Deciphering post-surgery gut microbial dynamics in colorectal cancer through multi- cohort machine learning

Deciphering post-surgery gut microbial dynamics in colorectal cancer through multi- cohort machine learning | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Deciphering post-surgery gut microbial dynamics in colorectal cancer through multi- cohort machine learning Mutebi John Kenneth, Chin-Chia Wu, Chuan-Yin Fang, Michael W Y Chan, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7866815/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Surgical resection remains the primary treatment for colorectal cancer (CRC), yet its influence on the postoperative gut microbiota remains incompletely understood. In this study, we analyzed the gut microbial communities before and after surgery from our study cohort and integrated findings from four independent CRC datasets to enhance robustness. Our results revealed that post-surgery samples had a reduced microbial diversity but were enriched with commensal taxa, suggesting a potential re-establishment of beneficial microbiota following tumor removal. Leveraging machine learning and Explainable Artificial Intelligence (XAI) through SHapley Additive exPlanations (SHAP), we identified potential postoperative microbial biomarkers, notably Akkermansia , among the dominant commensal bacteria enriched in post-surgery. Collectively, these findings highlight suggest that surgical resection may promote a favorable shift in gut microbial composition and this could guide targeted microbial modulation to improve postoperative recovery. Our study lays the groundwork for microbiota-informed strategies aimed at improving clinical outcomes in CRC patients after surgery. Colorectal cancer Surgical resection Postoperative gut microbiota Commensal restoration Microbial biomarkers Explainable AI Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction Colorectal cancer (CRC) is the second leading cause of cancer mortality worldwide [1], and is influenced by various factors, including heredity, lifestyle, diet, and the host’s gastrointestinal microbiota [2]. The human gastrointestinal tract is inhabited by a vast assemblage of microorganisms, including protozoa, archaea, fungi, viruses, and bacteria [3]. This gut microbiota plays a significant role in maintaining the host’s health by regulating and modulating intestinal homeostasis and local immune response [4]. Reduced microbial diversity, often associated with gut dysbiosis, has been linked to an increased risk of developing CRC [5]. Therefore, maintaining a delicate balance of the gut microbiota is not only essential for promoting overall health but also for preventing the risk of developing CRC. Dysbiosis promotes opportunistic bacterial overgrowth, generate carcinogenic bacterial metabolites, trigger chronic inflammation, and impair mucosal barrier integrity, thereby contributing to CRC carcinogenesis [6, 7]. Additionally, certain bacteria, such as Enterotoxigenic Bacteroides fragilis (ETBF), Polyketide synthase positive (pks+) Escherichia coli , Fusobacterium nucleatum , and Enterococcus faecalis , are reportedly more abundant in CRC patients than in healthy controls, and have been associated with CRC development and progression [8–10]. This evidence supports recent efforts by scientists to harness gut microbiota as biomarkers for the diagnosis and treatment of CRC. Surgical resection is the primary treatment for CRC, particularly in patients with localized or locally advanced disease [11, 12]. However, colorectal surgery is associated with various postoperative complications, including Clostridioides difficile infection (CDI) [13], anastomotic leak, surgical site infections (SSIs), and postoperative ileus (POI), which increases patient morbidity and mortality [14]. A recent study reported that colorectal surgery has among the highest rates of SSIs of about 23% and a weighted mean of 11% [15], despite optimal and appropriate surgical techniques and sterile conditions. Gut microbiota has been proposed to have a causative role in these complications among CRC patients. For example, collagen degradation and Matrix Metalloproteinase-9 (MMP-9) activation by Enterococcus faecalis have been linked to anastomotic leak [16, 17], while increased abundance of inflammation-promoting bacteria has been linked to POI [18, 19]. This implies that gut microbiota can also be harnessed for targeted prevention and treatment of post-surgery complications in CRC patients. Recent studies have described distinct alterations in the gut microbiota of CRC patients compared with healthy individuals [20, 21]. As a result, microbial signatures with potential biomarker value for diagnosis, prognosis, and therapeutic stratification have been identified [22–24]. However, little is known about how CRC treatments, especially colorectal resection, can influence gut microbiota. Because gut microbiota can initiate and promote CRC tumorigenesis [25, 26], surgical removal of the tumor is expected to affect the associated microbiota [27]. However, it remains unclear whether surgical resection restores beneficial microbiota and normal gut homeostasis. Patients undergoing colorectal surgery often receive bowel preparation, including oral antibiotics, polyethylene glycol, and perioperative intravenous antibiotics, which can affect their gut microbial composition [28–30]. Additionally, surgery disrupts the gut’s anaerobic ecosystem, reducing obligate anaerobes and allowing facultative anaerobes to persist, which may contribute to postoperative dysbiosis [31]. Since microbial composition influences intestinal and anastomotic tissue healing [32, 33], recurrence [34, 35], and chemoresistance [36, 37], characterizing microbiota dynamics after surgery is critical. Therefore, this study was conducted to analyze fecal samples of CRC patients before and after colorectal resection, aiming to determine microbial changes associated with post-surgery in CRC patients. We hypothesized that postoperative fecal samples would contain potential microbial biomarkers that can be harnessed to improve the overall colorectal surgery outcomes. We applied machine learning (ML), eXplainable Artificial Intelligence (XAI), and SHapley Additive exPlanations (SHAP) to identify microbial features with potential relevance for postoperative CRC management. The findings of this study could foster a new era of personalized and precision medicine that is based on microbial modulation for CRC management. 2. Materials and Methods 2.1 Study design and sample collection This prospective study included 15 consecutive patients with CRC scheduled for surgery at Buddhist Dalin Tzu Chi General Hospital between June 2023 and December 2024. Eligible participants were aged 42–85 years (Table 1 ) and were able to provide written informed consent. CRC was defined as primary malignant epithelial colorectal tumors or adenomas larger than 10 mm, according to the Taiwan Society of Colon and Rectal Surgeons (TSCRS) [38]. Diagnosis was confirmed by colonoscopy to ascertain the CRC status of every participant. Table 1 Demographic overview of our study population. Variable Healthy Controls CRC Patients Number 10 15 Age, year, mean (IQR) 52.3 (25.0) 63.0 (8.5) Sex (Female/Male), n 5\5 7\8 BMI, mean (IQR) 21.9 (6.0) 23.8 (3.0) Tumor Location NA Rectum Participants with distant metastases, coexisting bowel disorders, prior emergency surgery, or receipt of chemotherapy, radiotherapy, targeted therapy, or antibiotics/probiotics within three months before enrollment were excluded. Preoperative stool samples were collected 7 days before colonoscopy (categorized as Before Surgery) and postoperative samples were collected at least three months later (categorized as After Surgery). All patients underwent standard mechanical bowel preparation with polyethylene glycol the day before surgery and received prophylactic antibiotics immediately before surgery [39]. Surgical resections were performed according to oncological principles of TSCRS. The study protocol was approved by the Institutional Review Board (IRB) of Buddhist Dalin Tzu Chi General Hospital and conducted in accordance with the Declaration of Helsinki [40]. 2.2 DNA extraction, sequencing and data processing Microbial DNA was extracted from stool samples using the QIAamp Fast DNA Stool Mini Kit (QIAGEN, Germany) according to the manufacturer’s instructions. The concentration and purity of the extracted DNA were assessed with a NanoDrop® spectrophotometer (Thermo Fisher Scientific, USA), and the integrity was confirmed by PCR amplification and agarose gel electrophoresis of the 16S rRNA gene (~ 1500 bp) [41]. The full-length 16S rRNA gene (V1–V9 regions) was amplified with universal primers 27F (5’-AGRGTTTGATYMTGGCTCAG-3’) and 1492R (5’-GGYTACCTTGTTACGACTT-3’) [42]. Amplicons were sequenced on the PacBio Sequel platform (Pacific Biosciences, USA) following the manufacturer’s protocol [43]. Circular consensus sequence (CCS) reads were generated from raw PacBio data and analyzed using QIIME2 (v2025.7) as previously reported [44]. Low-quality and chimeric sequences were trimmed with DADA2 (q2-DADA2). High-quality reads were clustered into amplicon sequence variants (ASVs) at 98% similarity [45], and taxonomy was assigned against the National Center for Biotechnology Information (NCBI) reference database [46]. Taxonomic profiles were collapsed to the genus level and normalized with respect to the cumulative microbial count for each subject. 2.3 Machine learning modeling and identification of post-surgery–associated microbial markers. We applied machine learning (ML) models to identify potential gut microbial biomarkers for colorectal post-surgery. Specifically, we trained and evaluated these models using random forest (RF), support vector machine (SVM) and XGBoost algorithms. These algorithms have been previously implemented to identify microbial biomarkers in studies involving genomic datasets [47–49]. To enlarge our dataset, four external sequence read datasets (DA1–DA4; Table S1 ) were retrieved from the NCBI Sequence Read Archive (SRA; https://www.ncbi.nlm.nih.gov/sra ), for independent studies from China [50], Spain [51], Singapore [52], and South Korea [53], respectively. The inclusion criteria were the availability of pre- and post-surgery microbiome data, associated metadata, sufficient sequencing depth, and high-quality sequence reads. Datasets DA1 and DA3 included both CRC patients and healthy controls; for consistency, only CRC patients were analyzed. Similarly, DA2 comprised 111 participants with or without colorectal anastomotic leakage (CAL). However, non-CAL patients with paired pre- and post-surgery microbiome data were analyzed. Furthermore, DA4 consisted of CRC patients with or without ileostomy and only patients from the non-ileostomy control group with paired pre- and post-surgery samples were analyzed. In all studies, the V3–V4 region of the 16S rRNA gene was amplified with universal bacterial primers and DNA amplicons were paired-end sequenced on the Illumina platform. These raw paired-end sequences were processed in QIIME2 (v2025.7) pipeline as described in Section 2.2 above. The resulting ASVs were taxonomically assigned against the NCBI reference database [46], and collapsed up to the genus level. These genera were normalized in accordance with the cumulative microbial count for each subject and standardized by z-score transformation. Sparse features (> 90% zero values) were excluded to reduce noise and then, normalized genus level data was fed into machine learning models for analysis. The analysis was implemented in Python (v3.11.4) using SciKit-learn library (v1.7.0; https://scikit-learn.org/stable/ ) in Visual Studio Code (v1.100) [54]. The dataset was split randomly into training (70%) and validation (30%) sets. Model hyperparameters were optimized using grid search with 20 repetitions of 5-fold stratified cross-validation. For SVM, the radial basis function (rbf) kernel was tuned for C (1–50) and gamma (0.001–0.1) [55]. For RF, the maximum tree depth (None, 3, 5) and number of estimators (50, 100, 150, 200, 250) were tested [56]. For XGBoost, maximum tree depth (None, 3, 5), colsample_bytree (0.1–0.9), and number of estimators (50, 100, 150, 200, 250) were optimized [57]. The primary evaluation metric was the area under the receiver operating characteristic curve (AUC); however, additional metrics including accuracy, precision, recall, and F1-score were assessed [58]. After comparing these metrics, we determined that RF was the most effective classifier for before and after surgery based on gut microbiota data. Therefore, it was utilized for subsequent model predictions and interpretability. 2.4 SHapley Additive exPlanations (SHAP) interpretability analysis To evaluate the contribution of individual gut microbial taxa to model predictions, we employed SHAP, an explainable artificial intelligence (XAI) framework based on cooperative game theory [59, 60]. SHAP provides both local (sample-specific) and global (feature-level) interpretability, making it suitable for identifying bacterial taxa that consistently influence the model outcomes [61]. Compared to conventional feature importance metrics, SHAP can account for nonlinear interactions and heterogeneous effects [62], which are common in microbiome data. SHAP analyses were conducted in Python (v3.11.4) using the shap package (v0.45.0), based on RF models. We applied the TreeExplainer function with the interventional feature perturbation setting, which reduces bias from correlated features and provides more robust estimates of contribution [63]. Feature importance rankings were obtained from the mean absolute SHAP values across all samples, and visualization plots were generated to facilitate interpretation. Consequently, several microbial genera were identified with consistent high contributions to model predictions. These genera were considered potential biomarkers of clinical importance towards post-colorectal surgery management. 2.5 Statistical analysis and data visualization Variations in the bacterial diversity based on the temporal gut microbial changes before and after surgery were assessed using Chao1 indices for alpha diversity [64], and unweighted UniFrac distances for beta diversity [65]. The top 20 bacterial species between the two groups were visualized using bar plots created in Microsoft Excel 2024. Functional metagenomic profiles were inferred with Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) [66], and correlations with species-level abundance were tested using statistical analysis of meta genomic profiles (STAMP) software (v2.1.3) [67]. Categorical variables were compared using the χ² test, and continuous variables with the Wilcoxon rank-sum test. Differences in diversity metrics were evaluated using the Kruskal–Wallis test and PERMANOVA. A p-value < 0.05 was considered statistically significant. Statistical analyses were conducted in SPSS v24 (IBM, USA). 3. Results 3.1 Variations in bacterial community structure and composition before and after colorectal surgery Bacterial community diversity before and after colorectal surgery was evaluated using both alpha and beta diversity metrics. Alpha diversity, as measured by the Chao1 index ( Fig. 1 A ) , was significantly higher in pre-surgery samples compared to post-surgery samples (Kruskal–Wallis test, p = 0.038), which indicated a reduction in within-sample richness following surgery. Conversely, beta diversity analysis based on unweighted UniFrac distances did not reveal statistically significant overall differences between pre- and post-surgery groups (PERMANOVA, p = 0.104). Nonetheless, principal coordinate analysis (PCoA) suggested partial but not complete separation between the two groups ( Fig. 1 B ) , reflecting some degree of inter-individual variability in bacterial community structure. The 16S rRNA sequencing revealed 322 classifiable bacterial genera in the fecal samples collected from study patients. To explore compositional shifts, we focused on the top 20 most abundant genera ( Fig. 2 ) . Post-surgery samples were enriched in Parabacteroides , Collinsella , and Akkermansia , while pre-surgery samples showed higher relative abundances of Fusobacterium . When compared with external reference datasets [47–49], pre-surgery samples were characterized by higher levels of Bacteroides and Streptococcus , whereas post-surgery samples were enriched in Phocaeicola , Faecalibacterium , and Ruminococcus ( Fig. 3 ) . These compositional differences were statistically supported (p = 0.044, Kruskal–Wallis test), implying that the variation in bacterial community could have been influenced by surgery. 3.2 Gut microbiota-based classification model for pre- and post-surgery colorectal surgery To assess the biomarker potential of gut microbiota for pre- and post-surgery classification, we trained three supervised classifiers (RF, SVM, XGB) and evaluated their performance using AUROC, accuracy, precision, recall, specificity, and F1 score. These evaluation metrics are represented as averages from 20 repetitions of five-fold cross-validation. As summarized in Table 2 , the RF classifier achieved the best overall performance, with an AUROC of 0.90, accuracy of 89%, and F1 score of 0.88. XGB also performed well (AUROC = 0.83, accuracy = 80%), while SVM achieved slightly higher accuracy (87%) but a lower AUROC (0.78). These findings suggest that tree-based models, particularly RF, are better suited for distinguishing pre- and post-surgery microbial profiles. Table 2 Performance comparison of Support Vector Machine (SVM), Random Forest (RF), and XGBoost (XGB) classifiers across evaluation metrics Classifier Accuracy Precision Recall F1 AUC ROC Average Precision SVM 0.7185 ± 0.1107 0.7579 ± 0.2129 0.5952 ± 0.2327 0.6276 ± 0.1836 0.7831 ± 0.1168 0.7973 ± 0.1157 RF 0.8447 ± 0.0953 0.9042 ± 0.1270 0.7457 ± 0.1847 0.7996 ± 0.1338 0.8875 ± 0.0865 0.9078 ± 0.0674 XGB 0.7750 ± 0.1012 0.7748 ± 0.1490 0.7274 ± 0.1612 0.7361 ± 0.1195 0.8316 ± 0.0948 0.8547 ± 0.0831 We further examined the RF model’s predicted probability distributions. As shown in Fig. 4 A, pre-surgery samples displayed broader and lower classification scores, whereas post-surgery samples exhibited more narrowly distributed, higher scores. This difference was statistically significant (Mann–Whitney U = 0.00, p = 0.005), indicating that a RF-based model can reliably separate the two groups based on their gut microbiota differences. 