Gut microbiome profiles in Chilean participants with colorectal adenomas: an exploratory 16S rRNA sequencing study Authors: Roxana González-Stegmaier, Troy Ejsmentewicz, Celia Podesta, Roddy Jorquera, Rodrigo Lagos, Carlos Barrientos, Edmundo Aravena, Macarena Larrain, José Pablo Aguilera, Evelyn Silva-Moreno, Franz Villarroel-Espíndola Published: 2026 ABSTRACT AbstractColorectal cancer (CRC) is the third most commonly diagnosed cancer globally and the second leading cause of cancer-related mortality. The disease typically progresses from premalignant adenomatous lesions. Recent studies highlight the pivotal role of the gut microbiota in colorectal carcinogenesis, suggesting a complex interplay between the microbiota and tumor development. This study analyzes gut microbiota profiles in a single-center cohort of Chilean individuals with and without adenomas, using 16S rRNA sequencing to interrogate microbial changes associated with the transition from healthy colon epithelium to precancerous lesions. The results show that the adenoma-associated microbiota is characterized by increased Proteobacteria abundance and decreased Firmicutes abundance compared with healthy controls. Participants with precancerous lesions exhibit increased abundance of Desulfovibrio, a genus of sulfate-reducing bacteria capable of producing hydrogen sulfide (H2S), a microbial metabolite that has been linked to DNA damage and epithelial stress responses. These findings provide a baseline of the microbial landscape in our population and identify biomarkers associated with premalignant lesions, highlighting microbiome analysis as a diagnostic tool for colorectal disease.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-57417-1.Keywords: Microbiome, Adenoma, Colorectal cancer, ChileanSubject terms: Biomarkers, Cancer, Gastroenterology, Microbiology, Oncology Introduction Colorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide, representing 9.6% of all cases. It is also the second leading cause of cancer-related deaths, contributing to 9.3% of total cases1. The global incidence and mortality rates for both sexes are 18.4 and 8.1 per 100,000 people, respectively. The highest mortality rates are observed in Asia (51.1%), Europe (27.4%), and Latin American countries (8.1%)2. Specifically in Latin America, the incidence and mortality rates are 16.9 and 8.2 per 100,000, respectively. In Chile, the incidence rate is 20.3 per 100,000, with 6,778 new cases reported for both sexes, making CRC the third leading cause of cancer-related deaths, with 3330 deaths in 20222. The development of CRC is a complex and multifactorial process influenced by a combination of genetic, epigenetic, dietary, environmental, and inflammatory factors3,4. Among these, the gut microbiome emerges as a critical mediator, linking environmental exposures to the formation of colorectal polyps. This connection is evident through the observed correlations between environmental factors and both compositional and functional changes in the colonic microbial community5. Alterations in the gut microbiome have been implicated in the development of colorectal polyps, with specific microbial species identified as potential contributors to oncogenesis5,6. The progression of CRC typically begins with dysplastic precursor lesions in the intestinal tract, which accumulate genetic mutations over time, ultimately driving their transformation into malignant tumors5. These precursor lesions are categorized into two types: tubular adenomas (TAs), which include tubular, tubulovillous, and villous subtypes, and sessile serrated adenomas (SSAs), encompassing both SSAs and traditional serrated adenomas5. Notably, colorectal polyps that develop through the conventional adenoma-carcinoma sequence, progressing from normal mucosa to early adenoma, advanced adenoma, high-grade dysplasia, and ultimately invasive carcinoma, are frequently associated with malignancy. Among these, TAs represent a significant subset and have been strongly associated with disruptions in the gut microbiome6. In this context, the human gut microbiota plays a pivotal role in the development of colorectal polyps. This complex microbial community comprises over 100 trillion microorganisms, predominantly from the phyla Firmicutes and Bacteroidetes, which together account for more than 90% of the endogenous bacteria found in healthy adults7,8. Although the Firmicutes-to-Bacteroidetes ratio has long been used as a broad descriptor of gut microbial composition, its biological and clinical interpretation remains limited. Because this ratio varies substantially among healthy individuals and is influenced by host-related, environmental, and methodological factors, it should not be considered a reliable indicator of microbiome dysbiosis or disease status9. The gut microbiota plays a crucial role in maintaining intestinal homeostasis by regulating mucosal barriers, immune responses, and metabolic functions10. Globally, more than 16% of cancer cases have been attributed to infectious agents, emphasizing the role of microorganisms in carcinogenesis11. Studies have shown that dysbiosis, or an imbalance in microbial composition, is associated with the development of CRC, with significant differences in microbiota composition observed between healthy individuals and patients with adenomas or CRC12. Fecal metagenomic profiling has linked specific gut microbial changes to CRC patients across independent cohorts worldwide. Stool samples from CRC patients showed enrichment of Fusobacterium nucleatum, Peptostreptococcus stomatis, Parvimonas micra, Porphyromonas asaccharolytica, Solobacterium