aver-\nage abundance of RS1 was 9.44%, and the average abun-\ndance of RS2 was 42.05%, P < 0.05, indicating a\nsignificant difference between the two groups. For Rhizo-\nbiales (p__Proteobacteria; c__Alphaproteobacteria), the\naverage abundance of RS1 was 2.85%, and the average\nabundance of RS2 was 1.10%, P < 0.05, indicating a sig-\nnificant difference between the two groups. For Caulo-\nbacterales (p__Proteobacteria; c__Alphaproteobacteria),\nthe average abundance of RS1 was 0.58%, and the aver-\nage abundance of RS2 was 0.13%, P < 0.05, indicating a\nsignificant difference between the two groups. The\nFig. 4 Analysis of species differences and differences between species. a Beta diversity heatmap: The numbers in the boxes in the figure are the\ndissimilarity coefficients between the samples. The disparity coefficients are smaller, the difference in species diversity is smaller. In the sam e box,\nthe upper and lower values represent Weighted Unifrac and Unweighted Unifrac distance. b NMDS:Each point in the figure represents a sample.\nThe distance between the points represents the degree of difference, and the same group of samples is represented by the same color. c Analysis\nof species differences between T_test groups at the phylum level. d Analysis of species differences between T_test groups at the class level. e\nAnalysis of species differences between T_test groups at the order level. f Cladogram:In the cladogram, the circle radiating from the inside to the\noutside represents the classification level from the phylum to the genus (or species). Species with no significant difference are uniformly colored\nyellow. Different species of Biomarker follow the group for coloring. The red nodes indicate the microbial groups that play an important role in\nthe red group, and the green nodes indicate the microbial groups that play an important role in the green group. g The LDA value\ndistribution histogram\nTable 4 Significantness test table for community structure\ndifferences between groups\nGroup A Observed-delta Expected-delta Significance\nRS1-RS2 0.03 0.80 0.82 0.036\nWang et al. BMC Microbiology           (2020) 20:98 Page 7 of 11\n\naverage abundance of RS1 for Chthoniobacterales (p__\nVerrucomicrobia; c__Verrucomicrobiae) was 0.13%, and\nthe average abundance of RS2 group was 0.04%, P < 0.05,\nindicating a significant difference between the two\ngroups (Fig. 4E).\nLEfSe (LDA (Linear Discriminant Analysis) effect size)\nis used to compare two or more groups. The LDA value\ndistribution histogram shows the species with an LDA\nscore greater than the set value (the default setting is 4),\ni.e., the biomarker with statistical differences between\nthe groups. This study showed that the species with sig-\nnificant differences in abundance in the different groups\nwere c_Gammaproteobacteria, f_Xanthomonadaceae,\nand o_Xanthomonadales, and the length of the histo-\ngram bar represents the size of the difference species\n(i.e., LDA score). C_Gammaproteobacteria was enriched\nin the RS2 group, and f_Xanthomonadaceae and o_\nXanthomonadales were enriched in the RS1 group; the\nLDA score showed f_Xanthomonadaceae > o_Xanthomo-\nnadales, indicating great influence of f_Xanthomonada-\nceae in the RS1 group (Fig. 4F, G ).\nDiscussion\nHigh-throughput sequencing technology is a presently\nwidely used sequencing technology, which can quickly\nand accurately sequence a large number of samples sim-\nultaneously, and obtain a large amount of data. High-\nthroughput sequencing is particularly important in the\ngut flora. This technology has the advantages of high ac-\ncuracy and high sequencing depth, which can detect low\nabundance or unknown bacteria, and further obtain\nmore accurate and comprehensive flora information.\nThe intestinal flora is a special system that coexists with\nthe human body. This has a large number and variety,\nand is an ecosystem with a high density of cells. In re-\ncent years, as the research on the intestinal flora become\nmore and more intensive, studies have revealed that the\nintestinal flora is correlated to a variety of diseases, in-\ncluding intestinal tumors, autism, obesity and diabetes.