3.3 Model explainability and feature importance To identify the microbial taxa driving classification in post-surgery for potential biomarkers, we applied SHAP to the RF model. The top 10 genera ranked by average absolute SHAP values are shown in Fig. 4 B. Among these, Akkermansia , Collinsella , Dorea , and Anaerostipes were strong positive contributors to post-surgery classification, whereas lower relative abundances of Faecalibacterium and Blautia were associated with higher prediction scores. Other genera, including Parabacteroides, Roseburia , Senegalimassilia , and Bilophila , contributed less to model performance. To further validate these findings, we compared the relative abundances of selected genera among the ranked top 10 between pre- and post-surgery groups (Fig. 5 ) . Post-surgery, unlike pre-surgery samples showed higher statistically significant (p < 0.05) mean abundances of Akkermansia (6.8 Vs 2.5%), Blautia (4.1% Vs 1.9%), Faecalibacterium (5.3% Vs 2.4%), and Parabacteroides (6.7% Vs 1.1%). Conversely, pre-surgery samples had higher abundances of Bacteroides (5.8% Vs 1.5%), and Streptococcus (4.5% Vs 0.5%) as comparted to post-surgery. 3.4 Comparative analysis of microbial functional pathways before and after surgery We examined microbial functional pathways using RF feature importance analysis and compared them between pre-and post-surgery groups. A RF classifier was trained with 100 trees using a deviance splitting criterion, a minimum leaf size of 2, and feature selection at each split set to 0.577 sqrt (nvars). The classifier achieved an AUROC of 0.87, indicating good discriminatory performance. The top-ranked pathways identified by RF included several biosynthetic and metabolic processes ( Fig. S1 ). Among these, queuosine biosynthesis, lipid IVA biosynthesis, fatty acid elongation, and acetylene degradation were more enriched in post-surgery samples, whereas pathways linked to bacterial cell wall biosynthesis (e.g., UDP-N-acetylmuramoyl-pentapeptide biosynthesis II) were relatively enriched in pre-surgery samples. Comparative abundance testing indicated that queuosine biosynthesis, a pathway implicated in maintaining translational fidelity and epithelial homeostasis, was significantly elevated post-surgery (p < 0.05). In contrast, peptidoglycan biosynthesis pathways, important for structural integrity and bacterial proliferation, showed higher relative abundance both pre- and post-surgery, suggesting their general role in supporting microbiota re-establishment after surgical perturbation. 4. Discussion The gut microbiota plays a central role in CRC, acting not only as a contributor to disease development [68], but also as a therapeutic target [69], as wells as a potential diagnosis and prognosis biomarker [70]. Alternations as a result of colorectal surgery would influence the patient’s wound healing, and susceptibility to complications such as surgical site infection, anastomotic leakage, and recurrence [71]. Therefore, understanding and leveraging these alternations may improve post-surgery outcomes and the overall CRC patient survival. In this study, we observed that post colorectal surgery was associated with a significant reduction in gut microbial richness (Fig. 1 A, p = 0.038), whereas overall community structure showed no statistically significant differences (Fig. 1 B, p = 0.104). This decline in alpha diversity suggests that surgical intervention could have disrupted the ecological balance of the gut microbiota, likely due to preoperative bowel preparation [15], perioperative antibiotics [72], oxygen exposure for obligate microbes, and mucosal injury [68]. Such disruptions have been consistently reported in previous studies [15, 73, 74], where perioperative interventions transiently reduce microbial diversity, which could impair the resilience of the gut ecosystem against pathogens [75]. Although β-diversity differences were not statistically significant, ordination analyses revealed subtle compositional separation between pre- and post-surgery samples. Similar trends have been documented in CRC cohorts, where surgery-induced perturbations reshaped the microbiota without producing large-scale shifts in global community structure [50, 52, 76]. This could be indicating that core microbial taxa may be resilient to surgical stress [77, 78]. Taken altogether, these findings highlight that while the overall microbial landscape remains relatively stable, surgery induces measurable ecological stress that reduces diversity with minimal alteration to core community compositions. Analysis of the top 20 abundant genera in both groups revealed that Prevotella , Alistipes , Fusobacterium , and Collinsella were highly enriched in pre-surgery samples (Fig. 2 ) as compared to post-surgery. Several of these taxa are well-recognized for their association with colorectal carcinogenesis [79]. For instance, Prevotella , an anaerobic Gram-negative bacterium, has been linked to malignant progression of CRC and attenuation of the anticancer efficacy of FOLFOX chemotherapy [80]. Alistipes has been described as a pro-inflammatory taxon with potential pathogenic roles in CRC development [81], while Fusobacterium is a consistently reported CRC-associated genus implicated in tumor initiation and immune modulation [82–85]. Similarly, higher abundances of Collinsella have been observed in CRC patients compared with healthy controls [21, 86]. Together, these patterns underscore that the pre-surgery gut environment is enriched in taxa with pro-carcinogenic potential. To validate these findings, we further examined the four independent external CRC microbiome datasets [50, 87–89], which consistently demonstrated enrichment of pro-CRC bacteria such as Bacteroides , Streptococcus , and Enterococcus in pre-surgery samples (Fig. 3 , p = 0.0062). This concordance strengthens the evidence that CRC patients harbor higher levels of pathogenic and pro-inflammatory taxa before surgical intervention [90, 91]. Conversely, post-surgery samples were enriched with genera considered beneficial to gut and host health, including Parabacteroides and Akkermansia (Fig. 3 ). Studies widely indicate Akkermansia as a next-generation probiotic with protective roles in gastrointestinal disorders, including CRC [92–95]. Similarly, Parabacteroides has been shown to exert anti-inflammatory effects and attenuate tumorigenesis in experimental models [96, 97]. Our analysis of the four independent datasets supported these observations with higher abundances of Phocaeicola , Blautia , Faecalibacterium , and Ruminococcus in post-surgery samples; which are frequently associated with anti-inflammatory activity and maintenance of gut homeostasis [98–103]. These findings suggest that colorectal surgery not only reduces pro-CRC taxa but may also promotes the expansion of potentially beneficial bacteria. Such shifts could have clinical implications, as enrichment of probiotic and anti-inflammatory taxa may contribute to improved immune responses, wound healing, and resilience against postoperative complications, including surgical site infections, anastomotic leakage, and recurrence [104–107]. We applied three supervised machine learning models RF, XGB, and SVM to genus-level abundance data for pre- and post-surgery predictions. Among these, the RF model achieved the highest predictive performance (AUROC = 0.90), outperforming XGB (0.83) and SVM (0.78) (Table 2 ). These findings are consistent with previous studies demonstrating that tree-based classifiers are particularly effective in handling the complex, non-linear relationships characteristic of microbiome data [108, 109]. The predicted probability distributions from the RF model further demonstrated a clear separation between pre- and post-surgery samples (Fig. 4 A), highlighting its robustness in capturing microbial differences associated with surgical intervention. To enhance interpretability of our RF model, we applied SHAP analysis, which highlighted the genera most influential in driving model predictions [110–113]. These taxa were considered potential post-surgery biomarkers (Fig. 4 A). Several of them were commensal bacteria and further analysis of their relative abundance based on 16s rRNA sequencing results indicated that they were indeed enriched in post-surgery samples (Fig. 5 A - D ). For example, Akkermansia emerged as the strongest positive predictor, aligning with its proposed role as a next-generation probiotic with beneficial effects on gut health and CRC outcomes [114, 115]. However, among these top 10 two bacterial features were enriched in pre-surgery samples (Fig. 5 E - F ). This showed that surgical resection could potentially lead to restoration of beneficial taxa in the gut, which would reduce the abundance of opportunistic taxa. Recent studies have reported a similar observation in which the abundance of short chain fatty acids (SCFAs) producers as well as other commensal bacteria were increased post-surgery [52, 116]. These findings not only demonstrate the predictive power of machine learning to identify potential markers but also provide their biological interpretability which could inform postoperative monitoring and, potentially, microbiota-based therapeutic strategies. Pathway analysis revealed distinct functional signatures between pre- and post-surgery samples. The top 10 features identified by RF were predominantly associated with biosynthetic processes. In particular, post-surgery samples ( Fig. S1 ) were enriched for pathways such as queuosine biosynthesis (linked to translational fidelity and epithelial homeostasis) [117], peptidoglycan biosynthesis IV (Enterococcus faecium) (structural integrity) [118], and L-rhamnose degradation I (bacterial carbon source utilization) [119]. These functions may suggest metabolic adaptations that could support the re-establishment of gut microbial communities following surgical disruption [120]. However, this study had some limitations. Firstly, the pre and post-surgery microbial analyses were based on 16S rRNA gene sequencing, which captures only bacterial taxonomic profiles and does not provide comprehensive information on other microbial components or functional pathways. Future studies using shotgun metagenomic sequencing would allow a broader characterization of the gut microbiome and enable stronger correlations with clinical outcomes. Although our model accurately distinguished pre- and post-surgery samples and identified potential post-surgery biomarkers, their functional relevance remains to be experimentally validated. Future independent patient cohorts will be essential to establish their clinical relevance for CRC prognosis. Finally, our relatively small cohort may have limited the statistical power to adequately examine the associations between the two groups. Larger patient cohorts will be essential to strengthen the robustness and validity of future analyses. Conclusion This study characterized gut microbiota changes in CRC patients before and after surgery. Microbial diversity was higher pre-surgery but declined post-surgery, likely reflecting the effects of intensive bowel preparation. Post-surgery samples were enriched with commensal taxa, suggesting a potential restoration of beneficial microbiota. Such microbial shifts may be harnessed to improved gut health, thereby supporting recovery and lowering the risk of postoperative complications. Using explainable AI with SHAP, we identified potential biomarkers, including Akkermansia , which was as key positive contributor in post-surgery. However, the limited cohort size and lack of clinical validation of the identified biomarkers may restrict the generalizability of these findings. Future studies with larger, clinically evaluated cohorts will be essential to confirm the relevance of these biomarkers for them to be translated into tools that can guide CRC prognosis and treatment. Declarations Conflict of interest The authors have no conflicts of interest Data Availability Statement The data presented in this study are available on request from the corresponding author. Ethical approval and consent The study was approved by the Institutional Ethics Committee of Dalin Tzu-Chi Hospital (B11304006), and all participants provided written informed consent. Funding This research was supported by Hualien Tzu-Chi General Hospital, Dalin Tzu-Chi Hospital (DTCRD114(2)-C-04), and Ditmanson Medical Foundation Chiayi Christian Hospital-National Chung Cheng University Joint Research Program (CYCH-CCU-2024-06). Author contributions Mutebi John Kenneth : Writing-Original Draft, Investigation, Conceptualization, Data Curation, Validation, Formal Analysis, and Writing-reviewing and Editing. Chuan-Yin Fang : Conceptualization, Investigation, Validation, Writing-Reviewing and Editing, Resources, and Supervision. Chin-Chia Wu : Conceptualization, Investigation, Validation, Writing-Reviewing and Editing, Resources, and Supervision. Michael W Y Ch an : Formal Analysis, Conceptualization, and Writing-reviewing and Editing Bing-Mu Hsu: Conceptualization, Investigation, Project administration, Funding acquisition, Writing-reviewing and Editing, Resources, and Supervision. References Klimeck L, Heisser T, Hoffmeister M, Brenner H (2023) Colorectal cancer: A health and economic problem. Best practice & research clinical gastroenterology 66:101839 Matsuda T, Fujimoto A, Igarashi Y (2025) Colorectal Cancer: Epidemiology, Risk Factors, and Public Health Strategies. Digestion 106(2):91-99. https://doi.org/10.1159/000543921 Afzaal M, Saeed F, Shah YA, Hussain M, Rabail R, Socol CT, Hassoun A, Pateiro M, Lorenzo JM, Rusu AV (2022) Human gut microbiota in health and disease: Unveiling the relationship. Frontiers in microbiology 13:999001 Andoh A (2016) Physiological Role of Gut Microbiota for Maintaining Human Health. Digestion 93(3):176-181. https://doi.org/10.1159/000444066 Singh G, Chaudhry Z, Boyadzhyan A, Sasaninia K, Rai V (2025) Dysbiosis and colorectal cancer: conducive factors, biological and molecular role, and therapeutic prospectives. Explor Target Antitumor Ther 6:1002329. https://doi.org/10.37349/etat.2025.1002329 Sánchez-Alcoholado L, Ramos-Molina B, Otero A, Laborda-Illanes A, Ordóñez R, Medina JA, Gómez-Millán J, Queipo-Ortuño MI (2020) The Role of the Gut Microbiome in Colorectal Cancer Development and Therapy Response. Cancers (Basel) 12(6). https://doi.org/10.3390/cancers12061406 Zou S, Fang L, Lee MH (2018) Dysbiosis of gut microbiota in promoting the development of colorectal cancer. Gastroenterol Rep (Oxf) 6(1):1-12. https://doi.org/10.1093/gastro/gox031 Oliero M, Hajjar R, Cuisiniere T, Fragoso G, Calvé A, Dagbert F, Loungnarath R, Sebajang H, Schwenter F, Wassef R (2022) Prevalence of pks+ bacteria and enterotoxigenic Bacteroides fragilis in patients with colorectal cancer. Gut Pathogens 14(1):51 Phipps AI, Hill CM, Lin G, Malen RC, Reedy AM, Kahsai O, Ammar H, Curtis K, Ma N, Randolph TW (2025) Fusobacterium nucleatum enrichment in colorectal tumor tissue: Associations with tumor characteristics and survival outcomes. Gastro Hep Advances:100644 Lee JB, Kim K-A, Cho HY, Kim D, Kim WK, Yong D, Lee H, Yoon SS, Han DH, Han YD (2021) Association between Fusobacterium nucleatum and patient prognosis in metastatic colon cancer. Scientific reports 11(1):20263 Paik J-H, Ryu C-G, Hwang D-Y (2023) Risk factors of recurrence in TNM stage I colorectal cancer. Annals of Surgical Treatment and Research 104(5):281-287 Li T, Liu Z, Bai F, Xiao H, Zhou H (2023) Surgical resection for second primary colorectal cancer: a population-based study. Front Med (Lausanne) 10:1167777. https://doi.org/10.3389/fmed.2023.1167777 Ong WL, Morarasu S, Lunca S, Pruna RM, Roata CE, Dimofte GM (2025) Impact of Clostridium difficile Infection Versus Colonization on Postoperative Outcomes After Oncological Colorectal Surgery: An Observational Single-Center Study With Propensity Score Analysis. J Surg Oncol 131(3):489-497. https://doi.org/10.1002/jso.27923 Pak H, Maghsoudi LH, Soltanian A, Gholami F (2020) Surgical complications in colorectal cancer patients. Annals of medicine and surgery 55:13-18 Nalluri-Butz H, Bobel MC, Nugent J, Boatman S, Emanuelson R, Melton-Meaux G, Madoff RD, Jahansouz C, Staley C, Gaertner WB (2022) A pilot study demonstrating the impact of surgical bowel preparation on intestinal microbiota composition following colon and rectal surgery. Scientific reports 12(1):10559 Shogan BD, Belogortseva N, Luong PM, Zaborin A, Lax S, Bethel C, Ward M, Muldoon JP, Singer M, An G, Umanskiy K, Konda V, Shakhsheer B, Luo J, Klabbers R, Hancock LE, Gilbert J, Zaborina O, Alverdy JC (2015) Collagen degradation and MMP9 activation by Enterococcus faecalis contribute to intestinal anastomotic leak. Sci Transl Med 7(286):286ra268. https://doi.org/10.1126/scitranslmed.3010658 Edomskis P, Goudberg MR, Sparreboom CL, Menon AG, Wolthuis AM, D'Hoore A, Lange JF (2021) Matrix metalloproteinase-9 in relation to patients with complications after colorectal surgery: a systematic review. Int J Colorectal Dis 36(1):1-10. https://doi.org/10.1007/s00384-020-03724-6 Shogan BD, Chen J, Duchalais E, Collins D, Chang M, Krull K, Krezalek MA, Larson DW, Walther-Antonio MR, Chia N (2020) Alterations of the rectal microbiome are associated with the development of postoperative ileus in patients undergoing colorectal surgery. Journal of Gastrointestinal Surgery 24(7):1663-1672 Jin Y, Geng R, Liu Y, Liu L, Jin X, Zhao F, Feng J, Wei Y (2020) Prediction of Postoperative Ileus in Patients With Colorectal Cancer by Preoperative Gut Microbiota. Front Oncol 10:526009. https://doi.org/10.3389/fonc.2020.526009 