moorei, and Gemella morbillorum, with a parallel reduction of Roseburia intestinalis and Eubacterium rectale, establishing that these alterations have potential value for non-invasive CRC detection13. Cross-cohort meta-analyses integrating geographically diverse populations have confirmed the enrichment of these taxa signatures and further demonstrated consistent depletion of Faecalibacterium prausnitzii, Roseburia intestinalis, and Ruminococcus bromii, alongside enrichment of the choline trimethylamine-lyase gene cutC, supporting a link between CRC-associated microbiota and altered metabolic pathways14. A large meta-analysis of multiple cohorts from five countries established a core CRC-associated microbial signature, including Fusobacterium nucleatum, Parvimonas micra, and Peptostreptococcus stomatis, and extended these findings by enrichment of Hungatella hathewayi and depletion of Roseburia hominis and Coprococcus comes15. Multi-omics analyses in a Japanese colonoscopy cohort have shown that species such as Atopobium parvulum and Actinomyces odontolyticus are already enriched at the adenoma stage, while Fusobacterium nucleatum and other oral-associated species accumulate progressively during the transition to invasive carcinoma. At the same time, Faecalibacterium prausnitzii and Bifidobacterium species declined from the adenoma stage onward, coinciding with increases in branched-chain amino acids and the secondary bile acid deoxycholate16. Microbial species frequently detected in precancerous lesions and implicated in colorectal carcinogenesis include Escherichia coli (E. coli)17, Bacteroides fragilis (B. fragilis)18,19, and Streptococcus gallolyticus20. Additionally, shifts in Fusobacterium mortiferum and pro-inflammatory bacteria (e.g., Bilophilia and Desulfovibrio), along with decreases in short-chain fatty acid (SCFA)-producing bacteria such as Faecalibacterium prausnitzii and Bifidobacterium pseudocatenulatum, have been observed in adenoma patients21,22. The interactions between the microbiota and its host play a crucial role in understanding gastrointestinal diseases. In recent years, advances in sequencing technologies, metagenomics, and bioinformatics have provided new tools to study gut microorganisms. However, knowledge about the colon-associated microbiome in the Chilean population and its role in precancerous lesions remains limited. To address this gap, this study aimed to characterize the microbiota associated with colonoscopy findings in participants with precancerous lesions (polyps) using 16S rRNA sequencing. The results revealed a specific microbial signature associated with the healthy gut, showing significance when compared to those with precancerous lesions. Notably, an increase in the Proteobacteria phylum and a decrease in Firmicutes were observed in participants with adenoma compared to the control group. Furthermore, the study identified the exclusive presence and increased abundance of the genus Desulfovibrio in stool samples from Chilean participants with colonic precancerous lesions. This study highlights distinct microbial shifts in participants with intestinal precancerous lesions, underscoring the potential of microbiome and selected microorganisms as targets for understanding early stages of colorectal disease. Materials and methods Study participants Participants in this study include men and women older than 18 years, with or without a diagnosis of a gastrointestinal premalignant lesion (polyp) at the time of colonoscopy (N = 27). Each participant underwent a routine colonoscopy assessment in the endoscopy unit of Fundación Arturo López Pérez (FALP) OECI Cancer Center between 2021 and 2023. Colonoscopy indications included screening, abdominal pain, rectal bleeding, polyp surveillance, and suspected neoplasia. Recruitment was performed in a clinical endoscopy setting; the control group consisted of colonoscopy-negative participants not population-based healthy volunteers. Control participants had no relevant colonoscopic findings and did not require biopsy, whereas the precancerous group included participants with histologically confirmed low-grade adenomas. Each participant was informed, signed a consent letter, and authorized access to clinical data, pathology reports, and 16S rRNA gene sequencing data from stool samples, according to the instructions from a physician. All biological material and its associated information were encrypted according to our internal procedure to ensure confidentiality and restrict access to patients’ information. All experimental protocols of this study were approved by the Ethics Committee of Fundación Arturo López Pérez (Santiago, Chile) (protocol ID: 2020-009-RES-SIN-INT, 11/08/2020). Written informed consent was obtained from all participants. All methods were carried out in accordance with the Declaration of Helsinki and relevant guidelines and regulations. Samples The consent letter involved the collection of a stool sample 10 days before the colonoscopy procedure. Participants with diarrhea episodes or antibiotic/probiotic consumption within the last 30 days before the stool sampling were excluded from the study. Stool specimens were freshly collected in a sterile plastic container without preservatives added, then 100 mg were transferred to a sterile tube, stabilized with buffer containing 500 mM TRIS, pH 9.0; 150 mM EDTA; 10 mM NaCl (0,2 µM filtered), and mixed thoroughly to obtain a stool homogenate to a final concentration of 100 mg/ml, following methodology of Deuter et al.23. Extraction and purification of DNA DNA was extracted from 100 mg of stool using