\nThe intestinal flora can change through the interaction\nof hormones in the body and in vitro, and this can affect\nthe ecological balance in the body. If the intestinal flora\nin the body becomes imbalanced, this would cause an\nadverse effect on intestinal function. The literature indi-\ncates that the use of exogenous hormones can cause the\nimbalance of the intestinal flora, and increase the diver-\nsity of the flora. The adult gut flora contains approxi-\nmately 1000 different bacterial species, in which thick-\nwalled bacteria (such as Clostridium, Enterococcus and\nLactobacillus), Bacteroidetes (such as Prevotella and Bac-\nteroidetes)a n d Actinobacteria (such as Bifidobacteria)\nare the major members [ 22]. In recent years, the deter-\nmination of whether the intestinal flora can regulate es-\ntrogen and its metabolites has attracted the attention of\nscholars. Glucuronidase in the intestinal flora can pro-\nmote the reabsorption of estrogen, and the level of estro-\ngen is closely correlated to the occurrence of uterine\nfibroids [ 23]. The study [ 24] conducted by Plottel et al.\nand other studies have found that a variety of bacteria in\nthe intestinal flora are associated with estrogen metabol-\nism, and all genes in the bacteria that are capable of me-\ntabolizing estrogen were collectively referred to as the\nestrobolome. The high bacterial enzyme activity of the\nestrobolome causes the free estrogen in the entero-\nhepatic circulation to significantly rise, thereby forming\nan endogenous hormonal environment. This endogen-\nous hormonal environment significantly increases the\nrisk of hormone-dependent tumors, including breast and\nendometrial cancers, through direct or indirect effects\n[25]. The study conducted by Guo et al. revealed the re-\nlationship between PCOS (polycystic ovary syndrome)\nand the intestinal flora. Estrone and E2 levels were lower\nin the PCOS group than in the normal control group,\nand this shows that the intestinal flora can affect the oc-\ncurrence and treatment of PCOS [ 26].\nThe present study focused on the changes in the intes-\ntinal flora of patients with factor hysteromyoma before\nand after total hysterectomy. Uterine fibroids are benign\ntumors in women, who are highly dependent on estro-\ngen and progesterone in the body. Although the uterus\ndoes not secrete hormones, the uterus is the main recep-\ntor organ of estrogen in the body. The anatomical struc-\nture of the uterus and ovary is closely correlated. Except\nfor the ovarian blood supply from the ovarian artery, a\nconsiderable part of the blood supply comes from the\nascending branch of the uterine artery. The scope of the\ntotal hysterectomy includes the uterus and its surround-\ning ligaments. At the same time, due to some unavoid-\nable thermal damage during the operation, which can\ndamage the surrounding vascular tissues, the patient ’s\novarian blood supply is affected after the total hysterec-\ntomy, thereby leading to a decrease in sex hormone se-\ncretion [ 27]. The ovary has two functions of ovulation\nand secretion of hormones. The indicators that respond\nto ovarian function include E2, FSH and AMH. If ovar-\nian function is reduced, E2 and AMH are reduced, and\nFSH is increases. The statistics of the present study sug-\ngest that the levels of E2 and AMH in patients undergo-\ning abdominal hysterectomy were significantly higher in\nthe preoperative group than in the postoperative group,\nwhile the levels of FSH in the preoperative group were\nlower than those in the postoperative group, and the dif-\nferences between these two groups were statistically sig-\nnificant. This indicates that hysterectomy damages the\nfunction of ovarian secretion to a certain extent, and\nthat the estrogen reduction is more significant. A study\n[28] also reported that patients with uterine fibroids\nhave a greater effect on ovarian function after\nWang et al. BMC Microbiology           (2020) 20:98 Page 8 of 11\n\nhysterectomy, when compared to patients with uterine\nfibroids removal.