Gut translocation of antimicrobial resistant pathogens in patients undergoing haematopoietic stem cell transplantation in India

Gut translocation of antimicrobial resistant pathogens in patients undergoing haematopoietic stem cell transplantation in India | 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 Gut translocation of antimicrobial resistant pathogens in patients undergoing haematopoietic stem cell transplantation in India Archana Madhav, Jobin John Jacob, Sushil Selvarajan, Sanika Kulkarni, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7763072/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Patients undergoing haematopoietic stem cell transplantation (HSCT) in low- and middle-income countries face high rates of antimicrobial resistant (AMR) bloodstream infections (BSIs), but the origins and dynamics of these infections remain poorly understood. Results We prospectively studied 81 HSCT patients in India, collecting 252 longitudinal stool samples and applying an enrichment-based metagenomic approach to selectively recover priority AMR pathogens. This strategy enabled high-resolution metagenome-assembled genomes and strain tracking. We identified a marked depletion of gut resistomes during pre-engraftment, followed by a rapid rebound with engraftment, driven by expansion of plasmid-borne carbapenemase and ESBL genes ( blaNDM , blaOXA ). Using whole genome sequencing, Hi-C metagenomics, and strain-level comparisons (> 99.9% ANI), we directly linked gut-colonising organisms to five culture-confirmed AMR-BSI episodes, providing genomic evidence of gut translocation. Patients with BSIs had high mortality (> 40%), underlining the clinical impact of these events. Conclusions We demonstrate that stool-based enrichment metagenomics is a practical and cost-effective approach for non-invasive monitoring of gut recovery and AMR risk after HSCT. Our findings provide the first direct genomic evidence of gut-derived AMR-BSIs in an LMIC cohort, highlighting translocation as a major driver of post-transplant mortality and a critical target for surveillance and intervention. Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Between 15–65% of all haematopoietic stem cell transplantation (HSCT) patients develop bacterial bloodstream infections (BSI), two-thirds of which are associated with mucosal barrier injury (MBI) ( 1 ). In 2020, a study from India reported that ~ 50% of all BSIs in HSCT patients were MBI-associated, termed MBI laboratory-confirmed BSIs ( 2 ). While studies often attribute these episodes to bacterial translocation (BT) from the gut, in critically ill patients, this inference is largely derived via intestinal barrier integrity. BT is a transient occurrence and plays an important role in immune modulation ( 3 , 4 ), however, in immunosuppressed HSCT patients with intestinal barrier dysfunction, this immunomodulatory function of the gut microbiota can contribute to BT and ultimately, bacteraemia ( 5 , 6 ). Gut microbiomes of HSCT patients are enriched with pathobiont Pseudomonadota (Proteobacteria) and Bacillota (Firmicutes) and depleted of commensal short-chain fatty acid-producing Bacteroidota (Bacteroidetes) ( 5 , 7 ). Among the Pseudomonadota, Enterobacteriaceae are the most common aetiologic agents of Gram-negative bacteraemia in HSCT patients ( 8 ), accounting for ~ 75% of culture-positive BSIs post-HSCT in India ( 2 ). Risk of Enterobacteriaceae BSIs is amplified by their ability to accumulate antimicrobial resistance (AMR) determinants, which may be exacerbated by stress induced by intestinal inflammation ( 9 ), with the gut microbiome acting as a resistance gene reservoir ( 10 ). Sardzikova et al. reported that multi-drug resistant (MDR) E. coli , Klebsiella pneumoniae , Enterococcus faecium and Staphylococcus aureus were the leading contributors to the resistome in the HSCT gut ( 11 ). In high income countries (HICs), 9–44% of BSIs in HSCT patients are associated with extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae ( 6 ), with higher mortality rates associated with carbapenem-resistant Enterobacteriaceae BSIs ( 12 ). Low-middle income countries (LMICs) consistently report a higher prevalence of AMR in almost all bacteria, yet syndrome-specific data (e.g., AMR in patients with BSIs post-HSCT) are scarce ( 13 ). The cocktail of immunosuppressants and antimicrobials administered post-HSCT also reduces the sensitivity of culture-based diagnostics of BSI, thus limiting the application of antimicrobial susceptibility testing (AST) ( 14 ). The growing challenge posed by AMR in LMICs ( 15 ), coupled with the limited HSCT patient data in this region, underscores the need for further investigation, where