the Applied Biosystem MagMaxMicrobiome Ultra Nucleic Acid Isolation Kit, according to the manufacturer’s instructions. The quality of the extract was verified on a 0.8% agarose gel stained with GelRed (Biotium), its concentration was measured on the Qubit 3.0 Fluorometer, using the Qubit dsDNA 1X Assay kit (Q33231, Invitrogen), and its purity was determined by the ratio of absorbance at 260 nm and 280 nm using a spectrophotometer (Cytation 5, BioTek). DNA was stored at -80 °C until it was used. 16S rRNA gene sequencing Three nanograms of stool total DNA were used for amplification of the V2, V4, and V8 hypervariable regions of the 16S rRNA gene using one primer set included in the Ion 16S Metagenomics Kit (A26216, Thermo Fisher Scientific). The following conditions were used for amplification of the hypervariable regions: one cycle of 95 °C for 10 min, followed by twenty-five cycles of 95 °C for 30s, 58 °C for 30s, 72 °C for 20s, and a hold at 72 °C for 7 min, with a final hold at 4 °C. The PCR products were verified by a 2.5% agarose gel and purified using Agencourt AMPure XP Reagent (Beckman Coulter, Inc., Atlanta, Georgia) following the manufacturer’s instructions. 100 ng of PCR products from each sample were used to create the DNA library with Ion Plus Fragment Library Kit (4471252, Ion Torrent). The following ligate and nick repair is performed using the Ion Xpress Barcode Adapters 1–16 Kit (4471250, Ion Torrent), with one cycle of 25 °C for 15 min, followed by one cycle of 72 °C for 5 min, and a hold at 4 °C. The amplicon libraries were purified using Agencourt AMPure XP Reagent (Beckman Coulter, Inc., Atlanta, Georgia) following the manufacturer’s instructions. The final concentration of amplicon libraries was quantified using the Ion Universal Library Quantitation Kit (A26217, Thermo Fisher Scientific) on the QuantStudio 3 thermocycler (Applied Biosystems, Waltham, MA, USA). Quantified libraries were diluted to 20 pM and pooled in equimolar ratios for normalization. Template preparation, chip loading, and sequencing Emulsion PCR, bead enrichment, and chip loading were automatically performed on an Ion Chef instrument (Thermo Fisher Scientific) using Ion 510 & Ion 520 & Ion 530 kit (A34019, Thermo Fisher Scientific). The Ion 510 chip (A34288, Thermo Fisher Scientific) was used. A planned run was created for each chip within Ion Torrent Suite Software v.5.12.2 (Thermo Fisher Scientific). From pooled libraries, 25 µl was transferred to library sample tubes in Chef reagent cartridges. A control sample containing E. coli DNA (4458450, Applied Biosystems) was used as an internal control. Finally, sequencing was carried out on the Ion GeneStudio S5 Plus Sequencer (Thermo Fisher Scientific) using Ion 510 & Ion 520 & Ion 530 kit with read length set at 400 and number of flows at 850. Bioinformatic data and statistical analyses Raw data were primarily processed with GeneStudio S5 sequencer and Ion Reporter server (Thermo Fisher Scientific). Microbiome profiles were predicted using the Ion Reporter 16S Metagenomics workflow (w1.1) using the default parameters installed on Ion Reporter software v5.12 (Thermo Fisher Scientific) following the methodology used by Know et al.24. The Ion Reporter workflow performs an analysis of the microbial composition of samples using the Quantitative Insights into Microbial Ecology (QIIME) software. The obtained sequences were mapped to the Greengenes v.13.5 database for taxonomic assignment into operational taxonomic units (OTUs). Taxonomic identification was based on the V2, V4, and V8 hypervariable regions of the 16S rRNA gene. Alpha diversity was represented by the Chao1 index and the Shannon index. Principal Component Analysis (PCA) was performed to explore patterns in microbial composition and assess global differences between control and precancerous groups, using relative abundance data at the family, genus, and species levels. Between-group differences were assessed using the Mann-Whitney test. A two-way ANOVA with mixed-effects model was applied to compare the mean relative abundance of bacterial taxa between study groups (control vs. precancerous), with bacterial taxon as the within-subjects factor. This approach accounts for the within-subject correlation structure arising from multiple taxa measured within the same individual, who share the same host environment and microbial ecosystem. Post-hoc comparisons were performed using a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli. The critical value for statistical significance was established as p < 0.05. The specific significance of some assays was marked with asterisks, and their values were indicated in each figure. Fisher’s contingency was used to assess whether the proportion of participants with high Proteobacteria abundance differed between the control and precancerous groups. An orthogonal methodology was considered for genus and species level to complement the previous PCA analysis, microbiome differential abundance analysis between the precancerous and control groups (N = 27) was estimated using Deicode25 and ANCOM-BC26 plugins accessed through QIIME2 platform27 at the genus and species levels, with default parameters and a minimum taxonomic abundance of 0.05, this orthogonal methodology takes into account the compositional nature of microbiome data26. Visualizations of Log Fold Change (Log2FC) results, including p-values, q-values and 95% confidence intervals, are represented (Supplementary Table S2 and S3). Raw 16S rRNA sequencing data are available at the Sequence Read Archive (SRA) under accession number PRJNA1381109. Results Adenoma