\nSome literatures have suggested that the richness of\nthe intestinal flora is closely correlated to systemic estro-\ngen. The richness of the flora at the phylum level does\nnot affect the content of estrogen and estrogen metabo-\nlites in the body. However, a large number of bacteria at\nthe family and species level regulate the content of estro-\ngen. In particular, Clostridium and Pneumococcus have\nthe most significant effect on estrogen metabolism [ 29].\nIn recent years, there have been many studies on the in-\ntestinal flora and various systemic diseases, but there are\nfew literatures on the changes of the intestinal flora after\ntotal hysterectomy. According to the results of the\npresent study, the flora coverage of these two groups of\nsamples reached more than 98%, indicating that the flora\ncoverage was good. From the perspective of the diversity\nof the flora, the analysis of the alpha diversity analysis\nindex (Shannon index, Simpson index, ACE index and\nChao1 index) indicated that there was no statistically\nsignificant difference between the RS1 group and RS2\ngroup. This shows that there were no significant differ-\nences between these two groups of microbiome alpha di-\nversity. Although the estrogen level decreased and the\novarian function was reduced after the total hysterec-\ntomy, the diversity of the intestinal flora before and after\nsurgery was less different for patients.\nFor these patients, the level of estrogen in the body de-\ncreased after the total hysterectomy. In order to further\nexplore the predominant strains of these two groups be-\nfore and after surgery, further exploration was con-\nducted in the present study. Based on the composition\nof the flora, the dominant strains in the RS1 and RS2\ngroups were identified. At the phylum level, the top\nthree dominant strains in the RS1 and RS2 groups were\nBacteroidetes, Proteobacteria and Firmicutes. The abun-\ndance of Bacteroidetes RS1 was significantly higher than\nthat of RS2, and the abundance of Proteobacteria RS1\nwas significantly lower than that of RS2. Gammaproteo-\nbacteria, Bacteroidia and Clostridia dominated at the\nlevel of the class. Enterobacteriales, Bacteroidales and\nClostridiales were the dominant species at the order\nlevel. At the phylum level, the species with differences\nbetween the RS1 and RS2 groups was Proteobacteria.A t\nthe class level, the species with statistically significant\ndifferences between these two groups were Gammapro-\nteobacteria (Proteobacteria) and Alphaproteobacteria\n(Proteobacteria). At the level of the order, the species\nwith statistically significant differences between these\ntwo groups were Enterobacteriales (p__Proteobacteria,\nc__Gammaproteobacteria), Rhizobiales (p__Proteobac-\nteria, c__Alphaproteobacteria)a n d Caulobacterales (p__\nProteobacteria, c__Alphaproteobacteria). Some studies\nhave revealed that a decrease in estrogen level leads to a\ndecrease in the diversity of the intestinal flora and a re-\nduction in the abundance of thick-walled bacteria, in-\ncluding Clostridium [30]. The decrease in estrogen levels\nin the present study lead to the increase in abundance of\nFirmicutes, the decrease in diversity of Bacteroidetes, and\nthe increase in species diversity of Proteobacteria. How-\never, a study conducted in 2014 [ 31] revealed that the\nabundance of estrogen and its metabolites, and the intes-\ntinal flora in the phyla, class and genus categories were\ncorrelated. Furthermore, it was noted that Clostridiales\nand Ruminococcaceae under Firmicutes are positively\ncorrelated with estrogen metabolites, but negatively cor-\nrelated with Bacteroidetes. In general, for patients who\nreceived total hysterectomy, the composition of the in-\ntestinal flora changes with the increase in Proteobacteria.