Zhao L, Fang Y, Zhang J, Wei C, Ji H, Zhao J, Wang D, Tang D (2024) Changes in Intestinal Microbiota and Their Relationship With Patient Characteristics in Colorectal Cancer. Clin Med Insights Oncol 18:11795549241307632. https://doi.org/10.1177/11795549241307632 Sheng Q, Du H, Cheng X, Cheng X, Tang Y, Pan L, Wang Q, Lin J (2019) Characteristics of fecal gut microbiota in patients with colorectal cancer at different stages and different sites. Oncology letters 18(5):4834-4844 Oh HH, Joo YE (2020) Novel biomarkers for the diagnosis and prognosis of colorectal cancer. Intest Res 18(2):168-183. https://doi.org/10.5217/ir.2019.00080 Kim CW, Cha JM, Kwak MS (2021) Identification of Potential Biomarkers and Biological Pathways for Poor Clinical Outcome in Mucinous Colorectal Adenocarcinoma. Cancers 13(13):3280 Herlo LF, Salcudean A, Sirli R, Iurciuc S, Herlo A, Nelson-Twakor A, Alexandrescu L, Dumache R (2024) Gut Microbiota Signatures in Colorectal Cancer as a Potential Diagnostic Biomarker in the Future: A Systematic Review. Int J Mol Sci 25(14). https://doi.org/10.3390/ijms25147937 Li S, Konstantinov SR, Smits R, Peppelenbosch MP (2017) Bacterial Biofilms in Colorectal Cancer Initiation and Progression. Trends Mol Med 23(1):18-30. https://doi.org/10.1016/j.molmed.2016.11.004 Qu R, Zhang Y, Ma Y, Zhou X, Sun L, Jiang C, Zhang Z, Fu W (2023) Role of the gut microbiota and its metabolites in tumorigenesis or development of colorectal cancer. Advanced Science 10(23):2205563 Tsigalou C, Paraschaki A, Bragazzi NL, Aftzoglou K, Bezirtzoglou E, Tsakris Z, Vradelis S, Stavropoulou E (2023) Alterations of gut microbiome following gastrointestinal surgical procedures and their potential complications. Front Cell Infect Microbiol 13:1191126. https://doi.org/10.3389/fcimb.2023.1191126 Weaver L, Troester A, Jahansouz C (2024) The Impact of Surgical Bowel Preparation on the Microbiome in Colon and Rectal Surgery. Antibiotics 13(7):580 Elbarmelgi MY, Shafik AA, Abd ElSamee AK, Tamer M (2024) Impact of pre-operative mechanical bowel preparation in preventing post-operative anastomotic leak: A meta-analysis. Asian J Surg. https://doi.org/10.1016/j.asjsur.2024.11.006 Petrou NA, Kontovounisios C (2022) The Use of Mechanical Bowel Preparation and Oral Antibiotic Prophylaxis in Elective Colorectal Surgery: A Call for Change in Practice. Cancers 14(23):5990 Ohigashi S, Sudo K, Kobayashi D, Takahashi T, Nomoto K, Onodera H (2013) Significant changes in the intestinal environment after surgery in patients with colorectal cancer. J Gastrointest Surg 17(9):1657-1664. https://doi.org/10.1007/s11605-013-2270-x Chen Y, Wu N, Yan X, Kang L, Ou G, Zhou Z, Xu C, Feng J, Shi T (2025) Impact of gut microbiota on colorectal anastomotic healing (Review). Mol Clin Oncol 22(6):52. https://doi.org/10.3892/mco.2025.2847 Hajjar R, Gonzalez E, Fragoso G, Oliero M, Alaoui AA, Calvé A, Vennin Rendos H, Djediai S, Cuisiniere T, Laplante P, Gerkins C, Ajayi AS, Diop K, Taleb N, Thérien S, Schampaert F, Alratrout H, Dagbert F, Loungnarath R, Sebajang H, Schwenter F, Wassef R, Ratelle R, Debroux E, Cailhier JF, Routy B, Annabi B, Brereton NJB, Richard C, Santos MM (2023) Gut microbiota influence anastomotic healing in colorectal cancer surgery through modulation of mucosal proinflammatory cytokines. Gut 72(6):1143-1154. https://doi.org/10.1136/gutjnl-2022-328389 Huo R-X, Wang Y-J, Hou S-B, Wang W, Zhang C-Z, Wan X-H (2022) Gut mucosal microbiota profiles linked to colorectal cancer recurrence. World journal of gastroenterology 28(18):1946 Gaines S, Shao C, Hyman N, Alverdy J (2018) Gut microbiome influences on anastomotic leak and recurrence rates following colorectal cancer surgery. Journal of British Surgery 105(2):e131-e141 Pandey K, Umar S (2021) Microbiome in drug resistance to colon cancer. Curr Opin Physiol 23. https://doi.org/10.1016/j.cophys.2021.100472 Gan Y, Yang H, Wang M, Li J (2025) Advances in drug resistance and resistance mechanisms of four colorectal cancer-associated gut microbiota. PeerJ 13:e19535 Chen H-H, Ke T-W, Huang C-W, Jiang J-K, Chen C-C, Hsieh Y-Y, Teng H-W, Lin B-W, Liang Y-H, Su Y-L (2021) Taiwan society of colon and rectal surgeons consensus on mCRC treatment. Frontiers in oncology 11:764912 Yoshida T, Homma S, Ichikawa N, Ohno Y, Miyaoka Y, Matsui H, Imaizumi K, Ishizu H, Funakoshi T, Koike M (2023) Preoperative mechanical bowel preparation using conventional versus hyperosmolar polyethylene glycol-electrolyte lavage solution before laparoscopic resection for colorectal cancer (TLUMP test): a phase III, multicenter randomized controlled non-inferiority trial. Journal of gastroenterology 58(9):883-893 Association WM (2025) World Medical Association Declaration of Helsinki: ethical principles for medical research involving human participants. Jama 333(1):71-74 Kai S, Matsuo Y, Nakagawa S, Kryukov K, Matsukawa S, Tanaka H, Iwai T, Imanishi T, Hirota K (2019) Rapid bacterial identification by direct PCR amplification of 16S rRNA genes using the MinION™ nanopore sequencer. FEBS Open Bio 9(3):548-557. https://doi.org/10.1002/2211-5463.12590 Srinivasan R, Karaoz U, Volegova M, MacKichan J, Kato-Maeda M, Miller S, Nadarajan R, Brodie EL, Lynch SV (2015) Use of 16S rRNA gene for identification of a broad range of clinically relevant bacterial pathogens. PloS one 10(2):e0117617 Buetas E, Jordán-López M, López-Roldán A, D’Auria G, Martínez-Priego L, De Marco G, Carda-Diéguez M, Mira A (2024) Full-length 16S rRNA gene sequencing by PacBio improves taxonomic resolution in human microbiome samples. BMC genomics 25(1):310 Asif A, Koner S, Chen J-S, Hussain A, Huang S-W, Hussain B, Hsu B-M (2024) Uncovering the microbial community structure and physiological profiles of terrestrial mud volcanoes: A comprehensive metagenomic insight towards their trichloroethylene biodegradation potentiality. Environmental Research 258:119457 Gupta S, Mortensen MS, Schjørring S, Trivedi U, Vestergaard G, Stokholm J, Bisgaard H, Krogfelt KA, Sørensen SJ (2019) Amplicon sequencing provides more accurate microbiome information in healthy children compared to culturing. Communications biology 2(1):291 Goldfarb T, Kodali Vamsi K, Pujar S, Brover V, Robbertse B, Farrell Catherine M, Oh D-H, Astashyn A, Ermolaeva O, Haddad D, Hlavina W, Hoffman J, Jackson John D, Joardar Vinita S, Kristensen D, Masterson P, McGarvey Kelly M, McVeigh R, Mozes E, Murphy Michael R, Schafer Susan S, Souvorov A, Spurrier B, Strope Pooja K, Sun H, Vatsan Anjana R, Wallin C, Webb D, Brister J R, Hatcher E, Kimchi A, Klimke W, Marchler-Bauer A, Pruitt Kim D, Thibaud-Nissen F, Murphy Terence D (2024) NCBI RefSeq: reference sequence standards through 25 years of curation and annotation. Nucleic Acids Research 53(D1):D243-D257. https://doi.org/10.1093/nar/gkae1038 Ai D, Pan H, Han R, Li X, Liu G, Xia LC (2019) Using Decision Tree Aggregation with Random Forest Model to Identify Gut Microbes Associated with Colorectal Cancer. Genes 10(2):112 Zheng Y, Fang Z, Xue Y, Zhang J, Zhu J, Gao R, Yao S, Ye Y, Wang S, Lin C, Chen S, Huang H, Hu L, Jiang GN, Qin H, Zhang P, Chen J, Ji H (2020) Specific gut microbiome signature predicts the early-stage lung cancer. Gut Microbes 11(4):1030-1042. https://doi.org/10.1080/19490976.2020.1737487 Wu H, Li Y, Jiang Y, Li X, Wang S, Zhao C, Yang X, Chang B, Yang J, Qiao J (2025) Machine learning prediction of obesity-associated gut microbiota: identifying Bifidobacterium pseudocatenulatum as a potential therapeutic target. Frontiers in Microbiology 15:1488656 Cong J, Zhu H, Liu D, Li T, Zhang C, Zhu J, Lv H, Liu K, Hao C, Tian Z (2018) A pilot study: changes of gut microbiota in post-surgery colorectal cancer patients. Frontiers in microbiology 9:2777 Hernández‐González PI, Barquín J, Ortega‐Ferrete A, Patón V, Ponce‐Alonso M, Romero‐Hernández B, Ocaña J, Caminoa A, Conde‐Moreno E, Galeano J (2023) Anastomotic leak in colorectal cancer surgery: Contribution of gut microbiota and prediction approaches. Colorectal Disease 25(11):2187-2197 Png C-W, Chua Y-K, Law J-H, Zhang Y, Tan K-K (2022) Alterations in co-abundant bacteriome in colorectal cancer and its persistence after surgery: a pilot study. Scientific Reports 12(1):9829 Lee SY, Park H-M, Kim CH, Kim HR (2023) Dysbiosis of gut microbiota during fecal stream diversion in patients with colorectal cancer. Gut Pathogens 15(1):40 Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V (2011) Scikit-learn: Machine learning in Python. the Journal of machine Learning research 12:2825-2830 Abdullah DM, Abdulazeez AM (2021) Machine learning applications based on SVM classification a review. Qubahan Academic Journal 1(2):81-90 Salman HA, Kalakech A, Steiti A (2024) Random forest algorithm overview. Babylonian Journal of Machine Learning 2024:69-79 Chen T, Guestrin C (2016) Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. pp 785-794 Rainio O, Teuho J, Klén R (2024) Evaluation metrics and statistical tests for machine learning. Sci Rep 14(1):6086. https://doi.org/10.1038/s41598-024-56706-x Lundberg SM, Lee S-I (2017) A unified approach to interpreting model predictions. Advances in neural information processing systems 30 Lundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, Katz R, Himmelfarb J, Bansal N, Lee S-I (2020) From local explanations to global understanding with explainable AI for trees. Nature machine intelligence 2(1):56-67 Salih AM, Raisi‐Estabragh Z, Galazzo IB, Radeva P, Petersen SE, Lekadir K, Menegaz G (2025) A perspective on explainable artificial intelligence methods: SHAP and LIME. Advanced Intelligent Systems 7(1):2400304 Chen Y, Ye Y, Liu X, Yin C, Jones CA (2025) Examining the nonlinear and spatial heterogeneity of housing prices in urban Beijing: an application of GeoShapley. Habitat International 162:103439 Deb D, Smith RM (2021) Application of random forest and SHAP tree explainer in exploring spatial (in) justice to aid urban planning. ISPRS International Journal of Geo-Information 10(9):629 Deng D, Zhao L, Song H, Wang H, Cao H, Cui H, Zhou Y, Cui R (2025) Microbiome analysis of gut microbiota in patients with colorectal polyps and healthy individuals. Sci Rep 15(1):7126. https://doi.org/10.1038/s41598-025-91626-4 Yang Z, Xu F, Li H, He Y (2021) Beyond samples: a metric revealing more connections of gut microbiota between individuals. Computational and Structural Biotechnology Journal 19:3930-3937 Ijoma GN, Nkuna R, Mutungwazi A, Rashama C, Matambo TS (2021) Applying PICRUSt and 16S rRNA functional characterisation to predicting co-digestion strategies of various animal manures for biogas production. Scientific reports 11(1):19913 Parks DH, Tyson GW, Hugenholtz P, Beiko RG (2014) STAMP: statistical analysis of taxonomic and functional profiles. Bioinformatics 30(21):3123-3124 Wong CC, Yu J (2023) Gut microbiota in colorectal cancer development and therapy. Nature reviews Clinical oncology 20(7):429-452 Hu Y, Zhou P, Deng K, Zhou Y, Hu K (2024) Targeting the gut microbiota: a new strategy for colorectal cancer treatment. Journal of Translational Medicine 22(1):915 Li G, Zhao D, Ouyang B, Chen Y, Zhao Y (2025) Intestinal microbiota as biomarkers for different colorectal lesions based on colorectal cancer screening participants in community. Frontiers in Microbiology 16:1529858 Agnes A, Puccioni C, D'Ugo D, Gasbarrini A, Biondi A, Persiani R (2021) The gut microbiota and colorectal surgery outcomes: facts or hype? A narrative review. BMC Surg 21(1):83. https://doi.org/10.1186/s12893-021-01087-5 Văcărean-Trandafir IC, Amărandi R-M, Ivanov IC, Dragoș LM, Mențel M, Iacob Ş, Muşină A-M, Bărgăoanu E-R, Roată CE, Morărașu Ș (2025) Impact of antibiotic prophylaxis on gut microbiota in colorectal surgery: insights from an Eastern European stewardship study. Frontiers in Cellular and Infection Microbiology 14:1468645 Weaver L, Troester A, Jahansouz C (2024) The Impact of Surgical Bowel Preparation on the Microbiome in Colon and Rectal Surgery. Antibiotics (Basel) 13(7). https://doi.org/10.3390/antibiotics13070580 Kohn J, Troester A, Ziegert Z, Frebault J, Boatman S, Martell M, Nalluri-Butz H, Bobel MC, Goffredo P, Johnson AJ, Jahansouz C, Staley C, Gaertner WB (2025) The Role of Surgical and Perioperative Factors in Shaping Gut Microbiome Recovery After Colorectal Surgery. Antibiotics 14(9):881 Koliarakis I, Athanasakis E, Sgantzos M, Mariolis-Sapsakos T, Xynos E, Chrysos E, Souglakos J, Tsiaoussis J (2020) Intestinal Microbiota in Colorectal Cancer Surgery. Cancers (Basel) 12(10). https://doi.org/10.3390/cancers12103011 Deng X, Li Z, Li G, Li B, Jin X, Lyu G (2018) Comparison of microbiota in patients treated by surgery or chemotherapy by 16S rRNA sequencing reveals potential biomarkers for colorectal cancer therapy. Frontiers in Microbiology 9:1607 Safarchi A, Al-Qadami G, Tran CD, Conlon M (2025) Understanding dysbiosis and resilience in the human gut microbiome: biomarkers, interventions, and challenges. Front Microbiol 16:1559521. https://doi.org/10.3389/fmicb.2025.1559521 Stavrou G, Kotzampassi K (2017) Gut microbiome, surgical complications and probiotics. Ann Gastroenterol 30(1):45-53. https://doi.org/10.20524/aog.2016.0086 Zhuang Y-P, Zhou H-L, Chen H-B, Zheng M-Y, Liang Y-W, Gu Y-T, Li W-T, Qiu W-L, Zhou H-G (2023) Gut microbiota interactions with antitumor immunity in colorectal cancer: From understanding to application. Hou XY, Zhang P, Du HZ, Gao YQ, Sun RQ, Qin SY, Tian Y, Li J, Zhang YX, Chu WH, Zhang ZJ, Xu FG (2021) Prevotella contributes to individual response of FOLFOX in colon cancer. Clin Transl Med 11(9):e512. https://doi.org/10.1002/ctm2.512 Fu J, Li G, Li X, Song S, Cheng L, Rui B, Jiang L (2024) Gut commensal Alistipes as a potential pathogenic factor in colorectal cancer. Discov Oncol 15(1):473. https://doi.org/10.1007/s12672-024-01393-3 Ou S, Wang H, Tao Y, Luo K, Ye J, Ran S, Guan Z, Wang Y, Hu H, Huang R (2022) Fusobacterium nucleatum and colorectal cancer: From phenomenon to mechanism. Front Cell Infect Microbiol 12:1020583. https://doi.org/10.3389/fcimb.2022.1020583 Dadgar-Zankbar L, Elahi Z, Shariati A, Khaledi A, Razavi S, Khoshbayan A (2024) Exploring the role of Fusobacterium nucleatum in colorectal cancer: Implications for tumor proliferation and chemoresistance. Cell Communication and Signaling 22(1):547 Wang N, Fang J-Y (2023) Fusobacterium nucleatum, a key pathogenic factor and microbial biomarker for colorectal cancer. Trends in Microbiology 31(2):159-172 Wu J, Li Q, Fu X (2019) Fusobacterium nucleatum contributes to the carcinogenesis of colorectal cancer by inducing inflammation and suppressing host immunity. Translational oncology 12(6):846-851 Yuan D, Tao Y, Wang H, Wang J, Cao Y, Cao W, Pan S, Yu Z (2022) A comprehensive analysis of the microbiota composition and host driver gene mutations in colorectal cancer. Investigational New Drugs 40(5):884-894 Hernández-González PI, Barquín J, Ortega-Ferrete A, Patón V, Ponce-Alonso M, Romero-Hernández B, Ocaña J, Caminoa A, Conde-Moreno E, Galeano J, Campo RD, García-Pérez JC (2023) Anastomotic leak in colorectal cancer surgery: Contribution of gut microbiota and prediction approaches. Colorectal Dis 25(11):2187-2197. https://doi.org/10.1111/codi.16733 Png CW, Chua YK, Law JH, Zhang Y, Tan KK (2022) Alterations in co-abundant bacteriome in colorectal cancer and its persistence after surgery: a pilot study. Sci Rep 12(1):9829. https://doi.org/10.1038/s41598-022-14203-z Lee SY, Park HM, Kim CH, Kim HR (2023) Dysbiosis of gut microbiota during fecal stream diversion in patients with colorectal cancer. Gut Pathog 15(1):40. https://doi.org/10.1186/s13099-023-00566-9 Ilozumba MN, Gomez MF, Lin T, Himbert C, Round JL, Zac Stephens W, Warby CA, Hardikar S, Li CI, Figueiredo JC (2025) Pre-surgery gut microbial diversity and abundance are associated with post-surgery onset of cachexia in colorectal cancer patients: the ColoCare Study. Cancer Causes & Control:1-18 Cintoni M, Palombaro M, Zoli E, D’Agostino G, Pulcini G, Leonardi E, Raoul P, Rinninella E, De Maio F, Capristo E, Gasbarrini A, Mele MC (2025) The Interplay Between the Gut Microbiota and Colorectal Cancer: A Review of the Literature. Microorganisms 13(6):1410 Jian H, Liu Y, Wang X, Dong X, Zou X (2023) Akkermansia muciniphila as a Next-Generation Probiotic in Modulating Human Metabolic Homeostasis and Disease Progression: A Role Mediated by Gut-Liver-Brain Axes? Int J Mol Sci 24(4). https://doi.org/10.3390/ijms24043900 Zhai Q, Feng S, Arjan N, Chen W (2019) A next generation probiotic, Akkermansia muciniphila. Crit Rev Food Sci Nutr 59(19):3227-3236. https://doi.org/10.1080/10408398.2018.1517725 Jan T, Negi R, Sharma B, Kumar S, Singh S, Rai AK, Shreaz S, Rustagi S, Chaudhary N, Kaur T (2024) Next generation probiotics for human health: An emerging perspective. Heliyon 10(16) Lalowski P, Zielińska D (2024) The Most Promising Next-Generation Probiotic Candidates—Impact on Human Health and Potential Application in Food Technology. Fermentation 10(9):444 Koh GY, Kane AV, Wu X, Crott JW (2020) Parabacteroides distasonis attenuates tumorigenesis, modulates inflammatory markers and promotes intestinal barrier integrity in azoxymethane-treated A/J mice. Carcinogenesis 41(7):909-917. https://doi.org/10.1093/carcin/bgaa018 Yu SY, Xie YH, Qiu YW, Chen YX, Fang JY (2019) Moderate alteration to gut microbiota brought by colorectal adenoma resection. Journal of Gastroenterology and Hepatology 34(10):1758-1765 Choi J, Choi YR, Jeong MK, Song HH, Yu JS, Song SH, Park JH, Kim MJ, Park H, Ham YL, Han SH, Kim DJ, Lee DY, Suk KT (2025) Phocaeicola dorei ameliorates progression of steatotic liver disease by regulating bile acid, lipid, inflammation and proliferation. Gut Microbes 17(1):2539448. https://doi.org/10.1080/19490976.2025.2539448 Xiang J, Chai N, Li L, Hao X, Linghu E (2024) Alterations of Gut Microbiome in Patients with Colorectal Advanced Adenoma by Metagenomic Analyses. The Turkish Journal of Gastroenterology 35(11):859 Dikeocha IJ, Al-Kabsi AM, Chiu H-T, Alshawsh MA (2022) Faecalibacterium prausnitzii Ameliorates Colorectal Tumorigenesis and Suppresses Proliferation of