presence correlates with specific genus not affecting the overall gut microbiological diversity We analyzed and compared the microbiological composition of fecal samples, representing the gut microbiome, from 27 participants who underwent colonoscopy at a single center (Table 1). Based on the clinical report from the gastroenterologist’s colonoscopy and the histopathological diagnosis from the pathologist, the participants were divided into two groups. The control group included 14 participants (78.60% women) with no significant findings during colonoscopy and who did not require a biopsy, showing a mean age of 56.8 ± 13.8 (95% CI 49.6–64.0) years old, with no intestinal lesions or history of diarrhea. On the other hand, the precancerous group included 13 participants (53.80% women) with a mean age of 58.2 ± 12.2 (95% CI 51.6–64.9) years and diagnosed with the presence of polyps (sessile or adenomatous) on the surface of the colon or rectum. All observed polyps were removed during the colonoscopy procedure for histological assessment, showing low-grade dysplasia (Table 2). The details of 16S rRNA sequencing data, including the total and valid reads per sample, were reported in supplementary data (Table S1). The average of valid reads was 64,386, with a minimum of 30,792 reads. A minimum of 10,000 reads per sample was set as sampling depth for the analysis. PCA was performed to explore patterns in microbial composition and assess global differences between control and precancerous groups, using relative abundance data at the family, genus, and species levels; however, the score plot did not separate the two groups and only shows high variability among samples from the precancerous group (Figure S1A). Additionally, a biplot of the bacterial family, genus, and species distribution for the entire cohort showed that families such as Enterobacteriaceae, Ruminococcaceae, and Fusobacteriaceae exerted a strong influence on the data structure. This is reflected by their longer vectors projecting toward the positive side of PC1 (Figure S1B). Table 1Clinical and demographic characteristics of patients.CategoryID patientAge (years)Age (mean, SD)GenderColonoscopy findingsBiopsyControl12836656.8 ± 13.8MaleNoneNo129028FemaleNoneNo129261MaleNoneNo129953FemaleNoneNo130178FemaleNoneNo130352FemaleNoneNo130469FemaleNoneNo130835FemaleNoneNo136163MaleNoneNo136567FemaleNoneNoS59045FemaleNoneNoA16261FemaleNoneNoA43066FemaleNoneNoB68151FemaleNoneNoPrecancerous12916158.2 ± 12.2MaleMultiple unspecified polypsYes129469FemaleSingle unspecified polypYes129570MaleSingle unspecified polypYes130642FemaleSingle unspecified polypYes131366FemaleMultiple unspecified polypsYes131472MaleSingle unspecified polypYes131647FemaleMultiple unspecified polypsYes135539MaleSingle unspecified polypYesL43068MaleSingle unspecified polypYesG92254FemaleSingle unspecified polypYesM85671MaleSingle unspecified polypYesF10755FemaleSingle unspecified polypYesS23443FemaleSingle unspecified polypYes Clinical and demographic characteristics of patients. Table 2Characteristics of precancerous group.Clinical findingsParticipantsTotal cases13Individuals with single lesion10 (76.9%)Individuals with multiple lesions3 (23.1%)Polyps ≥ 0.5 cm7 (53.8%)Polyp location Sigmoid colon5 (38.5%) Descending colon3 (23.1%) Transverse colon2 (15.3%) Tranverse and descending colon1 (7.7%) Ascending and tranverse colon1 (7.7%) Rectum1 (7.7%) Tubular adenoma with low-grade dysplasia13 (100%) Characteristics of precancerous group. The composition of the gut microbiome for each group was examined at different taxonomic levels, including its global diversity and distribution within the studied cases. The microbiome profile from all participants showed the four phyla typically found in a healthy human gut, dominated mainly by Firmicutes and Bacteroidetes, followed by Actinobacteria and Proteobacteria (Fig. 1A). Gut microbiome profile and diversity in stools from participants with precancerous lesions compared to a healthy control group. From the total valid counts, the average relative abundance at the phylum level for each group was calculated (A). The contribution of the two major phyla in each group was expressed as the sum between them (B) and the Firmicutes to Bacteroidetes ratio (C). Alpha diversity, as estimated by the Chao1 and Shannon indices, and beta-diversity, as measured by Bray–Curtis dissimilarity using PCoA analysis, were assessed at the genus (D) and species (E) levels. (F) Bars display the contribution of the observed phyla within each sequenced stool sample, and (G) displays the contribution of the grouped cases within each phylum. The results are presented as mean ± SD Asterisks indicate significant differences between the precancerous and control groups, analyzed by two-way ANOVA. **q-value < 0.05; *q-value < 0.01. ns = not significant. The total contribution of Bacteroidetes and Firmicutes was 83.4% in the control group and 73.4% in the precancerous group, without statistically significant differences (Fig. 1B). Although the Firmicutes to Bacteroidetes (F/B) ratio was lower in the precancerous group compared to the control group, this difference did not reach statistical significance (Fig. 1C). Regarding microbiome diversity, this was assessed in terms of the relative abundance within the specific clinical groups at the genus and species levels. In both cases, the alpha diversity results displayed no significant differences for the Shannon and Chao1 index between the two groups evaluated, including the beta diversity, calculated based on Bray-Curtis dissimilarity index, which showed no differences when it was