\nThe MRPP analysis revealed that the differences be-\ntween the RS1 and RS2 groups were greater than the\nintra-group differences, and that the differences between\nthese groups were statistically significant. Based on the\nUnifrac distance for PoCA (Principal Co-ordinates Ana-\nlysis) analysis, the PC1 factor expressed in 39.2%, and\nthe NMDS could accurately reflect the degree of differ-\nence between these two groups of samples. At the same\ntime, the level of estrogen in the body decreased after\nthe total hysterectomy. It was assumed that the total\nhysterectomy was the cause of the intestinal flora.\nThe level of estrogen in the body can change the intes-\ntinal flora, but the manner in which the intestinal flora\nregulates estrogen metabolism in the body remains un-\nclear. Therefore, further research is needed. Although\nthe present study proposed the decrease in estrogen\nlevel and a series of changes in the intestinal flora after\nthe total hysterectomy, there were still limitations in the\npresent study that needs further exploration. The\npresent study lacks a comparative analysis of sex hor-\nmones and the intestinal flora in patients with uterine fi-\nbroids and healthy women. At the same time, due to the\nsmall sample size, there may be some bias in the experi-\nmental results. In the future, the sample size needs to be\nexpanded for a more in-depth research.\nConclusion\nIn conclusion, transabdominal hysterectomy can reduce\nestrogen levels in the body, and reduce the diversity and\nabundance of the intestinal flora before and after sur-\ngery, but the main difference was the increase in Proteo-\nbacteria. In the future, more multi-center and large-\nsample studies are needed to obtain more accurate con-\nclusions, and further investigate the interaction mechan-\nism between the intestinal flora and estrogen. A more\nrigorous and reliable scientific basis for patients with\nfactor hysteromyoma, undergoing total hysterectomy\nafter application of hormone replacement, when neces-\nsary, as well as dietary guidance, should be provided.\nWang et al. BMC Microbiology           (2020) 20:98 Page 9 of 11\n\nMethods\nPatient enrollment and sample collection\nThis was a case-control study that included women aged\n40–45 years who underwent transabdominal hysterec-\ntomy due to uterine fibroids from September 2018 to\nDecember 2018 in the Department of Obstetrics and\nGynecology, Shengjing Hospital of China Medical Uni-\nversity. Inclusion criteria were: patients with uterine fi-\nbroids with ultrasound and clinical diagnosis; no\nhypertension, diabetes, heart disease or other comorbidi-\nties; no menopause; no previous abdominal surgery or\nintestinal disease history; consistent intensity of anti-\nbiotic use after surgery; and no significant change in\nbowel habits after surgery. The patients were divided\ninto the preoperative RS1 group, and the postoperative\nRS2 group. Blood for E2, AMH and FSH tests and stool\nspecimens were collected from the RS1 and RS2 groups.\nThis study was approved by the Ethics Committee of\nShengjing Hospital of China Medical University.\nSample collection, DNA extraction and 16S rRNA gene\namplicon sequencing\nFresh stools were collected from 15 patients before\nand at one month after the surgery. Approximately 5\ng of the middle part of the stool was placed in a ster-\nile dry test tube containing pure ethanol, and stored\nin a freezer at − 80 °C. The CTAB (Cetyltrimethylam-\nmonium Ammonium Bromide) or SDS (Sodium dode-\ncyl sulfate) method was used to extract the genomic\nDNA of the sample, and the purity and concentration\nof the DNA were detected by agarose gel electrophor-\nesis. Afterwards, an appropriate amount of the sample\nwas collected in a centrifuge tube, and the sample\nwas diluted to 1 ng/ μl with sterile water. Then, these\nsamples were amplified by PCR (polymerase chain re-\naction). Next, equal concentration mixing was per-\nformed, according to the PCR product concentration.