HCT116 Colorectal Cancer Cells. Biomedicines 10(5):1128 Xu F, Li Q, Wang S, Dong M, Xiao G, Bai J, Wang J, Sun X (2023) The efficacy of prevention for colon cancer based on the microbiota therapy and the antitumor mechanisms with intervention of dietary Lactobacillus. Microbiology Spectrum 11(5):e00189-00123 Zhang X, Yu D, Wu D, Gao X, Shao F, Zhao M, Wang J, Ma J, Wang W, Qin X (2023) Tissue-resident Lachnospiraceae family bacteria protect against colorectal carcinogenesis by promoting tumor immune surveillance. Cell host & microbe 31(3):418-432. e418 Liu X, Mao B, Gu J, Wu J, Cui S, Wang G, Zhao J, Zhang H, Chen W (2021) Blautia-a new functional genus with potential probiotic properties? Gut Microbes 13(1):1-21. https://doi.org/10.1080/19490976.2021.1875796 Togo C, Zidorio AP, Gonçalves V, Botelho P, de Carvalho K, Dutra E (2021) Does Probiotic Consumption Enhance Wound Healing? A Systematic Review. Nutrients 14(1). https://doi.org/10.3390/nu14010111 Siddharthan R, Chapek M, Warren M, Martindale R (2018) Probiotics in prevention of surgical site infections. Surgical Infections 19(8):781-784 Liu PC, Yan YK, Ma YJ, Wang XW, Geng J, Wang MC, Wei FX, Zhang YW, Xu XD, Zhang YC (2017) Probiotics Reduce Postoperative Infections in Patients Undergoing Colorectal Surgery: A Systematic Review and Meta-Analysis. Gastroenterol Res Pract 2017:6029075. https://doi.org/10.1155/2017/6029075 Liu PC, Yan YK, Ma YJ, Wang XW, Geng J, Wang MC, Wei FX, Zhang YW, Xu XD, Zhang YC (2017) Probiotics reduce postoperative infections in patients undergoing colorectal surgery: a systematic review and meta‐analysis. Gastroenterology research and practice 2017(1):6029075 Wang XW, Liu YY (2020) Comparative study of classifiers for human microbiome data. Med Microecol 4. https://doi.org/10.1016/j.medmic.2020.100013 Teixeira M, Silva F, Ferreira RM, Pereira T, Figueiredo C, Oliveira HP (2024) A review of machine learning methods for cancer characterization from microbiome data. NPJ Precision Oncology 8(1):123 Zhou Y, Han W, Feng Y, Wang Y, Liu X, Sun T, Xu J (2025) Revealing gut microbiota biomarkers associated with melanoma immunotherapy response and key bacteria-fungi interaction relationships: evidence from metagenomics, machine learning, and SHAP methodology. Front Immunol 16:1539653. https://doi.org/10.3389/fimmu.2025.1539653 Ma J, Fang Y, Li S, Zeng L, Chen S, Li Z, Ji G, Yang X, Wu W (2025) Interpretable machine learning algorithms reveal gut microbiome features associated with atopic dermatitis. Frontiers in Immunology 16:1528046 Novielli P, Romano D, Magarelli M, Bitonto PD, Diacono D, Chiatante A, Lopalco G, Sabella D, Venerito V, Filannino P (2024) Explainable artificial intelligence for microbiome data analysis in colorectal cancer biomarker identification. Frontiers in Microbiology 15:1348974 Rynazal R, Fujisawa K, Shiroma H, Salim F, Mizutani S, Shiba S, Yachida S, Yamada T (2023) Leveraging explainable AI for gut microbiome-based colorectal cancer classification. Genome biology 24(1):21 Tabatabaei SAS, Ghadim HY, Alaei S, Abdolvand F, Mazaheri H, Shamsi F, SarveAhrabi Y, Behrouzi A (2025) The association between the health of the intestines and the human body with Akkermansia muciniphila. The Microbe:100352 Faghfuri E, Gholizadeh P (2024) The role of Akkermansia muciniphila in colorectal cancer: A double-edged sword of treatment or disease progression? Biomedicine & Pharmacotherapy 173:116416 Cronin P, Murphy CL, Barrett M, Ghosh TS, Pellanda P, O’Connor EM, Zulquernain SA, Kileen S, McCourt M, Andrews E (2022) Colorectal microbiota after removal of colorectal cancer. NAR cancer 4(2):zcac011 Ehrenhofer-Murray AE (2025) Queuine: A bacterial nucleobase shaping translation in eukaryotes. Journal of Molecular Biology:168985 Ramos Y, Sansone S, Morales DK (2021) Sugarcoating it: Enterococcal polysaccharides as key modulators of host-pathogen interactions. PLoS Pathog 17(9):e1009822. https://doi.org/10.1371/journal.ppat.1009822 Xie S, Ma J, Lu Z (2024) Bacteroides thetaiotaomicron enhances oxidative stress tolerance through rhamnose-dependent mechanisms. Front Microbiol 15:1505218. https://doi.org/10.3389/fmicb.2024.1505218 Li D, Wei R, Zhang X, Gong S, Wan M, Wang F, Li J, Chen M, Liu R, Wan Y (2024) Gut commensal metabolite rhamnose promotes macrophages phagocytosis by activating SLC12A4 and protects against sepsis in mice. Acta Pharmaceutica Sinica B 14(7):3068-3085 Additional Declarations No competing interests reported. Supplementary Files SuplementaryData.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 15 Mar, 2026 Reviews received at journal 21 Dec, 2025 Reviewers agreed at journal 15 Dec, 2025 Reviewers agreed at journal 12 Nov, 2025 Reviewers invited by journal 28 Oct, 2025 Editor assigned by journal 18 Oct, 2025 Submission checks completed at journal 16 Oct, 2025 First submitted to journal 15 Oct, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7866815","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":539020032,"identity":"82fae13c-adbd-4d2c-a149-cfedc6d8aae2","order_by":0,"name":"Mutebi John Kenneth","email":"","orcid":"","institution":"National Chung Cheng University","correspondingAuthor":false,"prefix":"","firstName":"Mutebi","middleName":"John","lastName":"Kenneth","suffix":""},{"id":539020033,"identity":"8770fd48-ca76-46e1-978f-05c537f29084","order_by":1,"name":"Chin-Chia Wu","email":"","orcid":"","institution":"Dalin Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation","correspondingAuthor":false,"prefix":"","firstName":"Chin-Chia","middleName":"","lastName":"Wu","suffix":""},{"id":539020034,"identity":"69fb71b6-a590-4730-aae2-8f4f6eedbe2c","order_by":2,"name":"Chuan-Yin Fang","email":"","orcid":"","institution":"Ditmanson Medical Foundation Chia-Yi Christian Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chuan-Yin","middleName":"","lastName":"Fang","suffix":""},{"id":539020035,"identity":"46fa2fa2-29a7-4127-aeb4-9479dfe38270","order_by":3,"name":"Michael W Y Chan","email":"","orcid":"","institution":"Department of Biomedical Sciences, National Chung Cheng University","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"W Y","lastName":"Chan","suffix":""},{"id":539020036,"identity":"dd1477ec-c924-4a2d-839a-b2064f393d3f","order_by":4,"name":"Bing-Mu Hsu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAsklEQVRIiWNgGAWjYBACAwYeIFkB4UiQoOUMXIsBkVoY20jRYi529vBn3nl28gYHmA/e5mH4k9hASIvl7Lw0ad5tyYYbDrAlW/MwGBDWYnA7x4yZd9uBBIMDPGbSQC25xGgx/sw7B6SF/xvRWgykeRvAtrARp8Vydo6Z5JxjyYYzD7MZW84xMK4nqMVcOsf4w5saO3m+480Pb7ypkDMmpAMJMIPdSYKGUTAKRsEoGAW4AQAnJTUQ4RMIvAAAAABJRU5ErkJggg==","orcid":"","institution":"National Chung Cheng University","correspondingAuthor":true,"prefix":"","firstName":"Bing-Mu","middleName":"","lastName":"Hsu","suffix":""}],"badges":[],"createdAt":"2025-10-15 10:08:46","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7866815/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7866815/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95502215,"identity":"aedc7564-48c4-44ff-840f-8d23e0d8f89b","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"tif","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6065862,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/bad1787180d27cf1a79111c7.tif"},{"id":95502236,"identity":"8176ca3c-489d-474c-b876-4de043872b84","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":152347,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptCRCML.docx","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/ae1014e66dcae555e9a4d8c0.docx"},{"id":95502224,"identity":"f19bd9f1-32c6-41c5-b283-c8055038246d","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11784034,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/caa1ae0c5f9b65803064a9b9.tif"},{"id":95502219,"identity":"8f22a9c1-c2a9-483f-b168-659c7a5fec0e","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14196,"visible":true,"origin":"","legend":"","description":"","filename":"TableOne.docx","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/ae59a97dac1c16ab5bc9c71a.docx"},{"id":95502220,"identity":"5af6f096-51dc-4f7b-b413-ab32ee86571a","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13811,"visible":true,"origin":"","legend":"","description":"","filename":"TableTwo.docx","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/74e40e172897d5afc4b1a319.docx"},{"id":95529164,"identity":"fc4d68e1-aaf5-42b4-86f7-2e7f235d65fa","added_by":"auto","created_at":"2025-11-10 10:16:51","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4591638,"visible":true,"origin":"","legend":"","description":"","filename":"Fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/e44882d0dbd1104b2da5de1a.tif"},{"id":95502225,"identity":"1e9cef56-24e1-4c47-aee6-bf171d9bf65e","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"tif","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1628836,"visible":true,"origin":"","legend":"","description":"","filename":"Fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/e221e9ac4f9e286d8dad8ee2.tif"},{"id":95529294,"identity":"e941dfe3-2181-4764-be02-d11251521b8b","added_by":"auto","created_at":"2025-11-10 10:16:57","extension":"tif","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8155008,"visible":true,"origin":"","legend":"","description":"","filename":"Fig5.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/4c21c4b71f19b53657c6b39b.tif"},{"id":95502231,"identity":"7df156bc-aaff-4cf1-84f4-9c88a54fd462","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"json","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6691,"visible":true,"origin":"","legend":"","description":"","filename":"e87943cd77064caabe1b184dbdb6d675.json","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/691fadb9b39a2de60439b80f.json"},{"id":95502234,"identity":"aa28ead1-f110-467b-871c-c053cb6b3432","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":150879,"visible":true,"origin":"","legend":"","description":"","filename":"SuplementaryData.docx","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/99a5ab3aabd2da35da28fac2.docx"},{"id":95527980,"identity":"3bc7168b-77c2-4ff3-857a-f6ff11c629a3","added_by":"auto","created_at":"2025-11-10 10:15:19","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":254786,"visible":true,"origin":"","legend":"","description":"","filename":"e87943cd77064caabe1b184dbdb6d6751enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/7255632416d08bc5aca6e8d3.xml"},{"id":95528706,"identity":"756f8789-5053-420b-a9d0-8d90ce108e64","added_by":"auto","created_at":"2025-11-10 10:16:24","extension":"tif","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6065862,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/5517f20a4e4b6f235536848e.tif"},{"id":95502243,"identity":"208e9a64-aea8-4d92-83f9-3f7c9740f051","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"tif","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11784034,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/2db8ba07463d4ccd1ae995a3.tif"},{"id":95528889,"identity":"c74f56d0-888f-461a-a988-73a4c47e574c","added_by":"auto","created_at":"2025-11-10 10:16:34","extension":"tif","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4591638,"visible":true,"origin":"","legend":"","description":"","filename":"Fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/753b0b12c00a41b2ec267e74.tif"},{"id":95502239,"identity":"83c50760-0971-45c1-808b-5f7d508a16c0","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"tif","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1628836,"visible":true,"origin":"","legend":"","description":"","filename":"Fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/4196b107f01a8855ba10fed0.tif"},{"id":95528816,"identity":"a29b32b9-7be4-4500-86ce-0c3889a0b07c","added_by":"auto","created_at":"2025-11-10 10:16:31","extension":"tif","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8155008,"visible":true,"origin":"","legend":"","description":"","filename":"Fig5.tif","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/1f7cb26ff5993598b59abec3.tif"},{"id":95528651,"identity":"1d553c11-8bef-4ce4-9c5f-e6d75dc6f2e6","added_by":"auto","created_at":"2025-11-10 10:16:22","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":395326,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig1.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/e81c6025cbd60a385fdbf7df.png"},{"id":95502242,"identity":"8de09420-704d-4d6b-9c4e-4b1689088cb7","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":731990,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig2.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/809f233c538c576d21957e45.png"},{"id":95502229,"identity":"855c5753-9d12-4634-b63e-4f41d3a4513d","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":603879,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig3.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/5d76bc964d11cd03e19b4630.png"},{"id":95502246,"identity":"0448f5ed-fc1f-4ac0-be95-aa585c53e860","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":150130,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig4.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/a84376080e870225987d713f.png"},{"id":95529150,"identity":"da6f8486-ab21-4d1b-bda3-7d57f3ed0718","added_by":"auto","created_at":"2025-11-10 10:16:50","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":617592,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig5.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/431d90bd52f08cc3140de12f.png"},{"id":95502244,"identity":"9f5b5061-f91c-401b-9064-dfa24f7f620f","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"xml","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":236728,"visible":true,"origin":"","legend":"","description":"","filename":"e87943cd77064caabe1b184dbdb6d6751structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/d15de6efa63d0b6b7f6f3be0.xml"},{"id":95502247,"identity":"86d310e5-a604-4256-beb4-2df75e067101","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"html","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":267259,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/e8ef7c59ab515c982e1c48ba.html"},{"id":95502235,"identity":"9065283b-86b4-42aa-8fab-2e92ca416e27","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":9983905,"visible":true,"origin":"","legend":"\u003cp\u003eImpact of surgical resection in colorectal cancer on gut microbiota composition: (A) Alpha diversity (Chao1 index) and (B) beta diversity (unweighted UniFrac distances) illustrating differences in bacterial community structure between pre- and post-surgery samples\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/3730aed7b8e6ed5c4ee88802.png"},{"id":95502233,"identity":"f2e5b986-a818-41cf-b2c2-70c5b7f424af","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":29097134,"visible":true,"origin":"","legend":"\u003cp\u003eVariation in the relative abundance and composition of the top 20 gut bacterial taxa in pre- and post-surgery fecal samples of CRC patients in our study cohort\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/0884417156e659bf5e86b952.png"},{"id":95502237,"identity":"3a6b59ef-ee5b-4a7e-9862-a1be47e78ea1","added_by":"auto","created_at":"2025-11-10 05:35:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":11608759,"visible":true,"origin":"","legend":"\u003cp\u003eVariation in the relative abundance and composition of the top 20 gut bacterial taxa in pre- and post-surgery across the four external reference datasets\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/bc103262700460b6661f60c3.png"},{"id":95502217,"identity":"b07ec48a-fbaa-44a3-a68f-594fb6b5a4e0","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1946847,"visible":true,"origin":"","legend":"\u003cp\u003eGut microbiota–based classification of pre- and post-surgery CRC patients by Random forest. (A) Predicted probability distributions for pre- and post-surgery samples. (B) Top 10 post-surgery–enriched genera identified by average absolute SHAP values as potential microbial biomarkers.\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/4a0149a787d41d3fefb5d687.png"},{"id":95502230,"identity":"33155b88-98c9-41f0-bf59-01587a3d14b0","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":15270064,"visible":true,"origin":"","legend":"\u003cp\u003eComparative analysis of the relative abundance of selected bacterial genera ranked among the top 10 features by average absolute SHAP values between pre- and post-surgery groups.\u003c/p\u003e","description":"","filename":"Fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/0517941fb38ccb8a2350508d.png"},{"id":95502214,"identity":"30a9f8f1-a486-4c1e-9840-fc5197f0c36a","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":940893,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/9481e0cf-191f-4ea6-961f-bd812ddaa26f.pdf"},{"id":95502227,"identity":"c9a01370-2f66-425a-a66f-0016e3aef4ca","added_by":"auto","created_at":"2025-11-10 05:35:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":150879,"visible":true,"origin":"","legend":"","description":"","filename":"SuplementaryData.docx","url":"https://assets-eu.researchsquare.com/files/rs-7866815/v1/2e1c97f174d5570d7f581ef3.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Deciphering post-surgery gut microbial dynamics in colorectal cancer through multi- cohort machine learning","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eColorectal cancer (CRC) is the second leading cause of cancer mortality worldwide [1], and is influenced by various factors, including heredity, lifestyle, diet, and the host\u0026rsquo;s gastrointestinal microbiota [2]. The human gastrointestinal tract is inhabited by a vast assemblage of microorganisms, including protozoa, archaea, fungi, viruses, and bacteria [3]. This gut microbiota plays a significant role in maintaining the host\u0026rsquo;s health by regulating and modulating intestinal homeostasis and local immune response [4]. Reduced microbial diversity, often associated with gut dysbiosis, has been linked to an increased risk of developing CRC [5]. Therefore, maintaining a delicate balance of the gut microbiota is not only essential for promoting overall health but also for preventing the risk of developing CRC.