expressed as an analysis of multiple components (PCoA) at genus and species level (Fig. 1D and E, respectively). At the individual-sample level, the stacked bar plots showed that phylum-level composition varied across participants in both groups, although Firmicutes and Bacteroidetes remained the predominant phyla in most samples (Fig. 1F). At the phylum level, Firmicutes showed a mean relative abundance of 53.70% (95% CI: 47.1–60.3) in the control group and 41.90% (95% CI: 33.5–50.2) in the precancerous group, representing a significant reduction of 11.80% (p-value = 0.0320, q-value = 0.0336). In contrast, the Proteobacteria displayed a mean abundance of 7.00% (95% CI: 4.2–9.7) in control and 21.70% (95% CI: 12.4–30.9) in the precancerous group, with a statistically significant increase of 14.70% (p-value = 0.0073, q-value = 0.0076) (Fig. 1G). No significant differences were observed for the Bacteroidetes with 29.80% (95% CI: 25.4–34.2) in control group and 31.50% (95% CI: 26.9–36.1) in the precancerous group, and the Actinobacteria phylum with 9.20% (95% CI: 4.3–14.1) in control group and 3.10% in the precancerous group (95% CI: 0.9–5.3) (Fig. 1G). These results suggest that, while the overall composition of the gut microbiome retains similar characteristics, changes in specific phyla, such as Firmicutes and Proteobacteria, may be associated with adenoma development. Common microorganisms display a unique profile pattern according to colonic health Overall, at the genus level, 76 genera were observed, but only 57 (75.00%) of them were shared between both groups, and from the remaining 19 mutually excluded genera, 13 (68.00%) and 6 (32.00%) were exclusively found in the control and the precancerous group, respectively (Fig. 2A). In the overall genus-level profile, Faecalibacterium and Bacteroides were the most represented in both groups (Fig. 2A). Within the non-shared subset of genera exclusive to the precancerous group, Succinivibrio was the most represented, with a mean relative abundance of 3.40%; however, it should be noted that this reflects its share within the non-shared subset and does not indicate dominance of the overall stool community. Comparing between groups as compositional and as abundance-based data, differences were observed for some genera related to intestinal health, such as Faecalibacterium and Bifidobacterium, which showed a notable decrease in the precancerous group, with Faecalibacterium dropping from 28.20% to 19.90% and Bifidobacterium declining from 9.96% to 1.66% (Figure S2). Anaerostipes was significantly reduced in the precancerous group relative to the control group, with a mean relative abundance of 0.13% and 0.04%, respectively, a 70% decrease (p-value = 0.0297, q-value = 0.0312). On the other hand, Desulfovibrio was significantly enriched in the precancerous group, with a mean relative abundance of 3.34% compared with 0.86% in controls, a nearly four-fold increase (p-value = 0.0249, q-value = 0.0261) (Fig. 2A). Furthermore, genera associated with pathogenic activity were enriched in the precancerous group. Prevotella increased from 1.06% to 7.71% and Fusobacterium from 0.06% to 3.20% (Fig. 2A). However, these bacterial genera were not significant in the compositional analyses (Table S2). Specific shared bacterial taxa differentially enriched or depleted in stools by colonic condition. Mean relative abundance at genus (A) and species (B) levels was calculated for each group and compared between healthy controls and precancerous cases to identify shared taxa. The number of common microbial taxa are shown as Venn diagrams depicting the overlap of genera (A) and species (B). The significant taxa were individualized and compared at the genus (A) and species (B) levels between groups. All data are presented as mean ± SD. Asterisks indicate significant differences between precancerous and control groups, analyzed by two-way ANOVA. *q-value < 0.05. At the species-assignment level, 161 were observed, with 96 (60%) shared between groups, 38 (23.60%) found only in the control group, and 27 (16.80%) found exclusively in the precancerous group (Fig. 2B). Among these species, the most abundant microorganisms were Bacteroides barnesiae found only in the control group, with a mean of 1.40%, and for the precancerous group were Succinivibrio dextrinosolvens and Desulfovibrio D168, with abundances of 3.40% and 1.40%, respectively. Integrated analysis at the species level identified taxon-level differences in stool microbiota composition between the control and precancerous groups, combining relative abundance comparisons with compositional analysis. A marked decrease in Faecalibacterium prausnitzii was observed, with its mean abundance dropping from 45.20% in the control group to 26.30% in the precancerous group (p-value = 0.0267, q-value = 0.0280). Similarly, Clostridium spshowed a reduction from 1.07% to 0.26% (p-value = 0.0141, q-value = 0.0148), Eubacterium ramulus decreased from 0.21% to 0.04% (p-value = 0.0186, q-value = 0.0196), and Roseburia sp. decreased from 0.19% to 0.01% (p-value = 0.0037, q-value = 0.0039) (Fig. 2B). Compositional analysis supported this trend, indicating significant decreases in several species, including Clostridium sp. (Log2FC of -2.3, p-value = < 0.001, q-value = 0.0413), and Faecalibacterium prausnitzii (Log2FC of -0,68, p-value = 0,0006, q-value = 0,0685), exclusively in the precancerous group (Table S3). Furthermore, several species-level taxa showed higher mean relative abundance in the precancerous group, including Prevotella copri (8.30%), Bacteroides eggerthii (3.40%), and Fusobacterium mortiferum (2.50%) (Figures S3). Together with the lower abundance of taxa commonly associated with intestinal homeostasis, including Faecalibacterium prausnitzii and Bifidobacterium. These findings suggest taxon-level differences in stool microbiota composition among participants with adenomas. However, these results should be interpreted with caution, particularly for taxa detected in a limited number of participants and should be considered exploratory rather than evidence of a defined dysbiotic state or a causal role in adenoma development. When group-exclusive taxa were evaluated descriptively, control-specific genera were mainly represented by members of the Firmicutes phylum (86.50%), being the most abundant Bulleidia (31.99%), Tyzzerella (28.74%), Succinatimonas (15.86%), and Turicibacter (5.12%), followed by Faecalitalea (4.79%) from the Actinobacteria phylum (Fig. 3A). Whereas taxa detected only in the precancerous group included a higher contribution of Proteobacteria-associated members, such as, Succinivibrio (94.72%), followed by Mitsuokella (4.90%) from the Firmicutes phylum (Fig. 3B). In addition, we observed that 64.90% of the species detected in this group belonged to the phylum Proteobacteria, with Succinivibrio dextrinosolvens (48.73%) being the most abundant species, followed by Desulfovibrio D168 (16.16%). Another representative but less species included Acidaminococcus fermentans (11.03%), Bacteroides coprocola (10.91%), and Collinsella tanakaei (6.09%) (Fig. 3B). Proportional distribution of group-exclusive microbial taxa at the genus and species levels. Pie charts show the mean proportional contribution of taxa detected exclusively in the control group (A) or in the precancerous group (B). Percentages were calculated relative to the total reads assigned to non-shared taxa within each group. The Proteobacteria phylum is dramatically influenced by the histological health of the colonic epithelium Participants with adenomas displayed a unique microbiome profile characterized by a Firmicutes to Proteobacteria (F/P) ratio significantly lower than the control group (p-value = 0.0125) (Fig. 4A). Additionally, when these cases were segregated into low- and high-Proteobacteria relative abundance groups using the cohort median (0.30%) as cut-off. Participants with Proteobacteria abundance above this cut-off were more frequently represented in the precancerous group, as assessed by Fisher’s exact test (p-value = 0.0070) (Fig. 4B). Three specific classes of proteobacteria ( β, δ, and γ) were identified (Fig. 4C), and each class showed a differential relative abundance above the control group, 0.45%, 1.60%, and 1.90% for γ-proteobacteria, β-proteobacteria, and δ-proteobacteria, respectively (Fig. 4C). β-proteobacteria were 1.10% more abundant than γ-proteobacteria, and δ-proteobacteria were 1.40% more abundant than γ-proteobacteria (Fig. 4C). Furthermore, the relative abundance of δ-proteobacteria in the precancerous group was 3.3-fold more than the control group (p-value = 0.0135, q-value = 0.0141) (Fig. 4C). Specifically, in the precancerous group, the δ-proteobacteria genus was represented by Desulfovibrio, showing a mean relative abundance of 3.40%, and it was statistically significantly higher compared to the control group (0.86%, p-value = 0.0249, q-value = 0.0261) (Fig. 4D). Proteobacterial genera are the most meaningful bacterial taxa in stools from participants with a precancerous lesion. The contribution of the Proteobacteria phylum in the microbial stool composition was represented as the ratio between Firmicutes and Proteobacteria, presented as mean ± SD, and analyzed using the Mann-Whitney test, *p-value < 0.05 (A) and the frequency of cases with abundances above the median (high) for each group using a 2 × 2 contingency analysis (B). The relative abundance of each proteobacteria class was segregated by case and group and presented as mean ± SD, and analyzed with two-way ANOVA (C). The contribution by the Proteobacteria class was represented as a heat map, with the relative abundance value of each bacterial taxon normalized to a z-score, at the genus (D) and species level (E). Asterisks indicate significant differences as compared to the precancerous and control groups. *q-value < 0.05. ns = not significant. At the species-assignment level, Desulfovibrio intestinalis showed a significant difference between groups, with a mean relative abundance of 0.0011% in the controls and no detection in the precancerous group (p-value = 0.0042, q-value = 0.0044) (Fig. 4E). This result indicates that not all Desulfovibrio-assigned taxa followed the same pattern observed at the genus level. Although the Proteobacteria-focused analysis supports an increase in delta-proteobacteria and Desulfovibrio in participants with precancerous lesions. Discussion The integration of clinical data and the robustness of the metagenomics dataset allow us to establish a baseline of the gut microbiome measured from stools in those participants with a diagnosis of colon adenomas, and to compare it to a healthy control group. This study provides the first report on microbial communities associated with precancerous lesions from a transversal cohort of Chilean participants from a single center. Firmicutes, Bacteroidetes, Actinobacteria, Proteobacteria, Fusobacteria, and Verrucomicrobia are widely recognized as the dominant phyla in a healthy intestine28. Firmicutes and Bacteroidetes collectively constitute approximately 90% of the gut microbiota7,8. Our small cohort showed comparable results to those observed for bigger cohorts, confirming the representativity of the