\nAfter thorough mixing, 2% agarose gel electrophoresis\nw a su s e dt op u r i f yt h eP C Rp r o d u c t s .T h es e q u e n c e\nwith a main band size within 400 –450 bp was select,\nand this was tapped to recover the target band. The\nlibrary was constructed using the NEB Next® Ultra ™\nDNA Library Prep Kit for Illumina (New England\nBiolabs). The constructed library was subjected to\nQubit quantification and library detection. After the\nqualification, HiSeq was used for the on-line se-\nquencing [ 32, 33]. Blood samples were taken on an\nempty stomach at one day before the surgery, and\nat one month after the surgery, in order to detect\nthe E2, AMH and FSH. E2 and FSH were detected\nusing the Beckman Coulter UniCel DXI 800. The\ndetection method used was chemiluminescence.\nAMH was detected by ELISA (enzyme-linked im-\nmunosorbent assay).\nAnalysis of 16S rRNA gene sequences, bioinformatics and\nstatistical analyses\nThe sequences were analyzed using the QIIME [ 34]\n(Quantitative Insights Into Microbial Ecology) software\npackage, and in-house Perl scripts were used to analyze\nthe alpha-diversity (within samples) and beta-diversity\n(among samples). First, the reads were filtered using\nQIIME quality filters. Then, pick_de_novo_otus.py was\nuse to select the OTUs by making an OTU table. Se-\nquences with ≥97% similarity were assigned to the same\nOTUs. A representative sequence was selected for each\nOTU, and the RDP classifier [ 35] was used to annotate\nthe taxonomic information for each representative se-\nquence. OTUs that reached a 97-nucleotide similarity\nlevel were used for the alpha diversity (Shannon, Simp-\nson index) and richness analysis (ACE and Chao1).\nThen, rarefaction curves were generated based on these\nthree matrices. A metagenomic biomarker discovery ap-\nproach was employed with LEfSe (linear discriminant\nanalysis [LDA] coupled with effect size measurement),\nwhich performed with the nonparametric Wilcoxon\nsum-rank test. In order to mine deeper data of the mi-\ncrobial diversity of the differences between these sam-\nples, a significance test were conducted with some\nstatistical analysis methods, including t-test, LEfSe,\nANOSIM and MRPP. Beta diversity was used to explore\nthe differences between samples, and the Wilcoxon rank\nsum test was used to determine whether the differences\nin beta diversity between these groups were significant.\nThe SPSS 24.0 software was used for data processing. The\nmeasurement data were expressed as mean ± standard de-\nviation. T-test was used for comparisons between two\ngroups. The count data was used to indicate the ratio.\nChi-square test was used for comparisons between two\ngroups. P < 0.05 was considered statistically significant.\nAbbreviations\nTNF-α: Tumor Necrosis Factor- α; EGF: epidermal growth factor; VGEF: vascular\nendothelial growth factor; TGF: transforming growth factor; E2: estradiol;\nAMH: anti-Mullerian hormone; FSH: follicle-stimulating hormone;\nOTUs: Operational Taxonomic Units; UPGMA: Unweighted Pair-group Method\nwith Arithmetic Means; NMDS: Non-Metric Multi-Dimensional Scaling;\nMRPP: Multi Response Permutation Procedure; LDA: Linear Discriminant\nAnalysis; LEfSe: LDA effect size; PCOS: polycystic ovary syndrome;\nPoCA: Principal Co-ordinates Analysis; PCR: polymerase chain reaction;\nELISA: enzyme-linked immunosorbent assay; CTAB: Cetyltrimethylammonium\nAmmonium Bromide; SDS: Sodium dodecyl sulfate; QIIME: Quantitative\nInsights Into Microbial Ecology\nAcknowledgements\nNot applicable.\nAuthors’ contributions\nWTW,YBL and XXM conceived and designed the study. WTW,YBL,XP\ncollected the samples. WTW wrote the manuscript. XXM and QJW helped\ndraft and revised the manuscript. WWT,XXM and QJW modified the\nmanuscript to prepare its final version. QJW and XHH helped with statistics.\nAll authors read and approved the final manuscript.