\u003c/p\u003e\u003cp\u003eDysbiosis promotes opportunistic bacterial overgrowth, generate carcinogenic bacterial metabolites, trigger chronic inflammation, and impair mucosal barrier integrity, thereby contributing to CRC carcinogenesis [6, 7]. Additionally, certain bacteria, such as Enterotoxigenic \u003cem\u003eBacteroides fragilis\u003c/em\u003e (ETBF), Polyketide synthase positive (pks+) \u003cem\u003eEscherichia coli\u003c/em\u003e, \u003cem\u003eFusobacterium nucleatum\u003c/em\u003e, and \u003cem\u003eEnterococcus faecalis\u003c/em\u003e, are reportedly more abundant in CRC patients than in healthy controls, and have been associated with CRC development and progression [8\u0026ndash;10]. This evidence supports recent efforts by scientists to harness gut microbiota as biomarkers for the diagnosis and treatment of CRC.\u003c/p\u003e\u003cp\u003eSurgical resection is the primary treatment for CRC, particularly in patients with localized or locally advanced disease [11, 12]. However, colorectal surgery is associated with various postoperative complications, including \u003cem\u003eClostridioides difficile\u003c/em\u003e infection (CDI) [13], anastomotic leak, surgical site infections (SSIs), and postoperative ileus (POI), which increases patient morbidity and mortality [14]. A recent study reported that colorectal surgery has among the highest rates of SSIs of about 23% and a weighted mean of 11% [15], despite optimal and appropriate surgical techniques and sterile conditions.\u003c/p\u003e\u003cp\u003eGut microbiota has been proposed to have a causative role in these complications among CRC patients. For example, collagen degradation and Matrix Metalloproteinase-9 (MMP-9) activation by \u003cem\u003eEnterococcus faecalis\u003c/em\u003e have been linked to anastomotic leak [16, 17], while increased abundance of inflammation-promoting bacteria has been linked to POI [18, 19]. This implies that gut microbiota can also be harnessed for targeted prevention and treatment of post-surgery complications in CRC patients.\u003c/p\u003e\u003cp\u003eRecent studies have described distinct alterations in the gut microbiota of CRC patients compared with healthy individuals [20, 21]. As a result, microbial signatures with potential biomarker value for diagnosis, prognosis, and therapeutic stratification have been identified [22\u0026ndash;24]. However, little is known about how CRC treatments, especially colorectal resection, can influence gut microbiota. Because gut microbiota can initiate and promote CRC tumorigenesis [25, 26], surgical removal of the tumor is expected to affect the associated microbiota [27]. However, it remains unclear whether surgical resection restores beneficial microbiota and normal gut homeostasis.\u003c/p\u003e\u003cp\u003ePatients undergoing colorectal surgery often receive bowel preparation, including oral antibiotics, polyethylene glycol, and perioperative intravenous antibiotics, which can affect their gut microbial composition [28\u0026ndash;30]. Additionally, surgery disrupts the gut\u0026rsquo;s anaerobic ecosystem, reducing obligate anaerobes and allowing facultative anaerobes to persist, which may contribute to postoperative dysbiosis [31]. Since microbial composition influences intestinal and anastomotic tissue healing [32, 33], recurrence [34, 35], and chemoresistance [36, 37], characterizing microbiota dynamics after surgery is critical. Therefore, this study was conducted to analyze fecal samples of CRC patients before and after colorectal resection, aiming to determine microbial changes associated with post-surgery in CRC patients. We hypothesized that postoperative fecal samples would contain potential microbial biomarkers that can be harnessed to improve the overall colorectal surgery outcomes. We applied machine learning (ML), eXplainable Artificial Intelligence (XAI), and SHapley Additive exPlanations (SHAP) to identify microbial features with potential relevance for postoperative CRC management. The findings of this study could foster a new era of personalized and precision medicine that is based on microbial modulation for CRC management.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Study design and sample collection\u003c/h2\u003e\u003cp\u003eThis prospective study included 15 consecutive patients with CRC scheduled for surgery at Buddhist Dalin Tzu Chi General Hospital between June 2023 and December 2024. Eligible participants were aged 42\u0026ndash;85 years (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and were able to provide written informed consent. CRC was defined as primary malignant epithelial colorectal tumors or adenomas larger than 10 mm, according to the Taiwan Society of Colon and Rectal Surgeons (TSCRS) [38]. Diagnosis was confirmed by colonoscopy to ascertain the CRC status of every participant.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDemographic overview of our study population.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHealthy Controls\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCRC Patients\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eNumber\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e15\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eAge, year, mean (IQR)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e52.3 (25.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e63.0 (8.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eSex (Female/Male), n\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e5\\5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7\\8\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eBMI, mean (IQR)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e21.9 (6.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e23.8 (3.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eTumor Location\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRectum\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eParticipants with distant metastases, coexisting bowel disorders, prior emergency surgery, or receipt of chemotherapy, radiotherapy, targeted therapy, or antibiotics/probiotics within three months before enrollment were excluded. Preoperative stool samples were collected 7 days before colonoscopy (categorized as Before Surgery) and postoperative samples were collected at least three months later (categorized as After Surgery).\u003c/p\u003e\u003cp\u003eAll patients underwent standard mechanical bowel preparation with polyethylene glycol the day before surgery and received prophylactic antibiotics immediately before surgery [39]. Surgical resections were performed according to oncological principles of TSCRS. The study protocol was approved by the Institutional Review Board (IRB) of Buddhist Dalin Tzu Chi General Hospital and conducted in accordance with the Declaration of Helsinki [40].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 DNA extraction, sequencing and data processing\u003c/h2\u003e\u003cp\u003eMicrobial DNA was extracted from stool samples using the QIAamp Fast DNA Stool Mini Kit (QIAGEN, Germany) according to the manufacturer\u0026rsquo;s instructions. The concentration and purity of the extracted DNA were assessed with a NanoDrop\u0026reg; spectrophotometer (Thermo Fisher Scientific, USA), and the integrity was confirmed by PCR amplification and agarose gel electrophoresis of the 16S rRNA gene (~\u0026thinsp;1500 bp) [41].\u003c/p\u003e\u003cp\u003eThe full-length 16S rRNA gene (V1\u0026ndash;V9 regions) was amplified with universal primers 27F (5\u0026rsquo;-AGRGTTTGATYMTGGCTCAG-3\u0026rsquo;) and 1492R (5\u0026rsquo;-GGYTACCTTGTTACGACTT-3\u0026rsquo;) [42]. Amplicons were sequenced on the PacBio Sequel platform (Pacific Biosciences, USA) following the manufacturer\u0026rsquo;s protocol [43]. Circular consensus sequence (CCS) reads were generated from raw PacBio data and analyzed using QIIME2 (v2025.7) as previously reported [44]. Low-quality and chimeric sequences were trimmed with DADA2 (q2-DADA2). High-quality reads were clustered into amplicon sequence variants (ASVs) at 98% similarity [45], and taxonomy was assigned against the National Center for Biotechnology Information (NCBI) reference database [46]. Taxonomic profiles were collapsed to the genus level and normalized with respect to the cumulative microbial count for each subject.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Machine learning modeling and identification of post-surgery\u0026ndash;associated microbial markers.\u003c/h2\u003e\u003cp\u003eWe applied machine learning (ML) models to identify potential gut microbial biomarkers for colorectal post-surgery. Specifically, we trained and evaluated these models using random forest (RF), support vector machine (SVM) and XGBoost algorithms. These algorithms have been previously implemented to identify microbial biomarkers in studies involving genomic datasets [47\u0026ndash;49]. To enlarge our dataset, four external sequence read datasets (DA1\u0026ndash;DA4; \u003cb\u003eTable \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e) were retrieved from the NCBI Sequence Read Archive (SRA; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/sra\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/sra\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), for independent studies from China [50], Spain [51], Singapore [52], and South Korea [53], respectively. The inclusion criteria were the availability of pre- and post-surgery microbiome data, associated metadata, sufficient sequencing depth, and high-quality sequence reads.\u003c/p\u003e\u003cp\u003eDatasets DA1 and DA3 included both CRC patients and healthy controls; for consistency, only CRC patients were analyzed. Similarly, DA2 comprised 111 participants with or without colorectal anastomotic leakage (CAL). However, non-CAL patients with paired pre- and post-surgery microbiome data were analyzed. Furthermore, DA4 consisted of CRC patients with or without ileostomy and only patients from the non-ileostomy control group with paired pre- and post-surgery samples were analyzed. In all studies, the V3\u0026ndash;V4 region of the 16S rRNA gene was amplified with universal bacterial primers and DNA amplicons were paired-end sequenced on the Illumina platform.\u003c/p\u003e\u003cp\u003eThese raw paired-end sequences were processed in QIIME2 (v2025.7) pipeline as described in Section \u003cspan refid=\"Sec4\" class=\"InternalRef\"\u003e2.2\u003c/span\u003e above. The resulting ASVs were taxonomically assigned against the NCBI reference database [46], and collapsed up to the genus level. These genera were normalized in accordance with the cumulative microbial count for each subject and standardized by z-score transformation. Sparse features (\u0026gt;\u0026thinsp;90% zero values) were excluded to reduce noise and then, normalized genus level data was fed into machine learning models for analysis.\u003c/p\u003e\u003cp\u003eThe analysis was implemented in Python (v3.11.4) using SciKit-learn library (v1.7.0; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://scikit-learn.org/stable/\u003c/span\u003e\u003cspan address=\"https://scikit-learn.org/stable/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) in Visual Studio Code (v1.100) [54]. The dataset was split randomly into training (70%) and validation (30%) sets. Model hyperparameters were optimized using grid search with 20 repetitions of 5-fold stratified cross-validation. For SVM, the radial basis function (rbf) kernel was tuned for C (1\u0026ndash;50) and gamma (0.001\u0026ndash;0.1) [55]. For RF, the maximum tree depth (None, 3, 5) and number of estimators (50, 100, 150, 200, 250) were tested [56]. For XGBoost, maximum tree depth (None, 3, 5), colsample_bytree (0.1\u0026ndash;0.9), and number of estimators (50, 100, 150, 200, 250) were optimized [57]. The primary evaluation metric was the area under the receiver operating characteristic curve (AUC); however, additional metrics including accuracy, precision, recall, and F1-score were assessed [58]. After comparing these metrics, we determined that RF was the most effective classifier for before and after surgery based on gut microbiota data. Therefore, it was utilized for subsequent model predictions and interpretability.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 SHapley Additive exPlanations (SHAP) interpretability analysis\u003c/h2\u003e\u003cp\u003eTo evaluate the contribution of individual gut microbial taxa to model predictions, we employed SHAP, an explainable artificial intelligence (XAI) framework based on cooperative game theory [59, 60]. SHAP provides both local (sample-specific) and global (feature-level) interpretability, making it suitable for identifying bacterial taxa that consistently influence the model outcomes [61]. Compared to conventional feature importance metrics, SHAP can account for nonlinear interactions and heterogeneous effects [62], which are common in microbiome data.\u003c/p\u003e\u003cp\u003eSHAP analyses were conducted in Python (v3.11.4) using the shap package (v0.45.0), based on RF models. We applied the TreeExplainer function with the interventional feature perturbation setting, which reduces bias from correlated features and provides more robust estimates of contribution [63]. Feature importance rankings were obtained from the mean absolute SHAP values across all samples, and visualization plots were generated to facilitate interpretation. Consequently, several microbial genera were identified with consistent high contributions to model predictions. These genera were considered potential biomarkers of clinical importance towards post-colorectal surgery management.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Statistical analysis and data visualization\u003c/h2\u003e\u003cp\u003eVariations in the bacterial diversity based on the temporal gut microbial changes before and after surgery were assessed using Chao1 indices for alpha diversity [64], and unweighted UniFrac distances for beta diversity [65]. The top 20 bacterial species between the two groups were visualized using bar plots created in Microsoft Excel 2024. Functional metagenomic profiles were inferred with Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) [66], and correlations with species-level abundance were tested using statistical analysis of meta genomic profiles (STAMP) software (v2.1.3) [67]. Categorical variables were compared using the χ\u0026sup2; test, and continuous variables with the Wilcoxon rank-sum test. Differences in diversity metrics were evaluated using the Kruskal\u0026ndash;Wallis test and PERMANOVA. A p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. Statistical analyses were conducted in SPSS v24 (IBM, USA).\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Variations in bacterial community structure and composition before and after colorectal surgery\u003c/h2\u003e\u003cp\u003eBacterial community diversity before and after colorectal surgery was evaluated using both alpha and beta diversity metrics. Alpha diversity, as measured by the Chao1 index \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e, was significantly higher in pre-surgery samples compared to post-surgery samples (Kruskal\u0026ndash;Wallis test, p\u0026thinsp;=\u0026thinsp;0.038), which indicated a reduction in within-sample richness following surgery. Conversely, beta diversity analysis based on unweighted UniFrac distances did not reveal statistically significant overall differences between pre- and post-surgery groups (PERMANOVA, p\u0026thinsp;=\u0026thinsp;0.104). Nonetheless, principal coordinate analysis (PCoA) suggested partial but not complete separation between the two groups \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e, reflecting some degree of inter-individual variability in bacterial community structure.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe 16S rRNA sequencing revealed 322 classifiable bacterial genera in the fecal samples collected from study patients. To explore compositional shifts, we focused on the top 20 most abundant genera \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Post-surgery samples were enriched in \u003cem\u003eParabacteroides\u003c/em\u003e, \u003cem\u003eCollinsella\u003c/em\u003e, and \u003cem\u003eAkkermansia\u003c/em\u003e, while pre-surgery samples showed higher relative abundances of \u003cem\u003eFusobacterium\u003c/em\u003e. When compared with external reference datasets [47\u0026ndash;49], pre-surgery samples were characterized by higher levels of \u003cem\u003eBacteroides\u003c/em\u003e and \u003cem\u003eStreptococcus\u003c/em\u003e, whereas post-surgery samples were enriched in \u003cem\u003ePhocaeicola\u003c/em\u003e, \u003cem\u003eFaecalibacterium\u003c/em\u003e, and \u003cem\u003eRuminococcus\u003c/em\u003e \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. These compositional differences were statistically supported (p\u0026thinsp;=\u0026thinsp;0.044, Kruskal\u0026ndash;Wallis test), implying that the variation in bacterial community could have been influenced by surgery.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Gut microbiota-based classification model for pre- and post-surgery colorectal surgery\u003c/h2\u003e\u003cp\u003eTo assess the biomarker potential of gut microbiota for pre- and post-surgery classification, we trained three supervised classifiers (RF, SVM, XGB) and evaluated their performance using AUROC, accuracy, precision, recall, specificity, and F1 score. These evaluation metrics are represented as averages from 20 repetitions of five-fold cross-validation.\u003c/p\u003e\u003cp\u003eAs summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, the RF classifier achieved the best overall performance, with an AUROC of 0.90, accuracy of 89%, and F1 score of 0.88. XGB also performed well (AUROC\u0026thinsp;=\u0026thinsp;0.83, accuracy\u0026thinsp;=\u0026thinsp;80%), while SVM achieved slightly higher accuracy (87%) but a lower AUROC (0.78). These findings suggest that tree-based models, particularly RF, are better suited for distinguishing pre- and post-surgery microbial profiles.