genomic approaches could offer new insights and solutions. Here, we investigated AMR in the gut resistomes of HSCT patients at a tertiary care centre in a LMIC with a high AMR burden and identified episodes of BSI attributable to gastrointestinal BT. We utilised plate-sweep enrichment and shotgun metagenomics on stool samples for targeted characterisation of the HSCT gut resistome, focusing on priority AMR pathogens, with enhanced metagenome-assembled genome (MAG) recovery from stool samples. Methods Study design This study was a prospective, longitudinal study enrolling 81 participants undergoing allogeneic HSCT over a period of 7 months between May and December 2022 at the Department of Haematology, Christian Medical College (CMC), Vellore, Tamil Nadu, India. The study was a collaboration between the Cambridge Institute for Therapeutic Immunology and Infectious Disease (CITIID) and the Clinical Microbiology and Clinical Haematology departments at CMC Vellore. This study was granted ethical approval by the Institutional Review Board (IRB), Research and Ethics committee at CMC Vellore IRB Min No. 14605 dated 27.04.2022. Participant eligibility criteria and clinical data management Patients were eligible if they were admitted to the Department of Haematology at CMC Hospital, scheduled to undergo an allogeneic stem cell transplant and provided informed consent to participate. Informed consent forms were written in English and translated into Hindi and Tamil. Eligible patients (as per clinician’s assessment) or their legal representatives, were asked for informed consent, following which a clinical case report form was completed. Recorded clinical metadata included age, sex, diagnosis, donor age, donor sex, HLA match, type of transplant, preparative regimen, post-HSCT infection episodes, episodes of Clostridioides difficile colitis, GvHD and mucositis, date of absolute neutrophil count (ANC) recovery, total days of hospital stay, and outcome. Additional clinical information was obtained from the clinicians or nurses in-charge, accessed through the clinical workstation, or through discharge/death summaries. Sensitive clinical information was accessed through secure hospital servers and shared only with key personnel. Sample collection and anonymisation Stool samples were collected were collected in sterile stool sample collection containers and transferred to the Department of Clinical Microbiology at 4ºC. Once received, samples were labelled to indicate that it was a study sample and processed within 72 hours. Blood samples were collected and sent to the Microbiology department for culture and AST if a BSI was suspected by the clinician in-charge, for example, if the patient developed a fever post-HSCT. Stool samples were anonymised with a code that represented the patient number, timepoint relative to HSCT, day relative to HSCT, and sample number from that patient. For example, P02AT0353 represents patient 2, after transplant (AT), collected on day + 35, and the third sample from this patient. The blood sample IDs contained an additional code (Blood Adult – BA, Blood Paediatric – BP or Blood Venous – BV). This code was placed between the day relative to HSCT and sample number from that patient. For example, P02AT005BP2 represents a blood sample for patient 2, after transplant, collected on day + 5, blood paediatric, and the second sample from this patient. Stool sample processing Up to 25g (or mL, depending on consistency) of stool was transferred from the stool collection container into sterile 50mL Falcon tubes using Pasteur pipettes. If stool consistency was very hard (Type 1 or type 2 according to the Bristol stool chart classification) ( 33 ), 1g of autoclaved glass beads were added into the 50mL Falcon tube prior to transferring stool from the collection container. Once stool was transferred, the Falcon tube(s) were tightly closed and weighed. From the total weight, the weight of the Falcon tube (13.0g (Abdos) or 13.3g (Tarsons)) was subtracted from the total weight (and 1g in addition to the tube weight was subtracted from the total weight if glass beads were added). Once the stool sample was weighed, sterile PBS was added to the Falcon tubes; the volume added was equivalent to the weight of the stool, such that the stool could be resuspended in a 1:1 ratio. Samples were vortexed until Falcon tube contents were homogenised, and the stool was completely resuspended in PBS. To ensure sample homogenisation, the tube was shaken from time to time to move the glass beads around during vortexing. After vortexing, samples underwent a slow centrifuge at 400 g for 10 mins at 4°C. Serial dilutions and plating of stool samples The 50mL Falcon