\nWang et al. BMC Microbiology           (2020) 20:98 Page 10 of 11\n\nFunding\nThis work was supported by the National Natural Science Foundation of\nChina [grant numbers 81872123]; University innovation team of Liaoning\nProvince; Special Professor of Liaoning Province; “Major Special Construction\nPlan” for Discipline Construction of China Medical University in 2018[grant\nnumbers 3110118029] and Outstanding Scientific Fund of Shengjing Hospital\n[grant numbers 201601]. Funding bodies had no role in study design,data\ncollection,analysis,and writing manuscript.\nAvailability of data and materials\nThe datasets generated and analyzed are available from corresponding\nauthor on reasonable request.\nEthics approval and consent to participate\nThe study was approved by the Ethics Committee of Shengjing Hospital of\nChina Medical University. Informed consent was not needed because the\nstudy was retrospective. Patient identity and all the personal information\nwere confidential.\nConsent for publication\nNot applicable.\nCompeting interests\nThe authors declare that they have no competing interests.\nReceived: 26 November 2019 Accepted: 2 April 2020\nReferences\n1. Khan AT, Shehmar M, Gupta JK. Uterine fibroids: current perspectives. Int J\nWomen's Health. 2014;6:95 –114.\n2. Du X, Liu Y, Zhao C, Fang J, Wang X, Wei L. Changes of serum 25(OH) D3\nand IGF-1 levels in patients with thyroid nodules. BMC Endocr Disord. 2019;\n19(1):48.\n3. Fichera M, Török P, Tesarik J, Della Corte L, Rizzo G, Garzon S, Carlea A, Di\nAngelo AS, Zito G, Panella MM. Vitamin D, reproductive disorders and\nassisted reproduction: evidences and perspectives. Int J Food Sci Nutr. 2019:\n1–10.\n4. Joseph DS, Malik M, Nurudeen S, Catherino WH. Myometrial cells undergo\nfibrotic transformation under the influence of transforming growth factor\nbeta-3. Fertil Steril. 2010;93(5):1500 –8.\n5. Perrot D, Fernandez H, Levaillant JM, Capmas P. Quality assessment of pelvic\nultrasound for uterine myoma according to the CNGOF guidelines. J\nGynecol Obstet Hum Reprod. 2017;46(4):317 –21.\n6. Marsh EE, Chibber S, Wu J, Siegersma K, Kim J, Bulun S. Epidermal growth\nfactor-containing fibulin-like extracellular matrix protein 1 expression and\nregulation in uterine leiomyoma. Fertil Steril. 2016;105(4):1070 –5.\n7. Su C, Fan M, Lu L, Li P. Role of epidermal growth factor in pathogenesis of\nuterine leiomyomas. Asian Pac J Trop Med. 2015;8(5):378 –81.\n8. Yuan H, Wang C, Wang D, Wang Y. Comparing the effect of laparoscopic\nsupracervical and total hysterectomy for uterine fibroids on ovarian reserve\nby assessing serum anti-mullerian hormone levels: a prospective cohort\nstudy. J Minim Invasive Gynecol. 2015;22(4):637 –41.\n9. Malik M, Britten J, Cox J, Patel A, Catherino WH. Gonadotropin-releasing\nhormone analogues inhibit leiomyoma extracellular matrix despite presence\nof gonadal hormones. Fertil Steril. 2016;105(1):214 –24.\n10. Heintz-Buschart A, Wilmes P. Human gut microbiome: function matters.\nTrends Microbiol. 2018;26(7):563 –74.\n11. Quigley EM, Monsour HP. The gut microbiota and the liver: implications for\nclinical practice. Expert Rev Gastroenterol Hepatol. 2013;7(8):723 –32.\n12. McLean MH, Dieguez D Jr, Miller LM, Young HA. Does the microbiota play a\nrole in the pathogenesis of autoimmune diseases? Gut. 2015;64(2):332 –41.\n13. Lv L-X, Fang D-Q, Shi D, Chen D-Y, Yan R, Zhu Y-X, Chen Y-F, Shao L, Guo F-\nF, Wu W-R, et al. Alterations and correlations of the gut microbiome,\nmetabolism and immunity in patients with primary biliary cirrhosis. Environ\nMicrobiol. 2016;18(7):2272 –86.\n14. Federico A, Dallio M, Godos J, Loguercio C, Salomone F. Targeting gut-liver\naxis for the treatment of nonalcoholic steatohepatitis: translational and\nclinical evidence. Transl Res. 2016;167(1):116 –24.\n15. Hu J, Zhong X, Yan J, Zhou D, Qin D, Xiao X, Zheng Y, Liu Y. High-\nthroughput sequencing analysis of intestinal flora changes in ESRD and\nCKD patients. BMC Nephrol. 2020;21(1):12.\n16. Barlow GM, Yu A, Mathur R. Role of the gut microbiome in obesity and\ndiabetes mellitus. Nutr Clin Pract. 2015;30(6):787 –97.\n17. Ahmadmehrabi S, Tang WHW. Gut microbiome and its role in\ncardiovascular diseases. Curr Opin Cardiol. 2017;32(6):761 –6.