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance comparison of Support Vector Machine (SVM), Random Forest (RF), and XGBoost (XGB) classifiers across evaluation metrics\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eClassifier\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAccuracy\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePrecision\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eRecall\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eF1\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eAUC ROC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eAverage Precision\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eSVM\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e0.7185\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1107\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e0.7579\u0026thinsp;\u0026plusmn;\u0026thinsp;0.2129\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e0.5952\u0026thinsp;\u0026plusmn;\u0026thinsp;0.2327\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c5\"\u003e\u003cp\u003e0.6276\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1836\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c6\"\u003e\u003cp\u003e0.7831\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1168\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c7\"\u003e\u003cp\u003e0.7973\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1157\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRF\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e0.8447\u0026thinsp;\u0026plusmn;\u0026thinsp;0.0953\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e0.9042\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1270\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e0.7457\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1847\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c5\"\u003e\u003cp\u003e0.7996\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1338\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c6\"\u003e\u003cp\u003e0.8875\u0026thinsp;\u0026plusmn;\u0026thinsp;0.0865\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c7\"\u003e\u003cp\u003e0.9078\u0026thinsp;\u0026plusmn;\u0026thinsp;0.0674\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eXGB\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e0.7750\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1012\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e0.7748\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1490\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e0.7274\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1612\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c5\"\u003e\u003cp\u003e0.7361\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1195\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c6\"\u003e\u003cp\u003e0.8316\u0026thinsp;\u0026plusmn;\u0026thinsp;0.0948\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c7\"\u003e\u003cp\u003e0.8547\u0026thinsp;\u0026plusmn;\u0026thinsp;0.0831\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eWe further examined the RF model\u0026rsquo;s predicted probability distributions. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, pre-surgery samples displayed broader and lower classification scores, whereas post-surgery samples exhibited more narrowly distributed, higher scores. This difference was statistically significant (Mann\u0026ndash;Whitney U\u0026thinsp;=\u0026thinsp;0.00, p\u0026thinsp;=\u0026thinsp;0.005), indicating that a RF-based model can reliably separate the two groups based on their gut microbiota differences.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Model explainability and feature importance\u003c/h2\u003e\u003cp\u003eTo identify the microbial taxa driving classification in post-surgery for potential biomarkers, we applied SHAP to the RF model. The top 10 genera ranked by average absolute SHAP values are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB. Among these, \u003cem\u003eAkkermansia\u003c/em\u003e, \u003cem\u003eCollinsella\u003c/em\u003e, \u003cem\u003eDorea\u003c/em\u003e, and \u003cem\u003eAnaerostipes\u003c/em\u003e were strong positive contributors to post-surgery classification, whereas lower relative abundances of \u003cem\u003eFaecalibacterium\u003c/em\u003e and \u003cem\u003eBlautia\u003c/em\u003e were associated with higher prediction scores. Other genera, including Parabacteroides, \u003cem\u003eRoseburia\u003c/em\u003e, \u003cem\u003eSenegalimassilia\u003c/em\u003e, and \u003cem\u003eBilophila\u003c/em\u003e, contributed less to model performance.\u003c/p\u003e\u003cp\u003eTo further validate these findings, we compared the relative abundances of selected genera among the ranked top 10 between pre- and post-surgery groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Post-surgery, unlike pre-surgery samples showed higher statistically significant (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) mean abundances of \u003cem\u003eAkkermansia\u003c/em\u003e (6.8 Vs 2.5%), \u003cem\u003eBlautia\u003c/em\u003e (4.1% Vs 1.9%), \u003cem\u003eFaecalibacterium\u003c/em\u003e (5.3% Vs 2.4%), and \u003cem\u003eParabacteroides\u003c/em\u003e (6.7% Vs 1.1%). Conversely, pre-surgery samples had higher abundances of \u003cem\u003eBacteroides\u003c/em\u003e (5.8% Vs 1.5%), and \u003cem\u003eStreptococcus\u003c/em\u003e (4.5% Vs 0.5%) as comparted to post-surgery.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Comparative analysis of microbial functional pathways before and after surgery\u003c/h2\u003e\u003cp\u003eWe examined microbial functional pathways using RF feature importance analysis and compared them between pre-and post-surgery groups. A RF classifier was trained with 100 trees using a deviance splitting criterion, a minimum leaf size of 2, and feature selection at each split set to 0.577 sqrt (nvars). The classifier achieved an AUROC of 0.87, indicating good discriminatory performance. The top-ranked pathways identified by RF included several biosynthetic and metabolic processes (\u003cb\u003eFig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). Among these, queuosine biosynthesis, lipid IVA biosynthesis, fatty acid elongation, and acetylene degradation were more enriched in post-surgery samples, whereas pathways linked to bacterial cell wall biosynthesis (e.g., UDP-N-acetylmuramoyl-pentapeptide biosynthesis II) were relatively enriched in pre-surgery samples.\u003c/p\u003e\u003cp\u003eComparative abundance testing indicated that queuosine biosynthesis, a pathway implicated in maintaining translational fidelity and epithelial homeostasis, was significantly elevated post-surgery (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). In contrast, peptidoglycan biosynthesis pathways, important for structural integrity and bacterial proliferation, showed higher relative abundance both pre- and post-surgery, suggesting their general role in supporting microbiota re-establishment after surgical perturbation.\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe gut microbiota plays a central role in CRC, acting not only as a contributor to disease development [68], but also as a therapeutic target [69], as wells as a potential diagnosis and prognosis biomarker [70]. Alternations as a result of colorectal surgery would influence the patient\u0026rsquo;s wound healing, and susceptibility to complications such as surgical site infection, anastomotic leakage, and recurrence [71]. Therefore, understanding and leveraging these alternations may improve post-surgery outcomes and the overall CRC patient survival. In this study, we observed that post colorectal surgery was associated with a significant reduction in gut microbial richness (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, p\u0026thinsp;=\u0026thinsp;0.038), whereas overall community structure showed no statistically significant differences (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB, p\u0026thinsp;=\u0026thinsp;0.104). This decline in alpha diversity suggests that surgical intervention could have disrupted the ecological balance of the gut microbiota, likely due to preoperative bowel preparation [15], perioperative antibiotics [72], oxygen exposure for obligate microbes, and mucosal injury [68]. Such disruptions have been consistently reported in previous studies [15, 73, 74], where perioperative interventions transiently reduce microbial diversity, which could impair the resilience of the gut ecosystem against pathogens [75].\u003c/p\u003e\u003cp\u003eAlthough β-diversity differences were not statistically significant, ordination analyses revealed subtle compositional separation between pre- and post-surgery samples. Similar trends have been documented in CRC cohorts, where surgery-induced perturbations reshaped the microbiota without producing large-scale shifts in global community structure [50, 52, 76]. This could be indicating that core microbial taxa may be resilient to surgical stress [77, 78]. Taken altogether, these findings highlight that while the overall microbial landscape remains relatively stable, surgery induces measurable ecological stress that reduces diversity with minimal alteration to core community compositions.\u003c/p\u003e\u003cp\u003eAnalysis of the top 20 abundant genera in both groups revealed that \u003cem\u003ePrevotella\u003c/em\u003e, \u003cem\u003eAlistipes\u003c/em\u003e, \u003cem\u003eFusobacterium\u003c/em\u003e, and \u003cem\u003eCollinsella\u003c/em\u003e were highly enriched in pre-surgery samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) as compared to post-surgery. Several of these taxa are well-recognized for their association with colorectal carcinogenesis [79]. For instance, \u003cem\u003ePrevotella\u003c/em\u003e, an anaerobic Gram-negative bacterium, has been linked to malignant progression of CRC and attenuation of the anticancer efficacy of FOLFOX chemotherapy [80]. \u003cem\u003eAlistipes\u003c/em\u003e has been described as a pro-inflammatory taxon with potential pathogenic roles in CRC development [81], while \u003cem\u003eFusobacterium\u003c/em\u003e is a consistently reported CRC-associated genus implicated in tumor initiation and immune modulation [82\u0026ndash;85]. Similarly, higher abundances of \u003cem\u003eCollinsella\u003c/em\u003e have been observed in CRC patients compared with healthy controls [21, 86]. Together, these patterns underscore that the pre-surgery gut environment is enriched in taxa with pro-carcinogenic potential.\u003c/p\u003e\u003cp\u003eTo validate these findings, we further examined the four independent external CRC microbiome datasets [50, 87\u0026ndash;89], which consistently demonstrated enrichment of pro-CRC bacteria such as \u003cem\u003eBacteroides\u003c/em\u003e, \u003cem\u003eStreptococcus\u003c/em\u003e, and \u003cem\u003eEnterococcus\u003c/em\u003e in pre-surgery samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, p\u0026thinsp;=\u0026thinsp;0.0062). This concordance strengthens the evidence that CRC patients harbor higher levels of pathogenic and pro-inflammatory taxa before surgical intervention [90, 91].\u003c/p\u003e\u003cp\u003eConversely, post-surgery samples were enriched with genera considered beneficial to gut and host health, including \u003cem\u003eParabacteroides\u003c/em\u003e and \u003cem\u003eAkkermansia\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Studies widely indicate \u003cem\u003eAkkermansia\u003c/em\u003e as a next-generation probiotic with protective roles in gastrointestinal disorders, including CRC [92\u0026ndash;95]. Similarly, \u003cem\u003eParabacteroides\u003c/em\u003e has been shown to exert anti-inflammatory effects and attenuate tumorigenesis in experimental models [96, 97]. Our analysis of the four independent datasets supported these observations with higher abundances of \u003cem\u003ePhocaeicola\u003c/em\u003e, \u003cem\u003eBlautia\u003c/em\u003e, \u003cem\u003eFaecalibacterium\u003c/em\u003e, and \u003cem\u003eRuminococcus\u003c/em\u003e in post-surgery samples; which are frequently associated with anti-inflammatory activity and maintenance of gut homeostasis [98\u0026ndash;103]. These findings suggest that colorectal surgery not only reduces pro-CRC taxa but may also promotes the expansion of potentially beneficial bacteria. Such shifts could have clinical implications, as enrichment of probiotic and anti-inflammatory taxa may contribute to improved immune responses, wound healing, and resilience against postoperative complications, including surgical site infections, anastomotic leakage, and recurrence [104\u0026ndash;107].\u003c/p\u003e\u003cp\u003eWe applied three supervised machine learning models RF, XGB, and SVM to genus-level abundance data for pre- and post-surgery predictions. Among these, the RF model achieved the highest predictive performance (AUROC\u0026thinsp;=\u0026thinsp;0.90), outperforming XGB (0.83) and SVM (0.78) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). These findings are consistent with previous studies demonstrating that tree-based classifiers are particularly effective in handling the complex, non-linear relationships characteristic of microbiome data [108, 109]. The predicted probability distributions from the RF model further demonstrated a clear separation between pre- and post-surgery samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA), highlighting its robustness in capturing microbial differences associated with surgical intervention.\u003c/p\u003e\u003cp\u003eTo enhance interpretability of our RF model, we applied SHAP analysis, which highlighted the genera most influential in driving model predictions [110\u0026ndash;113]. These taxa were considered potential post-surgery biomarkers (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Several of them were commensal bacteria and further analysis of their relative abundance based on 16s rRNA sequencing results indicated that they were indeed enriched in post-surgery samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA \u003cb\u003e- D\u003c/b\u003e). For example, \u003cem\u003eAkkermansia\u003c/em\u003e emerged as the strongest positive predictor, aligning with its proposed role as a next-generation probiotic with beneficial effects on gut health and CRC outcomes [114, 115]. However, among these top 10 two bacterial features were enriched in pre-surgery samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE \u003cb\u003e- F\u003c/b\u003e). This showed that surgical resection could potentially lead to restoration of beneficial taxa in the gut, which would reduce the abundance of opportunistic taxa. Recent studies have reported a similar observation in which the abundance of short chain fatty acids (SCFAs) producers as well as other commensal bacteria were increased post-surgery [52, 116]. These findings not only demonstrate the predictive power of machine learning to identify potential markers but also provide their biological interpretability which could inform postoperative monitoring and, potentially, microbiota-based therapeutic strategies.\u003c/p\u003e\u003cp\u003ePathway analysis revealed distinct functional signatures between pre- and post-surgery samples. The top 10 features identified by RF were predominantly associated with biosynthetic processes. In particular, post-surgery samples (\u003cb\u003eFig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e) were enriched for pathways such as queuosine biosynthesis (linked to translational fidelity and epithelial homeostasis) [117], peptidoglycan biosynthesis IV (Enterococcus faecium) (structural integrity) [118], and L-rhamnose degradation I (bacterial carbon source utilization) [119]. These functions may suggest metabolic adaptations that could support the re-establishment of gut microbial communities following surgical disruption [120].\u003c/p\u003e\u003cp\u003eHowever, this study had some limitations. Firstly, the pre and post-surgery microbial analyses were based on 16S rRNA gene sequencing, which captures only bacterial taxonomic profiles and does not provide comprehensive information on other microbial components or functional pathways. Future studies using shotgun metagenomic sequencing would allow a broader characterization of the gut microbiome and enable stronger correlations with clinical outcomes. Although our model accurately distinguished pre- and post-surgery samples and identified potential post-surgery biomarkers, their functional relevance remains to be experimentally validated. Future independent patient cohorts will be essential to establish their clinical relevance for CRC prognosis. Finally, our relatively small cohort may have limited the statistical power to adequately examine the associations between the two groups. Larger patient cohorts will be essential to strengthen the robustness and validity of future analyses.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study characterized gut microbiota changes in CRC patients before and after surgery. Microbial diversity was higher pre-surgery but declined post-surgery, likely reflecting the effects of intensive bowel preparation. Post-surgery samples were enriched with commensal taxa, suggesting a potential restoration of beneficial microbiota. Such microbial shifts may be harnessed to improved gut health, thereby supporting recovery and lowering the risk of postoperative complications. Using explainable AI with SHAP, we identified potential biomarkers, including \u003cem\u003eAkkermansia\u003c/em\u003e, which was as key positive contributor in post-surgery. However, the limited cohort size and lack of clinical validation of the identified biomarkers may restrict the generalizability of these findings. Future studies with larger, clinically evaluated cohorts will be essential to confirm the relevance of these biomarkers for them to be translated into tools that can guide CRC prognosis and treatment.