tubes were retrieved from the centrifuge and 100µl of the supernatant was added into 900µl of sterile PBS to produce a 10x dilution. The 10x Eppendorf was closed tightly and inverted multiple times to ensure complete resuspension of the neat supernatant in the PBS. The 10x dilution was further diluted into a 100x dilution of neat stool. MacConkey agar plates were prepared without crystal violet, so they also grew Gram-positive organisms belonging to the Staphylococcus and Enterococcus genera. Plates were labelled with the sample ID and dilution factor and 50µl of 100x dilution was pipetted onto the plate and spread using a sterile L-spreader. Inoculated MacConkey plates were incubated at 37°C overnight. Stool sample plate sweeps Following overnight incubation, 5mL of sterile PBS was added to each plate. A complete sweep of the plate was collected by resuspending any grown colonies in PBS, aspirating the resuspended culture, and transferring the plate-sweep culture aspirate into clean 15mL Falcon tubes. Then, 1mL of culture was aliquoted into each of three Eppendorf tubes – destined for DNA extraction, formalin inactivation (if eligible) and glycerol stocks (2 tubes) in trypticase soy broth. Blood sample collection and processing Blood culture bottles were incubated in the BacT/ALERT® automated Blood Culture system. Positive samples were passaged either on blood agar or selective media chosen according to the pathogen detected to obtain pure isolates for WGS. Isolates were incubated overnight in nutrient broth at 37ºC, and a 1 mL aliquot of the overnight culture was used for DNA extraction and WGS. Formalin inactivation Stool samples collected at the nearest timepoint to a patient’s culture-confirmed BSI (n = 13) were formalin inactivated for Hi-C metagenomic sequencing. The bacterial suspension was pelleted by centrifugation at 10,000 rpm for 10 minutes at room temperature. Pellets were resuspended in 1X TE buffer using an equal volume to that of the original culture and centrifuged again; this wash step was repeated once more. The bacterial pellet was resuspended in TE buffer containing a final concentration of 2.5% formaldehyde (formalin solution). Samples were incubated at room temperature for 30 min, followed by an incubation at 4°C for 20 min. The formalin fixation reaction was quenched by adding ice-cold glycine added to a final concentration of 0.25M. The sample tubes were inverted a few times to ensure complete resuspension and incubated for 5 minutes at room temperature, followed by a 15-minute incubation on ice. The tubes were then centrifuged for 10 minutes at 10,000 rpm at 4°C. The bacterial pellet was washed twice in 2mL 1X TE and centrifuged in 2mL microcentrifuge tubes. After centrifugation, the supernatant was discarded, and bacterial pellets were frozen at -80°C. DNA extraction DNA extraction was done using the Wizard® Genomic DNA purification kit by Promega as per the manufacturer’s protocol. Briefly, 1mL overnight plate sweep aliquot was centrifuged at 13,000–16,000 g for 2 minutes to pellet the cells. The supernatant was removed and 600 µ l of nuclei lysis solution was added to the pellet and mixed by pipetting. The sample tubes were incubated at 80ºC for 5 minutes to lyse the cells, then cooled down to room temperature. RNAse solution (3 µ l) was added to the cell lysates, which were then mixed by inversion. The samples were incubated at 37ºC for 15–60 minutes and then cooled down to room temperature. 200 µ l of protein precipitation solution was added to the RNAse-treated cell lysates. Tubes were vortexed vigorously at a high speed for 20 seconds to mix the protein precipitation solution with the cell lysate. The samples were incubated on ice for 10 minutes, followed by centrifugation at 13,000–16,000 g for 5 minutes. After centrifugation, the supernatant containing the DNA was transferred to a clean 1.5mL microcentrifuge tube containing 600 µ l of room-temperature isopropanol. The samples were mixed gently by inversion until thread-like strands of DNA formed. The tubes were then centrifuged again at 13,000–16,000 g for 3 minutes. Pellets were air-dried before adding 100 µ l of DNA rehydration solution and incubating at 65ºC for one hour, periodically mixing the solution by gently tapping the tube. The DNA was then stored in the laboratory fridge at 4ºC awaiting shipment to Cambridge. Shipping and sequencing Blood DNA and stool DNA were shipped at 4ºC from CMC to CITIID. DNA concentrations were quantified on a Qubit™ 2.0 Fluorometer, and 50µl of DNA from each sample was transferred into barcoded snap-cap microcentrifuge tubes and shipped to Eurofins Genomics for standard library preparation and sequencing. DNA samples