population at the level of microbial communities for a healthy gut, including those frequently used to estimate dysbiosis of the gut microbiome29–32. At the levels of alpha and beta diversity, our results did not reveal significant differences to distinguish participants based on their clinical status (precancerous). Similar findings have been reported using feces and mucosal samples from patients with or without colonic lesions21,33–36. However, despite the absence of differentiation in terms of overall diversity, our data suggests notable differences in the relative abundance of specific genera and species that are shared between participants with precancerous lesions and healthy controls, as well as others that are mutually excluded. Regarding a comparative baseline for the Chilean gut microbiome, just two studies are reported37,38. In both cases, the data were collected from asymptomatic participants without a colonoscopy assessment and were assigned as healthy. In specific, the analysis of stool samples from a cohort of 41 young, asymptomatic, normal weight women volunteers with a mean age of 25 ± 4 years showed that the microbial profile of a healthy Chilean colon is composed primarily of Firmicutes (43.60 ± 9.20%) and Bacteroidetes (41.60 ± 13.10%), followed by Verrucomicrobia (8.5 ± 10.4%), Proteobacteria (2.80 ± 4.80%), Actinobacteria (1.80 ± 3.90%) and Euryarchaeota (1.40 ± 2.70%), with an estimated Firmicutes to Bacteroidetes ratio around 1.05 ± 0.737. In addition, using pyrosequencing of colonic mucosal biopsies from a mixed ethnic group of 12 healthy participants (5 Chileans and 7 Spanish) with a mean age of 55.9 ± 15.7 years revealed that the microbial composition within the tissue included the four major phyla, Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria38, not comparing with stool samples. Despite methodological differences, we observed the presence of the most characteristic phyla previously reported for Chilean people; therefore, these studies collectively provide valuable insights into the microbial composition of the Chilean population and its potential implications for health and disease. On the other hand, the role of the gut microbiome in the initiation and progression of colorectal cancer (CRC) has been evaluated in a cohort of 206 Chinese individuals, revealing significant shifts in microbial composition. The cohort included 59 participants with polyps or adenoma with low-grade dysplasia, 54 cases with adenoma and high-grade dysplasia, 51 CRC patients, and 42 healthy controls. Sequencing analysis of 16S rRNA showed a progressive decrease in the relative abundance of Bacteroidetes (52.14%-53.92% versus 52.46%–47.06%) and Firmicutes (35.88%-29.73% versus 24.27%–25.36%) from health to polyp to adenomas to CRC stages39. Beneficial bacteria, including butyrate-producing genera and species like Bifidobacteria, Roseburia, Faecalibacterium prausnitzii, and Eubacterium hallii, contribute to maintaining intestinal homeostasis, enhancing gut barrier integrity, and competing effectively with pathogenic bacteria, thereby promoting overall intestinal health12,40. Our results demonstrated a significant reduction in the relative abundance of several of these beneficial bacteria in individuals with precancerous lesions. This pattern was particularly evident among Firmicutes members with reported roles in butyrate production, dietary compound metabolism, and microbial cross-feeding. Members of the genus Roseburia are obligate anaerobic bacteria within the family Lachnospiraceae that ferment dietary complex carbohydrates to produce butyrate as their primary metabolic end-product41,42. In our cohort, Roseburia sp. was markedly reduced in the precancerous group, consistent with previous reports showing depletion of butyrate-producing Roseburia species as a defining feature of dysbiosis in ulcerative colitis43 and colorectal cancer15. Eubacterium ramulus was also significantly reduced; this strict anaerobe participates in the intestinal degradation of dietary flavonoids such as quercetin and luteolin44. At the genus level, Anaerostipes was similarly reduced, which may be relevant because Anaerostipes species can convert lactate and acetate into butyrate through microbial cross-feeding45. Clostridium sp. was also significantly reduced, consistent with previous evidence of a lower prevalence of Clostridium spp. In advanced colorectal adenoma, together with reduced fecal short-chain fatty acid (SCFA) production46. Moreover, our findings demonstrated an increase in the Proteobacteria phylum in all participants with a confirmed diagnosis of precancerous lesions. Specifically, we observed a significant increase in the deltaproteobacteria class, and in the genus Desulfovibrio. This finding is consistent with a study performed in 206 Chinese individuals to determine the role of the gut microbiome in the initiation and progression of colorectal cancer, showing that Proteobacteria abundance increases with the severity of the lesions and tumor progression39. A much bigger cohort, involving 233 patients with an adenoma bigger than 1 cm, has confirmed that these patients had a higher representation of multiple OTUs within the Bacteroidetes phyla and δ-proteobacteria class, including Bilophila, Desulfovibrio, Sutterella, and Mogibacterium genera, than the healthy group, which included more than 500 controls21. In this context, the enrichment of Desulfovibrio in Chilean participants with colorectal adenomas is relevant. Members of this genus include sulfate-reducing bacteria capable of metabolizing sulfur-containing compounds and producing hydrogen