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no conflicts of interest\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data presented in this study are available on request from the corresponding author.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval and consent\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was approved by the Institutional Ethics Committee of Dalin Tzu-Chi Hospital (B11304006), and all participants provided written informed consent.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by Hualien Tzu-Chi General Hospital, Dalin Tzu-Chi Hospital (DTCRD114(2)-C-04), and Ditmanson Medical Foundation Chiayi Christian Hospital-National Chung Cheng University Joint Research Program (CYCH-CCU-2024-06).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMutebi John Kenneth\u003c/strong\u003e: Writing-Original Draft, Investigation, Conceptualization, Data Curation, Validation, Formal Analysis, and Writing-reviewing and Editing.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eChuan-Yin Fang\u003c/strong\u003e: Conceptualization, Investigation, Validation, Writing-Reviewing and Editing, Resources, and Supervision.\u0026nbsp;\u003cstrong\u003eChin-Chia Wu\u003c/strong\u003e: Conceptualization, Investigation, Validation, Writing-Reviewing and Editing, Resources, and Supervision.\u0026nbsp;\u003cstrong\u003eMichael W Y Ch\u003c/strong\u003e\u003cstrong\u003ean\u003c/strong\u003e:\u0026nbsp;Formal Analysis, Conceptualization, and Writing-reviewing and Editing\u0026nbsp;\u003cstrong\u003eBing-Mu Hsu:\u003c/strong\u003e Conceptualization, Investigation, Project administration, Funding acquisition, Writing-reviewing and Editing, Resources, and Supervision.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eKlimeck L, Heisser T, Hoffmeister M, Brenner H (2023) Colorectal cancer: A health and economic problem. Best practice \u0026amp; research clinical gastroenterology 66:101839\u003c/li\u003e\n\u003cli\u003eMatsuda T, Fujimoto A, Igarashi Y (2025) Colorectal Cancer: Epidemiology, Risk Factors, and Public Health Strategies. Digestion 106(2):91-99. https://doi.org/10.1159/000543921\u003c/li\u003e\n\u003cli\u003eAfzaal M, Saeed F, Shah YA, Hussain M, Rabail R, Socol CT, Hassoun A, Pateiro M, Lorenzo JM, Rusu AV (2022) Human gut microbiota in health and disease: Unveiling the relationship. Frontiers in microbiology 13:999001\u003c/li\u003e\n\u003cli\u003eAndoh A (2016) Physiological Role of Gut Microbiota for Maintaining Human Health. Digestion 93(3):176-181. https://doi.org/10.1159/000444066\u003c/li\u003e\n\u003cli\u003eSingh G, Chaudhry Z, Boyadzhyan A, Sasaninia K, Rai V (2025) Dysbiosis and colorectal cancer: conducive factors, biological and molecular role, and therapeutic prospectives. Explor Target Antitumor Ther 6:1002329. https://doi.org/10.37349/etat.2025.1002329\u003c/li\u003e\n\u003cli\u003eS\u0026aacute;nchez-Alcoholado L, Ramos-Molina B, Otero A, Laborda-Illanes A, Ord\u0026oacute;\u0026ntilde;ez R, Medina JA, G\u0026oacute;mez-Mill\u0026aacute;n J, Queipo-Ortu\u0026ntilde;o MI (2020) The Role of the Gut Microbiome in Colorectal Cancer Development and Therapy Response. Cancers (Basel) 12(6). https://doi.org/10.3390/cancers12061406\u003c/li\u003e\n\u003cli\u003eZou S, Fang L, Lee MH (2018) Dysbiosis of gut microbiota in promoting the development of colorectal cancer. Gastroenterol Rep (Oxf) 6(1):1-12. https://doi.org/10.1093/gastro/gox031\u003c/li\u003e\n\u003cli\u003eOliero M, Hajjar R, Cuisiniere T, Fragoso G, Calv\u0026eacute; A, Dagbert F, Loungnarath R, Sebajang H, Schwenter F, Wassef R (2022) Prevalence of pks+ bacteria and enterotoxigenic Bacteroides fragilis in patients with colorectal cancer. Gut Pathogens 14(1):51\u003c/li\u003e\n\u003cli\u003ePhipps AI, Hill CM, Lin G, Malen RC, Reedy AM, Kahsai O, Ammar H, Curtis K, Ma N, Randolph TW (2025) Fusobacterium nucleatum enrichment in colorectal tumor tissue: Associations with tumor characteristics and survival outcomes. Gastro Hep Advances:100644\u003c/li\u003e\n\u003cli\u003eLee JB, Kim K-A, Cho HY, Kim D, Kim WK, Yong D, Lee H, Yoon SS, Han DH, Han YD (2021) Association between Fusobacterium nucleatum and patient prognosis in metastatic colon cancer. Scientific reports 11(1):20263\u003c/li\u003e\n\u003cli\u003ePaik J-H, Ryu C-G, Hwang D-Y (2023) Risk factors of recurrence in TNM stage I colorectal cancer. Annals of Surgical Treatment and Research 104(5):281-287\u003c/li\u003e\n\u003cli\u003eLi T, Liu Z, Bai F, Xiao H, Zhou H (2023) Surgical resection for second primary colorectal cancer: a population-based study. Front Med (Lausanne) 10:1167777. https://doi.org/10.3389/fmed.2023.1167777\u003c/li\u003e\n\u003cli\u003eOng WL, Morarasu S, Lunca S, Pruna RM, Roata CE, Dimofte GM (2025) Impact of Clostridium difficile Infection Versus Colonization on Postoperative Outcomes After Oncological Colorectal Surgery: An Observational Single-Center Study With Propensity Score Analysis. J Surg Oncol 131(3):489-497. https://doi.org/10.1002/jso.27923\u003c/li\u003e\n\u003cli\u003ePak H, Maghsoudi LH, Soltanian A, Gholami F (2020) Surgical complications in colorectal cancer patients. Annals of medicine and surgery 55:13-18\u003c/li\u003e\n\u003cli\u003eNalluri-Butz H, Bobel MC, Nugent J, Boatman S, Emanuelson R, Melton-Meaux G, Madoff RD, Jahansouz C, Staley C, Gaertner WB (2022) A pilot study demonstrating the impact of surgical bowel preparation on intestinal microbiota composition following colon and rectal surgery. Scientific reports 12(1):10559\u003c/li\u003e\n\u003cli\u003eShogan BD, Belogortseva N, Luong PM, Zaborin A, Lax S, Bethel C, Ward M, Muldoon JP, Singer M, An G, Umanskiy K, Konda V, Shakhsheer B, Luo J, Klabbers R, Hancock LE, Gilbert J, Zaborina O, Alverdy JC (2015) Collagen degradation and MMP9 activation by Enterococcus faecalis contribute to intestinal anastomotic leak. Sci Transl Med 7(286):286ra268. https://doi.org/10.1126/scitranslmed.3010658\u003c/li\u003e\n\u003cli\u003eEdomskis P, Goudberg MR, Sparreboom CL, Menon AG, Wolthuis AM, D\u0026apos;Hoore A, Lange JF (2021) Matrix metalloproteinase-9 in relation to patients with complications after colorectal surgery: a systematic review. Int J Colorectal Dis 36(1):1-10. https://doi.org/10.1007/s00384-020-03724-6\u003c/li\u003e\n\u003cli\u003eShogan BD, Chen J, Duchalais E, Collins D, Chang M, Krull K, Krezalek MA, Larson DW, Walther-Antonio MR, Chia N (2020) Alterations of the rectal microbiome are associated with the development of postoperative ileus in patients undergoing colorectal surgery. Journal of Gastrointestinal Surgery 24(7):1663-1672\u003c/li\u003e\n\u003cli\u003eJin Y, Geng R, Liu Y, Liu L, Jin X, Zhao F, Feng J, Wei Y (2020) Prediction of Postoperative Ileus in Patients With Colorectal Cancer by Preoperative Gut Microbiota. Front Oncol 10:526009. https://doi.org/10.3389/fonc.2020.526009\u003c/li\u003e\n\u003cli\u003eZhao L, Fang Y, Zhang J, Wei C, Ji H, Zhao J, Wang D, Tang D (2024) Changes in Intestinal Microbiota and Their Relationship With Patient Characteristics in Colorectal Cancer. Clin Med Insights Oncol 18:11795549241307632. https://doi.org/10.1177/11795549241307632\u003c/li\u003e\n\u003cli\u003eSheng Q, Du H, Cheng X, Cheng X, Tang Y, Pan L, Wang Q, Lin J (2019) Characteristics of fecal gut microbiota in patients with colorectal cancer at different stages and different sites. Oncology letters 18(5):4834-4844\u003c/li\u003e\n\u003cli\u003eOh HH, Joo YE (2020) Novel biomarkers for the diagnosis and prognosis of colorectal cancer. Intest Res 18(2):168-183. https://doi.org/10.5217/ir.2019.00080\u003c/li\u003e\n\u003cli\u003eKim CW, Cha JM, Kwak MS (2021) Identification of Potential Biomarkers and Biological Pathways for Poor Clinical Outcome in Mucinous Colorectal Adenocarcinoma. Cancers 13(13):3280\u003c/li\u003e\n\u003cli\u003eHerlo LF, Salcudean A, Sirli R, Iurciuc S, Herlo A, Nelson-Twakor A, Alexandrescu L, Dumache R (2024) Gut Microbiota Signatures in Colorectal Cancer as a Potential Diagnostic Biomarker in the Future: A Systematic Review. Int J Mol Sci 25(14). https://doi.org/10.3390/ijms25147937\u003c/li\u003e\n\u003cli\u003eLi S, Konstantinov SR, Smits R, Peppelenbosch MP (2017) Bacterial Biofilms in Colorectal Cancer Initiation and Progression. Trends Mol Med 23(1):18-30. https://doi.org/10.1016/j.molmed.2016.11.004\u003c/li\u003e\n\u003cli\u003eQu R, Zhang Y, Ma Y, Zhou X, Sun L, Jiang C, Zhang Z, Fu W (2023) Role of the gut microbiota and its metabolites in tumorigenesis or development of colorectal cancer. Advanced Science 10(23):2205563\u003c/li\u003e\n\u003cli\u003eTsigalou C, Paraschaki A, Bragazzi NL, Aftzoglou K, Bezirtzoglou E, Tsakris Z, Vradelis S, Stavropoulou E (2023) Alterations of gut microbiome following gastrointestinal surgical procedures and their potential complications. Front Cell Infect Microbiol 13:1191126. https://doi.org/10.3389/fcimb.2023.1191126\u003c/li\u003e\n\u003cli\u003eWeaver L, Troester A, Jahansouz C (2024) The Impact of Surgical Bowel Preparation on the Microbiome in Colon and Rectal Surgery. Antibiotics 13(7):580\u003c/li\u003e\n\u003cli\u003eElbarmelgi MY, Shafik AA, Abd ElSamee AK, Tamer M (2024) Impact of pre-operative mechanical bowel preparation in preventing post-operative anastomotic leak: A meta-analysis. Asian J Surg. https://doi.org/10.1016/j.asjsur.2024.11.006\u003c/li\u003e\n\u003cli\u003ePetrou NA, Kontovounisios C (2022) The Use of Mechanical Bowel Preparation and Oral Antibiotic Prophylaxis in Elective Colorectal Surgery: A Call for Change in Practice. Cancers 14(23):5990\u003c/li\u003e\n\u003cli\u003eOhigashi S, Sudo K, Kobayashi D, Takahashi T, Nomoto K, Onodera H (2013) Significant changes in the intestinal environment after surgery in patients with colorectal cancer. J Gastrointest Surg 17(9):1657-1664. https://doi.org/10.1007/s11605-013-2270-x\u003c/li\u003e\n\u003cli\u003eChen Y, Wu N, Yan X, Kang L, Ou G, Zhou Z, Xu C, Feng J, Shi T (2025) Impact of gut microbiota on colorectal anastomotic healing (Review). Mol Clin Oncol 22(6):52. https://doi.org/10.3892/mco.2025.2847\u003c/li\u003e\n\u003cli\u003eHajjar R, Gonzalez E, Fragoso G, Oliero M, Alaoui AA, Calv\u0026eacute; A, Vennin Rendos H, Djediai S, Cuisiniere T, Laplante P, Gerkins C, Ajayi AS, Diop K, Taleb N, Th\u0026eacute;rien S, Schampaert F, Alratrout H, Dagbert F, Loungnarath R, Sebajang H, Schwenter F, Wassef R, Ratelle R, Debroux E, Cailhier JF, Routy B, Annabi B, Brereton NJB, Richard C, Santos MM (2023) Gut microbiota influence anastomotic healing in colorectal cancer surgery through modulation of mucosal proinflammatory cytokines. Gut 72(6):1143-1154. https://doi.org/10.1136/gutjnl-2022-328389\u003c/li\u003e\n\u003cli\u003eHuo R-X, Wang Y-J, Hou S-B, Wang W, Zhang C-Z, Wan X-H (2022) Gut mucosal microbiota profiles linked to colorectal cancer recurrence. World journal of gastroenterology 28(18):1946\u003c/li\u003e\n\u003cli\u003eGaines S, Shao C, Hyman N, Alverdy J (2018) Gut microbiome influences on anastomotic leak and recurrence rates following colorectal cancer surgery. Journal of British Surgery 105(2):e131-e141\u003c/li\u003e\n\u003cli\u003ePandey K, Umar S (2021) Microbiome in drug resistance to colon cancer. Curr Opin Physiol 23. https://doi.org/10.1016/j.cophys.2021.100472\u003c/li\u003e\n\u003cli\u003eGan Y, Yang H, Wang M, Li J (2025) Advances in drug resistance and resistance mechanisms of four colorectal cancer-associated gut microbiota. PeerJ 13:e19535\u003c/li\u003e\n\u003cli\u003eChen H-H, Ke T-W, Huang C-W, Jiang J-K, Chen C-C, Hsieh Y-Y, Teng H-W, Lin B-W, Liang Y-H, Su Y-L (2021) Taiwan society of colon and rectal surgeons consensus on mCRC treatment. Frontiers in oncology 11:764912\u003c/li\u003e\n\u003cli\u003eYoshida T, Homma S, Ichikawa N, Ohno Y, Miyaoka Y, Matsui H, Imaizumi K, Ishizu H, Funakoshi T, Koike M (2023) Preoperative mechanical bowel preparation using conventional versus hyperosmolar polyethylene glycol-electrolyte lavage solution before laparoscopic resection for colorectal cancer (TLUMP test): a phase III, multicenter randomized controlled non-inferiority trial. Journal of gastroenterology 58(9):883-893\u003c/li\u003e\n\u003cli\u003eAssociation WM (2025) World Medical Association Declaration of Helsinki: ethical principles for medical research involving human participants. Jama 333(1):71-74\u003c/li\u003e\n\u003cli\u003eKai S, Matsuo Y, Nakagawa S, Kryukov K, Matsukawa S, Tanaka H, Iwai T, Imanishi T, Hirota K (2019) Rapid bacterial identification by direct PCR amplification of 16S rRNA genes using the MinION\u0026trade; nanopore sequencer. FEBS Open Bio 9(3):548-557. https://doi.org/10.1002/2211-5463.12590\u003c/li\u003e\n\u003cli\u003eSrinivasan R, Karaoz U, Volegova M, MacKichan J, Kato-Maeda M, Miller S, Nadarajan R, Brodie EL, Lynch SV (2015) Use of 16S rRNA gene for identification of a broad range of clinically relevant bacterial pathogens. PloS one 10(2):e0117617\u003c/li\u003e\n\u003cli\u003eBuetas E, Jord\u0026aacute;n-L\u0026oacute;pez M, L\u0026oacute;pez-Rold\u0026aacute;n A, D\u0026rsquo;Auria G, Mart\u0026iacute;nez-Priego L, De Marco G, Carda-Di\u0026eacute;guez M, Mira A (2024) Full-length 16S rRNA gene sequencing by PacBio improves taxonomic resolution in human microbiome samples. BMC genomics 25(1):310\u003c/li\u003e\n\u003cli\u003eAsif A, Koner S, Chen J-S, Hussain A, Huang S-W, Hussain B, Hsu B-M (2024) Uncovering the microbial community structure and physiological profiles of terrestrial mud volcanoes: A comprehensive metagenomic insight towards their trichloroethylene biodegradation potentiality. Environmental Research 258:119457\u003c/li\u003e\n\u003cli\u003eGupta S, Mortensen MS, Schj\u0026oslash;rring S, Trivedi U, Vestergaard G, Stokholm J, Bisgaard H, Krogfelt KA, S\u0026oslash;rensen SJ (2019) Amplicon sequencing provides more accurate microbiome information in healthy children compared to culturing. Communications biology 2(1):291\u003c/li\u003e\n\u003cli\u003eGoldfarb T, Kodali Vamsi K, Pujar S, Brover V, Robbertse B, Farrell Catherine M, Oh D-H, Astashyn A, Ermolaeva O, Haddad D, Hlavina W, Hoffman J, Jackson John D, Joardar Vinita S, Kristensen D, Masterson P, McGarvey Kelly M, McVeigh R, Mozes E, Murphy Michael R, Schafer Susan S, Souvorov A, Spurrier B, Strope Pooja K, Sun H, Vatsan Anjana R, Wallin C, Webb D, Brister J R, Hatcher E, Kimchi A, Klimke W, Marchler-Bauer A, Pruitt Kim D, Thibaud-Nissen F, Murphy Terence D (2024) NCBI RefSeq: reference sequence standards through 25 years of curation and annotation. Nucleic Acids Research 53(D1):D243-D257. https://doi.org/10.1093/nar/gkae1038\u003c/li\u003e\n\u003cli\u003eAi D, Pan H, Han R, Li X, Liu G, Xia LC (2019) Using Decision Tree Aggregation with Random Forest Model to Identify Gut Microbes Associated with Colorectal Cancer. Genes 10(2):112\u003c/li\u003e\n\u003cli\u003eZheng Y, Fang Z, Xue Y, Zhang J, Zhu J, Gao R, Yao S, Ye Y, Wang S, Lin C, Chen S, Huang H, Hu L, Jiang GN, Qin H, Zhang P, Chen J, Ji H (2020) Specific gut microbiome signature predicts the early-stage lung cancer. Gut Microbes 11(4):1030-1042. https://doi.org/10.1080/19490976.2020.1737487\u003c/li\u003e\n\u003cli\u003eWu H, Li Y, Jiang Y, Li X, Wang S, Zhao C, Yang X, Chang B, Yang J, Qiao J (2025) Machine learning prediction of obesity-associated gut microbiota: identifying Bifidobacterium pseudocatenulatum as a potential therapeutic target. Frontiers in Microbiology 15:1488656\u003c/li\u003e\n\u003cli\u003eCong J, Zhu H, Liu D, Li T, Zhang C, Zhu J, Lv H, Liu K, Hao C, Tian Z (2018) A pilot study: changes of gut microbiota in post-surgery colorectal cancer patients. Frontiers in microbiology 9:2777\u003c/li\u003e\n\u003cli\u003eHern\u0026aacute;ndez‐Gonz\u0026aacute;lez PI, Barqu\u0026iacute;n J, Ortega‐Ferrete A, Pat\u0026oacute;n V, Ponce‐Alonso M, Romero‐Hern\u0026aacute;ndez B, Oca\u0026ntilde;a J, Caminoa A, Conde‐Moreno E, Galeano J (2023) Anastomotic leak in colorectal cancer surgery: Contribution of gut microbiota and prediction approaches. Colorectal Disease 25(11):2187-2197\u003c/li\u003e\n\u003cli\u003ePng C-W, Chua Y-K, Law J-H, Zhang Y, Tan K-K (2022) Alterations in co-abundant bacteriome in colorectal cancer and its persistence after surgery: a pilot study. Scientific Reports 12(1):9829\u003c/li\u003e\n\u003cli\u003eLee SY, Park H-M, Kim CH, Kim HR (2023) Dysbiosis of gut microbiota during fecal stream diversion in patients with colorectal cancer. Gut Pathogens 15(1):40\u003c/li\u003e\n\u003cli\u003ePedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V (2011) Scikit-learn: Machine learning in Python. the Journal of machine Learning research 12:2825-2830\u003c/li\u003e\n\u003cli\u003eAbdullah DM, Abdulazeez AM (2021) Machine learning applications based on SVM classification a review. Qubahan Academic Journal 1(2):81-90\u003c/li\u003e\n\u003cli\u003eSalman HA, Kalakech A, Steiti A (2024) Random forest algorithm overview. Babylonian Journal of Machine Learning 2024:69-79\u003c/li\u003e\n\u003cli\u003eChen T, Guestrin C (2016) Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. pp 785-794\u003c/li\u003e\n\u003cli\u003eRainio O, Teuho J, Kl\u0026eacute;n R (2024) Evaluation metrics and statistical tests for machine learning. Sci Rep 14(1):6086. https://doi.org/10.1038/s41598-024-56706-x\u003c/li\u003e\n\u003cli\u003eLundberg SM, Lee S-I (2017) A unified approach to interpreting model predictions. Advances in neural information processing systems 30\u003c/li\u003e\n\u003cli\u003eLundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, Katz R, Himmelfarb J, Bansal N, Lee S-I (2020) From local explanations to global understanding with explainable AI for trees. Nature machine intelligence 2(1):56-67\u003c/li\u003e\n\u003cli\u003eSalih AM, Raisi‐Estabragh Z, Galazzo IB, Radeva P, Petersen SE, Lekadir K, Menegaz G (2025) A perspective on explainable artificial intelligence methods: SHAP and LIME. Advanced Intelligent Systems 7(1):2400304\u003c/li\u003e\n\u003cli\u003eChen Y, Ye Y, Liu X, Yin C, Jones CA (2025) Examining the nonlinear and spatial heterogeneity of housing prices in urban Beijing: an application of GeoShapley. Habitat International 