were sequenced on an Illumina NovaSeq 6000 S4 PE150 instrument, with a minimum of 10 million paired-end reads requested for metagenomic sequencing of stool samples and 100x coverage for WGS for genomes up to 10Mb. The sequencing reads for both stool and blood samples from the study can be found under BioProject accession PRJNA1072756. A total of 308 stool samples were collected. Of these, 22 patients were excluded from the study because they underwent autologous HSCTs or had undergone transplantation prior to the study start date. Two-hundred and ninety-eight (298) samples were sent for whole-metagenome shotgun sequencing at Eurofins Genomics. Twenty samples were not sequenced since they contained insufficient DNA to generate libraries that met the QC requirements for successful sequencing, containing less than 1nM/L of DNA. In addition, the stool sample collection sent for sequencing also contained samples belonging to patients for whom we retrieved incomplete metadata, or those who were excluded for reasons described in the patient eligibility section above. Once their corresponding stool samples were removed, the complete stool sample collection comprised a total of 252 samples from the 81 patients who were eligible for the study. For metagenomic Hi-C, 13 formalin-inactivated stool samples were sent to Phase Genomics in Seattle, USA for Hi-C library prep and sequencing according to the ProxiMeta ™ Hi-C Kit Protocol v4.5. Contigs assembled from metagenomic read data were sent to the bioinformatics team at Phase Genomics to overlay Hi-C connections and interpret the data. Two stool samples (P14AT0065 and P51AT1247) did not pass QC for Hi-C. Quality control and assembly of sequencing reads Raw sequencing reads from the enriched stool samples were run through the nf-core/mag pipeline v2.3.0 ( 34 ). Short reads underwent trimming and adapter removal to improve read quality, before removing PhiX reads using Bowtie2. All reads were then passed onto FastQC for final quality control. Metagenomic read assembly was performed using SPAdes v3.15.4, with gene prediction carried out by Prodigal ( 35 ). All assembled contigs underwent QC using QUAST ( 36 ). Although the nf-core/mag pipeline had downstream binning functions, we used Metawrap v1.3.0 ( 31 ) to produce consensus bins from three popular binning pipelines MetaBAT2 v2.12.1 ( 37 ), Maxbin2 v2.2.7 ( 38 ) and CONCOCT v1.1.0 ( 39 ). The threshold for consensus bins was set to a minimum of 70% completeness and < 5% contamination to extract bins of high purity. After extracting consensus bins, each bin was passed through GTDB-TK ( 40 ) for species identification. Taxonomic assignment Taxonomic assignment was performed directly on raw stool sample reads using Kraken v2.1.2 ( 41 ), followed by Bayesian re-estimation at genus level using Bracken v2.7.0 ( 42 ). Since we were using enriched samples, we found negligible levels of human DNA in our samples, however, residual human OTUs were removed with the “extract_kraken_reads.py” script. The Bracken outputs were combined using the “combine_kraken_reports.py” script and converted into a .biom format using kraken-biom v1.2.0 from the KrakenTools software suite ( 43 ). OTU tables were imported into RStudio v4.2.3 and visualisations for total taxonomic composition of the stool samples in the study collection were carried out using metacoder v0.3.6 ( 44 ). OTU IDs were merged with a dataframe from NCBI containing information for that ID at each taxonomic level to produce a data frame containing OTU IDs, abundance data and taxonomy information. This table was converted into a taxmap object which is specialised for community abundance data, using metacoder’s parse_tax_data function, by specifying which columns contain taxonomy information, the order they appear in, and whether the column headers are named by the taxonomic rank. Microbial diversity Beta diversity was calculated using the Bray-Curtis dissimilarity index with the vegdist function from the vegan package. Beta diversity was plotted with the non-metric multidimensional scaling (NMDS) technique, computed using the metaMDS function optimised at 999 iterations with 3 dimensions which had the lowest stress value at 0.07 (< 0.099 represents a ‘good’ fit ( 45 )). The NMDS scores were then extracted into a new data frame, which was merged with sample metadata and plotted. Analysis of similarities (ANOSIM) was calculated using the anosim function from the vegan package and the pairwise adonis test was performed using the pairwiseAdonis v0.4.1 ( 46 ) package on R. Antimicrobial resistance gene detection AMR genes were detected from whole-genome and metagenomic sequences using ABRIcate v1.0.1 using the CARD