\n18. Lew KN, Starkweather A, Cong X, Judge M. A mechanistic model of gut-\nbrain Axis perturbation and high-fat diet pathways to gut microbiome\nhomeostatic disruption, systemic inflammation, and type 2 diabetes. Biol\nRes Nurs. 2019;21(4):384 –99.\n19. Chung L, Thiele Orberg E, Geis AL, Chan JL, Fu K, DeStefano Shields CE,\nDejea CM, Fathi P, Chen J, Finard BB, et al. Bacteroides fragilis Toxin\nCoordinates a Pro-carcinogenic Inflammatory Cascade via Targeting of\nColonic Epithelial Cells. Cell Host Microbe. 2018;23(2):203 –14.e205.\n20. Singh RK, Chang H-W, Yan D, Lee KM, Ucmak D, Wong K, Abrouk M,\nFarahnik B, Nakamura M, Zhu TH, et al. Influence of diet on the gut\nmicrobiome and implications for human health. J Transl Med. 2017;15(1):73.\n21. Yang Y, Qu C, Liang S, Wang G, Han H, Chen N, Wang X, Luo Z, Zhong C,\nChen Y, et al. Estrogen inhibits the overgrowth of Escherichia coli in the rat\nintestine under simulated microgravity. Mol Med Rep. 2018;17(2):2313 –20.\n22. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Manichanh C, Nielsen T, Pons\nN, Levenez F, Yamada T, et al. A human gut microbial gene catalogue\nestablished by metagenomic sequencing. Nature. 2010;464(7285):59 –65.\n23. Claesson MJ, Cusack S, O'Sullivan O, Greene-Diniz R, de Weerd H, Flannery E,\nMarchesi JR, Falush D, Dinan T, Fitzgerald G, et al. Composition, variability,\nand temporal stability of the intestinal microbiota of the elderly. Proc Natl\nAcad Sci U S A. 2011;108(Suppl 1):4586 –91.\n24. Plottel CS, Blaser MJ. Microbiome and malignancy. Cell Host Microbe. 2011;\n10(4):324–35.\n25. Rajagopala SV, Vashee S, Oldfield LM, Suzuki Y, Venter JC, Telenti A, Nelson\nKE. The human microbiome and Cancer. Cancer Prev Res (Phila). 2017;10(4):\n226–34.\n26. Guo Y, Qi Y, Yang X, Zhao L, Wen S, Liu Y, Tang L. Association between\npolycystic ovary syndrome and gut microbiota. PLoS One. 2016;11(4):\ne0153196.\n27. Song T, Kim MK, Kim ML, Jung YW, Yun BS, Seong SJ, Kwon SH. Impact of\nopportunistic salpingectomy on anti-Müllerian hormone in patients\nundergoing laparoscopic hysterectomy: a multicentre randomised\ncontrolled trial. BJOG. 2017;124(2):314 –20.\n28. Wang H-Y, Quan S, Zhang R-L, Ye H-Y, Bi Y-l, Jiang Z-m, Ng EHY.\nComparison of serum anti-Mullerian hormone levels following hysterectomy\nand myomectomy for benign gynaecological conditions. Eur J Obstet\nGynecol Reprod Biol .2013; 171(2):368 –71.\n29. Flores R, Shi J, Fuhrman B, Xu X, Veenstra TD, Gail MH, Gajer P, Ravel J,\nGoedert JJ. Fecal microbial determinants of fecal and systemic estrogens\nand estrogen metabolites: a cross-sectional study. J Transl Med. 2012;10:253.\n30. Breban M. Gut microbiota and inflammatory joint diseases. Joint Bone\nSpine. 2016;83(6):645–9.\n31. Fuhrman BJ, Feigelson HS, Flores R, Gail MH, Xu X, Ravel J, Goedert JJ.\nAssociations of the fecal microbiome with urinary estrogens and estrogen\nmetabolites in postmenopausal women. J Clin Endocrinol Metab. 2014;\n99(12):4632–40.\n32. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA,\nTurnbaugh PJ, Fierer N, Knight R. Global patterns of 16S rRNA diversity at a\ndepth of millions of sequences per sample. Proc Natl Acad Sci U S A. 2011;\n108(Suppl 1):4516–22.\n33. Hess M, Sczyrba A, Egan R, Kim T-W, Chokhawala H, Schroth G, Luo S, Clark\nDS, Chen F, Zhang T, et al. Metagenomic discovery of biomass-degrading\ngenes and genomes from cow rumen. Science. 2011;331(6016):463 –7.\n34. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK,\nFierer N, Peña AG, Goodrich JK, Gordon JI, et al. QIIME allows analysis of high-\nthroughput community sequencing data. Nat Methods. 2010;7(5):335–6.\n35. Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid\nassignment of rRNA sequences into the new bacterial taxonomy. Appl\nEnviron Microbiol. 2007;73(16):5261 –7.\nPublisher’sN o t e\nSpringer Nature remains neutral with regard to jurisdictional claims in\npublished maps and institutional affiliations.\nWang et al. BMC Microbiology           (2020) 20:98 Page 11 of 11","source_license":"CC0","license_restricted":false}