162:103439\u003c/li\u003e\n\u003cli\u003eDeb D, Smith RM (2021) Application of random forest and SHAP tree explainer in exploring spatial (in) justice to aid urban planning. ISPRS International Journal of Geo-Information 10(9):629\u003c/li\u003e\n\u003cli\u003eDeng D, Zhao L, Song H, Wang H, Cao H, Cui H, Zhou Y, Cui R (2025) Microbiome analysis of gut microbiota in patients with colorectal polyps and healthy individuals. Sci Rep 15(1):7126. https://doi.org/10.1038/s41598-025-91626-4\u003c/li\u003e\n\u003cli\u003eYang Z, Xu F, Li H, He Y (2021) Beyond samples: a metric revealing more connections of gut microbiota between individuals. Computational and Structural Biotechnology Journal 19:3930-3937\u003c/li\u003e\n\u003cli\u003eIjoma GN, Nkuna R, Mutungwazi A, Rashama C, Matambo TS (2021) Applying PICRUSt and 16S rRNA functional characterisation to predicting co-digestion strategies of various animal manures for biogas production. Scientific reports 11(1):19913\u003c/li\u003e\n\u003cli\u003eParks DH, Tyson GW, Hugenholtz P, Beiko RG (2014) STAMP: statistical analysis of taxonomic and functional profiles. Bioinformatics 30(21):3123-3124\u003c/li\u003e\n\u003cli\u003eWong CC, Yu J (2023) Gut microbiota in colorectal cancer development and therapy. Nature reviews Clinical oncology 20(7):429-452\u003c/li\u003e\n\u003cli\u003eHu Y, Zhou P, Deng K, Zhou Y, Hu K (2024) Targeting the gut microbiota: a new strategy for colorectal cancer treatment. Journal of Translational Medicine 22(1):915\u003c/li\u003e\n\u003cli\u003eLi G, Zhao D, Ouyang B, Chen Y, Zhao Y (2025) Intestinal microbiota as biomarkers for different colorectal lesions based on colorectal cancer screening participants in community. Frontiers in Microbiology 16:1529858\u003c/li\u003e\n\u003cli\u003eAgnes A, Puccioni C, D\u0026apos;Ugo D, Gasbarrini A, Biondi A, Persiani R (2021) The gut microbiota and colorectal surgery outcomes: facts or hype? A narrative review. BMC Surg 21(1):83. https://doi.org/10.1186/s12893-021-01087-5\u003c/li\u003e\n\u003cli\u003eVăcărean-Trandafir IC, Amărandi R-M, Ivanov IC, Dragoș LM, Mențel M, Iacob Ş, Muşină A-M, Bărgăoanu E-R, Roată CE, Morărașu Ș (2025) Impact of antibiotic prophylaxis on gut microbiota in colorectal surgery: insights from an Eastern European stewardship study. Frontiers in Cellular and Infection Microbiology 14:1468645\u003c/li\u003e\n\u003cli\u003eWeaver L, Troester A, Jahansouz C (2024) The Impact of Surgical Bowel Preparation on the Microbiome in Colon and Rectal Surgery. Antibiotics (Basel) 13(7). https://doi.org/10.3390/antibiotics13070580\u003c/li\u003e\n\u003cli\u003eKohn J, Troester A, Ziegert Z, Frebault J, Boatman S, Martell M, Nalluri-Butz H, Bobel MC, Goffredo P, Johnson AJ, Jahansouz C, Staley C, Gaertner WB (2025) The Role of Surgical and Perioperative Factors in Shaping Gut Microbiome Recovery After Colorectal Surgery. Antibiotics 14(9):881\u003c/li\u003e\n\u003cli\u003eKoliarakis I, Athanasakis E, Sgantzos M, Mariolis-Sapsakos T, Xynos E, Chrysos E, Souglakos J, Tsiaoussis J (2020) Intestinal Microbiota in Colorectal Cancer Surgery. Cancers (Basel) 12(10). https://doi.org/10.3390/cancers12103011\u003c/li\u003e\n\u003cli\u003eDeng X, Li Z, Li G, Li B, Jin X, Lyu G (2018) Comparison of microbiota in patients treated by surgery or chemotherapy by 16S rRNA sequencing reveals potential biomarkers for colorectal cancer therapy. Frontiers in Microbiology 9:1607\u003c/li\u003e\n\u003cli\u003eSafarchi A, Al-Qadami G, Tran CD, Conlon M (2025) Understanding dysbiosis and resilience in the human gut microbiome: biomarkers, interventions, and challenges. Front Microbiol 16:1559521. https://doi.org/10.3389/fmicb.2025.1559521\u003c/li\u003e\n\u003cli\u003eStavrou G, Kotzampassi K (2017) Gut microbiome, surgical complications and probiotics. Ann Gastroenterol 30(1):45-53. https://doi.org/10.20524/aog.2016.0086\u003c/li\u003e\n\u003cli\u003eZhuang Y-P, Zhou H-L, Chen H-B, Zheng M-Y, Liang Y-W, Gu Y-T, Li W-T, Qiu W-L, Zhou H-G (2023) Gut microbiota interactions with antitumor immunity in colorectal cancer: From understanding to application. \u003c/li\u003e\n\u003cli\u003eHou XY, Zhang P, Du HZ, Gao YQ, Sun RQ, Qin SY, Tian Y, Li J, Zhang YX, Chu WH, Zhang ZJ, Xu FG (2021) Prevotella contributes to individual response of FOLFOX in colon cancer. Clin Transl Med 11(9):e512. https://doi.org/10.1002/ctm2.512\u003c/li\u003e\n\u003cli\u003eFu J, Li G, Li X, Song S, Cheng L, Rui B, Jiang L (2024) Gut commensal Alistipes as a potential pathogenic factor in colorectal cancer. Discov Oncol 15(1):473. https://doi.org/10.1007/s12672-024-01393-3\u003c/li\u003e\n\u003cli\u003eOu S, Wang H, Tao Y, Luo K, Ye J, Ran S, Guan Z, Wang Y, Hu H, Huang R (2022) Fusobacterium nucleatum and colorectal cancer: From phenomenon to mechanism. Front Cell Infect Microbiol 12:1020583. https://doi.org/10.3389/fcimb.2022.1020583\u003c/li\u003e\n\u003cli\u003eDadgar-Zankbar L, Elahi Z, Shariati A, Khaledi A, Razavi S, Khoshbayan A (2024) Exploring the role of Fusobacterium nucleatum in colorectal cancer: Implications for tumor proliferation and chemoresistance. Cell Communication and Signaling 22(1):547\u003c/li\u003e\n\u003cli\u003eWang N, Fang J-Y (2023) Fusobacterium nucleatum, a key pathogenic factor and microbial biomarker for colorectal cancer. Trends in Microbiology 31(2):159-172\u003c/li\u003e\n\u003cli\u003eWu J, Li Q, Fu X (2019) Fusobacterium nucleatum contributes to the carcinogenesis of colorectal cancer by inducing inflammation and suppressing host immunity. Translational oncology 12(6):846-851\u003c/li\u003e\n\u003cli\u003eYuan D, Tao Y, Wang H, Wang J, Cao Y, Cao W, Pan S, Yu Z (2022) A comprehensive analysis of the microbiota composition and host driver gene mutations in colorectal cancer. Investigational New Drugs 40(5):884-894\u003c/li\u003e\n\u003cli\u003eHern\u0026aacute;ndez-Gonz\u0026aacute;lez PI, Barqu\u0026iacute;n J, Ortega-Ferrete A, Pat\u0026oacute;n V, Ponce-Alonso M, Romero-Hern\u0026aacute;ndez B, Oca\u0026ntilde;a J, Caminoa A, Conde-Moreno E, Galeano J, Campo RD, Garc\u0026iacute;a-P\u0026eacute;rez JC (2023) Anastomotic leak in colorectal cancer surgery: Contribution of gut microbiota and prediction approaches. Colorectal Dis 25(11):2187-2197. https://doi.org/10.1111/codi.16733\u003c/li\u003e\n\u003cli\u003ePng CW, Chua YK, Law JH, Zhang Y, Tan KK (2022) Alterations in co-abundant bacteriome in colorectal cancer and its persistence after surgery: a pilot study. Sci Rep 12(1):9829. https://doi.org/10.1038/s41598-022-14203-z\u003c/li\u003e\n\u003cli\u003eLee SY, Park HM, Kim CH, Kim HR (2023) Dysbiosis of gut microbiota during fecal stream diversion in patients with colorectal cancer. Gut Pathog 15(1):40. https://doi.org/10.1186/s13099-023-00566-9\u003c/li\u003e\n\u003cli\u003eIlozumba MN, Gomez MF, Lin T, Himbert C, Round JL, Zac Stephens W, Warby CA, Hardikar S, Li CI, Figueiredo JC (2025) Pre-surgery gut microbial diversity and abundance are associated with post-surgery onset of cachexia in colorectal cancer patients: the ColoCare Study. Cancer Causes \u0026amp; Control:1-18\u003c/li\u003e\n\u003cli\u003eCintoni M, Palombaro M, Zoli E, D\u0026rsquo;Agostino G, Pulcini G, Leonardi E, Raoul P, Rinninella E, De Maio F, Capristo E, Gasbarrini A, Mele MC (2025) The Interplay Between the Gut Microbiota and Colorectal Cancer: A Review of the Literature. Microorganisms 13(6):1410\u003c/li\u003e\n\u003cli\u003eJian H, Liu Y, Wang X, Dong X, Zou X (2023) Akkermansia muciniphila as a Next-Generation Probiotic in Modulating Human Metabolic Homeostasis and Disease Progression: A Role Mediated by Gut-Liver-Brain Axes? Int J Mol Sci 24(4). https://doi.org/10.3390/ijms24043900\u003c/li\u003e\n\u003cli\u003eZhai Q, Feng S, Arjan N, Chen W (2019) A next generation probiotic, Akkermansia muciniphila. Crit Rev Food Sci Nutr 59(19):3227-3236. https://doi.org/10.1080/10408398.2018.1517725\u003c/li\u003e\n\u003cli\u003eJan T, Negi R, Sharma B, Kumar S, Singh S, Rai AK, Shreaz S, Rustagi S, Chaudhary N, Kaur T (2024) Next generation probiotics for human health: An emerging perspective. Heliyon 10(16)\u003c/li\u003e\n\u003cli\u003eLalowski P, Zielińska D (2024) The Most Promising Next-Generation Probiotic Candidates\u0026mdash;Impact on Human Health and Potential Application in Food Technology. Fermentation 10(9):444\u003c/li\u003e\n\u003cli\u003eKoh GY, Kane AV, Wu X, Crott JW (2020) Parabacteroides distasonis attenuates tumorigenesis, modulates inflammatory markers and promotes intestinal barrier integrity in azoxymethane-treated A/J mice. Carcinogenesis 41(7):909-917. https://doi.org/10.1093/carcin/bgaa018\u003c/li\u003e\n\u003cli\u003eYu SY, Xie YH, Qiu YW, Chen YX, Fang JY (2019) Moderate alteration to gut microbiota brought by colorectal adenoma resection. Journal of Gastroenterology and Hepatology 34(10):1758-1765\u003c/li\u003e\n\u003cli\u003eChoi J, Choi YR, Jeong MK, Song HH, Yu JS, Song SH, Park JH, Kim MJ, Park H, Ham YL, Han SH, Kim DJ, Lee DY, Suk KT (2025) Phocaeicola dorei ameliorates progression of steatotic liver disease by regulating bile acid, lipid, inflammation and proliferation. Gut Microbes 17(1):2539448. https://doi.org/10.1080/19490976.2025.2539448\u003c/li\u003e\n\u003cli\u003eXiang J, Chai N, Li L, Hao X, Linghu E (2024) Alterations of Gut Microbiome in Patients with Colorectal Advanced Adenoma by Metagenomic Analyses. The Turkish Journal of Gastroenterology 35(11):859\u003c/li\u003e\n\u003cli\u003eDikeocha IJ, Al-Kabsi AM, Chiu H-T, Alshawsh MA (2022) Faecalibacterium prausnitzii Ameliorates Colorectal Tumorigenesis and Suppresses Proliferation of HCT116 Colorectal Cancer Cells. Biomedicines 10(5):1128\u003c/li\u003e\n\u003cli\u003eXu F, Li Q, Wang S, Dong M, Xiao G, Bai J, Wang J, Sun X (2023) The efficacy of prevention for colon cancer based on the microbiota therapy and the antitumor mechanisms with intervention of dietary Lactobacillus. Microbiology Spectrum 11(5):e00189-00123\u003c/li\u003e\n\u003cli\u003eZhang X, Yu D, Wu D, Gao X, Shao F, Zhao M, Wang J, Ma J, Wang W, Qin X (2023) Tissue-resident Lachnospiraceae family bacteria protect against colorectal carcinogenesis by promoting tumor immune surveillance. Cell host \u0026amp; microbe 31(3):418-432. e418\u003c/li\u003e\n\u003cli\u003eLiu X, Mao B, Gu J, Wu J, Cui S, Wang G, Zhao J, Zhang H, Chen W (2021) Blautia-a new functional genus with potential probiotic properties? Gut Microbes 13(1):1-21. https://doi.org/10.1080/19490976.2021.1875796\u003c/li\u003e\n\u003cli\u003eTogo C, Zidorio AP, Gon\u0026ccedil;alves V, Botelho P, de Carvalho K, Dutra E (2021) Does Probiotic Consumption Enhance Wound Healing? A Systematic Review. Nutrients 14(1). https://doi.org/10.3390/nu14010111\u003c/li\u003e\n\u003cli\u003eSiddharthan R, Chapek M, Warren M, Martindale R (2018) Probiotics in prevention of surgical site infections. Surgical Infections 19(8):781-784\u003c/li\u003e\n\u003cli\u003eLiu PC, Yan YK, Ma YJ, Wang XW, Geng J, Wang MC, Wei FX, Zhang YW, Xu XD, Zhang YC (2017) Probiotics Reduce Postoperative Infections in Patients Undergoing Colorectal Surgery: A Systematic Review and Meta-Analysis. Gastroenterol Res Pract 2017:6029075. https://doi.org/10.1155/2017/6029075\u003c/li\u003e\n\u003cli\u003eLiu PC, Yan YK, Ma YJ, Wang XW, Geng J, Wang MC, Wei FX, Zhang YW, Xu XD, Zhang YC (2017) Probiotics reduce postoperative infections in patients undergoing colorectal surgery: a systematic review and meta‐analysis. Gastroenterology research and practice 2017(1):6029075\u003c/li\u003e\n\u003cli\u003eWang XW, Liu YY (2020) Comparative study of classifiers for human microbiome data. Med Microecol 4. https://doi.org/10.1016/j.medmic.2020.100013\u003c/li\u003e\n\u003cli\u003eTeixeira M, Silva F, Ferreira RM, Pereira T, Figueiredo C, Oliveira HP (2024) A review of machine learning methods for cancer characterization from microbiome data. NPJ Precision Oncology 8(1):123\u003c/li\u003e\n\u003cli\u003eZhou Y, Han W, Feng Y, Wang Y, Liu X, Sun T, Xu J (2025) Revealing gut microbiota biomarkers associated with melanoma immunotherapy response and key bacteria-fungi interaction relationships: evidence from metagenomics, machine learning, and SHAP methodology. Front Immunol 16:1539653. https://doi.org/10.3389/fimmu.2025.1539653\u003c/li\u003e\n\u003cli\u003eMa J, Fang Y, Li S, Zeng L, Chen S, Li Z, Ji G, Yang X, Wu W (2025) Interpretable machine learning algorithms reveal gut microbiome features associated with atopic dermatitis. Frontiers in Immunology 16:1528046\u003c/li\u003e\n\u003cli\u003eNovielli P, Romano D, Magarelli M, Bitonto PD, Diacono D, Chiatante A, Lopalco G, Sabella D, Venerito V, Filannino P (2024) Explainable artificial intelligence for microbiome data analysis in colorectal cancer biomarker identification. Frontiers in Microbiology 15:1348974\u003c/li\u003e\n\u003cli\u003eRynazal R, Fujisawa K, Shiroma H, Salim F, Mizutani S, Shiba S, Yachida S, Yamada T (2023) Leveraging explainable AI for gut microbiome-based colorectal cancer classification. Genome biology 24(1):21\u003c/li\u003e\n\u003cli\u003eTabatabaei SAS, Ghadim HY, Alaei S, Abdolvand F, Mazaheri H, Shamsi F, SarveAhrabi Y, Behrouzi A (2025) The association between the health of the intestines and the human body with Akkermansia muciniphila. The Microbe:100352\u003c/li\u003e\n\u003cli\u003eFaghfuri E, Gholizadeh P (2024) The role of Akkermansia muciniphila in colorectal cancer: A double-edged sword of treatment or disease progression? Biomedicine \u0026amp; Pharmacotherapy 173:116416\u003c/li\u003e\n\u003cli\u003eCronin P, Murphy CL, Barrett M, Ghosh TS, Pellanda P, O\u0026rsquo;Connor EM, Zulquernain SA, Kileen S, McCourt M, Andrews E (2022) Colorectal microbiota after removal of colorectal cancer. NAR cancer 4(2):zcac011\u003c/li\u003e\n\u003cli\u003eEhrenhofer-Murray AE (2025) Queuine: A bacterial nucleobase shaping translation in eukaryotes. Journal of Molecular Biology:168985\u003c/li\u003e\n\u003cli\u003eRamos Y, Sansone S, Morales DK (2021) Sugarcoating it: Enterococcal polysaccharides as key modulators of host-pathogen interactions. PLoS Pathog 17(9):e1009822. https://doi.org/10.1371/journal.ppat.1009822\u003c/li\u003e\n\u003cli\u003eXie S, Ma J, Lu Z (2024) Bacteroides thetaiotaomicron enhances oxidative stress tolerance through rhamnose-dependent mechanisms. Front Microbiol 15:1505218. https://doi.org/10.3389/fmicb.2024.1505218\u003c/li\u003e\n\u003cli\u003eLi D, Wei R, Zhang X, Gong S, Wan M, Wang F, Li J, Chen M, Liu R, Wan Y (2024) Gut commensal metabolite rhamnose promotes macrophages phagocytosis by activating SLC12A4 and protects against sepsis in mice. Acta Pharmaceutica Sinica B 14(7):3068-3085\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":false,"email":"","identity":"current-microbiology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Current Microbiology","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"VoR Journals","inReviewEnabled":false,"inReviewRevisionsEnabled":false},"keywords":"Colorectal cancer, Surgical resection, Postoperative gut microbiota, Commensal restoration, Microbial biomarkers, Explainable AI","lastPublishedDoi":"10.21203/rs.3.rs-7866815/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7866815/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSurgical resection remains the primary treatment for colorectal cancer (CRC), yet its influence on the postoperative gut microbiota remains incompletely understood. In this study, we analyzed the gut microbial communities before and after surgery from our study cohort and integrated findings from four independent CRC datasets to enhance robustness. Our results revealed that post-surgery samples had a reduced microbial diversity but were enriched with commensal taxa, suggesting a potential re-establishment of beneficial microbiota following tumor removal. Leveraging machine learning and Explainable Artificial Intelligence (XAI) through SHapley Additive exPlanations (SHAP), we identified potential postoperative microbial biomarkers, notably \u003cem\u003eAkkermansia\u003c/em\u003e, among the dominant commensal bacteria enriched in post-surgery. Collectively, these findings highlight suggest that surgical resection may promote a favorable shift in gut microbial composition and this could guide targeted microbial modulation to improve postoperative recovery. Our study lays the groundwork for microbiota-informed strategies aimed at improving clinical outcomes in CRC patients after surgery.\u003c/p\u003e","manuscriptTitle":"Deciphering post-surgery gut microbial dynamics in colorectal cancer through multi- cohort machine learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-10 05:35:33","doi":"10.21203/rs.3.rs-7866815/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-15T19:10:43+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-21T16:42:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"90285564436423175324086385650600150243","date":"2025-12-15T09:39:16+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"208697452763933557171469994818556467373","date":"2025-11-12T06:15:44+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-28T14:04:57+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-18T14:39:03+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-16T06:49:29+00:00","index":"","fulltext":""},{"type":"submitted","content":"Current Microbiology","date":"2025-10-15T10:05:11+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":false,"email":"","identity":"current-microbiology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Current Microbiology","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"VoR Journals","inReviewEnabled":false,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c439a978-e6ac-4a74-b4c4-efe45c112da4","owner":[],"postedDate":"November 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-22T07:35:46+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-10 05:35:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7866815","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7866815","identity":"rs-7866815","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}