sulfide (H2S)47. H2S has been reported to induce DNA damage and transcriptional changes in human intestinal epithelial cells48. However, this observation should be interpreted with caution. The cross-sectional design of this study precludes establishing causality, and the observed enrichment may reflect changes in the local microenvironment associated with adenomas rather than a direct role in their development. While previous studies have suggested a potential link between sulfate-reducing bacteria and colorectal carcinogenesis, the current evidence remains insufficient to support a direct mechanistic association with early neoplastic lesions. Further studies integrating functional approaches, such as metagenomics or metabolomics, are required to clarify the role of this genus in colorectal tumorigenesis. This study has several limitations. First, the number of sequenced cases was relatively small, and the recruited participants were from a single oncology center, where the availability of asymptomatic individuals without relevant colonoscopy findings was limited. This extended the recruitment period and limited the external validity of the results. Therefore, the control group should be interpreted as colonoscopy-negative participants recruited in a clinical endoscopy setting, rather than population-based healthy volunteers, and our findings should not be considered representative of the broader Chilean population. Second, the cohort showed heterogeneity in age and sex distribution between groups, both of which may influence stool microbiota composition. Because of the limited sample size, robust adjustment for these variables and other potential cofounder was not feasible. Third, relevant lifestyle and environmental factors, including habitual diet, fiber intake, body mass index, smoking, alcohol consumption, physical activity, sleep patterns, medication use, and other host-related variables, were not comprehensively available. Thus, some of the microbial differences observed between groups may reflect unmeasured exposure in addition to adenoma status. Finally, fecal microbiota captures only part of the intestinal microbial ecosystem and may not fully represent mucosa-associated communities located near colorectal lesions. In addition, species-level assignments based on 16S rRNA sequencing should be interpreted cautiously because of the limited taxonomic resolution of this approach. In this regard, there is no universal sequence-similarity threshold capable of accurately separating all taxonomic levels, as bacterial lineages evolve at different rates49. This limitation is particularly relevant in our study, since taxonomic assignment was based on selected hypervariable regions, V2, V4, and V8, rather than full-length 16S rRNA sequences or shotgun metagenomics. Therefore, some species-level classifications should be considered approximate taxonomic assignments rather than definitive identifications. Future multicenter studies with larger and demographically balanced cohorts, paired stool and mucosal samples, and detailed clinical, dietary, and lifestyle metadata will be needed to validate and extend these findings. Conclusions This is the first characterization of the gut microbiome from stools in Chilean people related to the histological and clinicopathological status of the colon epithelium at the moment of sampling. The observed profile segregates people diagnosed with the presence of low-grade dysplasia adenomas from individuals with a healthy colon based on the enrichment of the Proteobacteria phylum, which is driven by the increased abundance of Desulfovibrio, a member of the δ-proteobacteria class, suggesting that these microorganisms may be involved in the pathogenesis of polyps and may be useful as a potential non-invasive biomarker. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Acknowledgements Special thanks to the participants and their families. Also, thanks to Dr. Andres Hernandez for submitting the BioProject to NCBI. Author contributions RG-S, TE, RJ, and FV-E designed the study. CB, EA, ML, and JPA collected the samples and associated clinical information. TE performed the experiments and next-generation sequencing (NGS). TE, RG-S, RJ, RL, and FV-E analyzed and interpreted the data. RG-S drafted the manuscript. RG-S, TE, FV-E, and ES-M critically reviewed and approved the final manuscript. ES acquired the funding. All authors revised and approved the final manuscript. Funding This work was supported and financed by Fundación Arturo López Pérez (FALP) Cancer Center (LMT-2020 & LMT-2021) and the National Agency for Research and Development by the National Fund for Scientific and Technological Development, ANID-FONDECYT grant No. 1221415 (to FV-E). Data availability 16S rRNA gene sequencing data have been submitted to the NCBI BioProject database under accession number PRJNA1381109. Declarations Competing interests The authors declare no competing interests. Ethics approval The entire study was previously approved by the Scientific Ethics Committee (CEC) of the Fundación Arturo López Pérez, as the name “PROTOCOLO 2020-009-RES-SIN-INT”. Written informed consent was obtained from all participants. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Roxana González-Stegmaier and Troy Ejsmentewicz. Contributor Information Evelyn Silva-Moreno, Email: evelyn.silva@falp.org. Franz Villarroel-Espíndola, Email: villarroele.franz@gmail.com. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Data Availability Statement 16S rRNA gene sequencing data have been submitted to the NCBI BioProject database under accession number PRJNA1381109.