database as reference. The ABRIcate pipeline was run with default parameters under the assumption that enriched organisms would be sequenced at sufficient coverage to reliably detect AMR genes in them. AMR gene alpha/beta diversities were also calculated to probe for longitudinal variations. Antimicrobial drug classes were reclassified as ‘multidrug’ if the AMR gene detected conferred resistance to \(\:\ge\:\) 3 antimicrobial classes. Antimicrobial sensitivity testing Phenotypic antimicrobial sensitivity testing was performed on bacteria isolated from blood and stool using disk diffusion tests against 8 antimicrobials for cultured Gram-negative organisms and 4 antimicrobials for Gram-positive organisms. 6mm disks of Whatmann No. 1 filter paper were saturated with antimicrobial solutions. Gram-negative isolates were tested against Cefotaxime and Ceftazidime (third generation cephalosporins), Cefoperazone-Sulbactam (third generation cephalosporin with sulphonamide), Cefepime (fourth generation cephalosporin), Amikacin and Gentamicin (aminoglycosides), Ertapenem and Meropenem (carbapenems). Gram-positive organisms were tested against Linezolid (oxazolidinone), Ampicillin (penicillin), high-level Gentamicin (HLG; aminoglycoside) and Vancomycin (glycopeptide). Random forest classification Using the randomForest package v4.7-1.1 ( 47 ) on RStudio, the random forest was computed by classifying 1000 bootstrapped datasets on 500 decision trees with 3 variables considered at each split. The estimated out-of-bag (OOB) error rate was 9.9%, suggesting that 90.1% of the samples were classified correctly. The OOB error rates were then plotted against the number of decision trees to identify the optimal number of trees (~ 150) required to obtain the lowest OOB error rate. To select the optimal number of variables to consider at each node of the decision trees we used an iterative for loop to test the random forest using a different number of variables (i.e., 1–10) from the original dataset at each step in the decision trees. In this case, the lowest OOB error rate was identical when using 2, 3 or 4 variables at each step, so we selected the default value of sampling 3 variables. This random forest was then used to draw a multidimensional scaling (MDS) plot. InStrain analysis InStrain v1.8.0 ( 32 ) was used in this study to detect strains shared by multiple samples. Since InStrain was developed for WMS data, the program was run with default settings so that strain-level information could be detected from the enriched sequencing data with high confidence. For this analysis, the consensus genome bins extracted from the metaWRAP pipeline were used as reference genomes. Fasta headers in the consolidated bin files were modified to include the sample ID, such that each contig would have a unique identifier when concatenating to produce a single fasta file containing all genome bins extracted from the enriched stool sample collection. Concurrently, a scaffold-to-bin (.stb) file was produced using the ‘parse_stb.py’ script (packaged with InStrain) to associate contigs in the combined assembly file to their original genome bin files. Then, we built a bowtie index for the combined fasta file containing extracted genome bins. Next, all the quality-checked metagenomic and whole-genome short read sequencing data were mapped to the bowtie index to identify regions where reads aligned to the index. InStrain Profile was run on on each resulting .bam file, using the --database-mode flag. The InStrain profile command filters .bam files to only retain reads that map with sufficient quality and > 5x coverage, hence minimising mismatches and extracting unique mappings. The reads that pass these filters are considered for microdiversity analyses that calculate nucleotide diversity at each base, in addition to scaffold-level properties including coverage, and ANI between the reads and the reference genomes. Finally, the inStrain compare function was used to produce outputs containing scaffold- (comparisonsTable.tsv) and genome-level (genomeWide_compare.tsv) comparisons of the InStrain profile objects created for each sample in the step above. Pairwise comparisons in the genomewide comparisons table underwent hierarchal clustering to generate strain-level clusters. This output was merged with the GTDB-TK species-level classification output to combine species information with strain-level data. Microbial network analysis The genome-level output from the InStrain compare function was used to produce a microbial network, to illustrate shared strains between patient samples The data frame containing genome-level comparisons was filtered to include only genomes that mapped with \(\:\ge\:\) 99.999% ANI with \(\:\ge\:\) 25% of the metagenomic reads compared to the reference index. This step reduced the genome dataset from containing 44,066 to 1,353 genome-to-genome relationships. This data frame was subset to include only columns containing the two samples and the consensus bin to which they mapped. The consensus bins were then merged with species-level taxonomy information from GTDB-TK. The microbial network was created using the igraph v1.5.1 ( 48 ), tidygraph v1.2.3 ( 49 ) and ggraph v2.1.0 ( 50 ) packages on RStudio to produce a directed acyclic table graph object. Each node ( n = 98) represented a sample (blood or stool) and edges between nodes denote connections ( n = 174) between samples. A column called ‘degree’ was added to the node table, which represents the number of connections to each sample. This was then used to produce a network plot using ggraph under the ‘kk’ layout. Results Cohort summary and outcomes A summary of the baseline clinical characteristics of the allogeneic HSCT cohort can be found in Table 1 . Overall, 32/81 patients developed a clinically significant bacterial infection (38 episodes total) post-HSCT. Clostridioides difficile colitis was the most common infection post-HSCT ( n = 11), followed by Enterobacteriaceae ( n = 7) and coagulase negative Staphylococcus (CONS) ( n = 5). Time taken for neutrophil recovery was the strongest predictor of discharge post-HSCT (Extended data (ED) 1). Table 1 Baseline characteristics of the HSCT study cohort Baseline characteristics n (%) Age (years) < 18 38 (46.9%) 18–30 13 (16.0%) 31–40 13 (16.0%) 41–50 11 (13.6%) 51–60 6 (7.4%) Sex (% male) 53 (65.4%) Diagnosis Malignant 39 (48.1% ) Acute lymphoblastic leukaemia 13 (33.3%) Acute myeloid leukaemia 9 (23.1%) Myelodysplastic syndrome 8 (20.5%) Chronic myeloid leukaemia 5 (12.8%) Hodgkin's lymphoma 2 (5.1%) Polycythemia vera 1 (2.6%) NK T-cell lymphoma 1 (2.6%) Non-malignant 42 (51.9%) Thalassemia major 17 (40.5%) Severe aplastic anaemia 15 (35.7%) Fanconi anaemia 4 (9.5%) Sickle cell anaemia 1 (2.4%) Wiskott Aldrich syndrome 1 (2.4%) Mucopolysaccharidosis 1 (2.4%) Chronic granulomatous disease 1 (2.4%) Severe congenital neutropenia 1 (2.4%) Paroxysmal nocturnal haemoglobinuria 1 (2.4%) Type of allogeneic transplant Fully matched 51 (63.0%) Related donor 40 (49.4%) Unrelated donor 11 (13.6%) Haploidentical 30 (37.0%) Conditioning regimen Myeloablative 58 (71.6%) Reduced intensity 23 (28.4%) Outcome Discharged 66 (81.5%) Died 15 (18.5%) There were 10 episodes of culture-positive BSI prior to HSCT among seven patients (resolved before transplantation). Excluding catheter and central-line associated BSI, 21.0% (17/81) of patients experienced bacterial BSIs post-HSCT (20 total). The median time to culture-positive BSI was 12 days post-HSCT. We retrieved 19 bacterial BSI isolates, which were whole genome sequenced (WGS; 7/10 pre- and 12/20 post-HSCT). Gram-negative bacilli (GNB) caused 65.0% (13/20) of post-transplant BSI episodes. The mortality rate among patients who developed BSI post-HSCT was 41.2% (7/17), whereas the mortality rate among patients who developed any infectious complications post-HSCT was 37.5% (12/32). ED2 contains a summary of all clinically significant organisms detected post-HSCT. Bacterial taxa isolated from stool via plate-sweeps A median of three stool samples were collected from each patient (range 1–8), with at least three stool samples collected for 61/81 patients. Samples from the remaining 20 patients were included for AMR gene profiling. Stool samples (n = 252) were categorised into five groups based on the timepoint relative to HSCT: pre-conditioning (before day − 8; n = 12), conditioning week (day − 8 to day 0; n = 78), pre-engraftment (day + 1 to day + 15; n = 89), early post-engraftment (day + 16 to day + 100; n = 67) and late post-engraftment (day + 101 to day + 200; n = 6). 100x dilutions of homogenised neat stool were plated on MacConkey agar, incubated overnight; a plate sweep and DNA extraction was performed the following day. Figure 1 shows an overview of the bacterial taxa that were enriched in the MacConkey plate sweeps. The omission of crystal violet permitted the growth of Staphylococcus spp. and Enterococcus spp. and targeted Gram-negative organisms (i.e., E. coli , Klebsiella pneumoniae , Acinetobacter baumannii and Pseudomonas aeruginosa ). This approach selected for specific bacterial species commonly associated with AMR, denoted by the focused colour dispersion within specific branches of Pseudomonadota (Proteobacteria; Fig. 1 A) and Bacillota (Firmicutes; Fig. 1 B). The 252 stool samples generated between 1,762,443–33,901,474 reads per sample, with Kraken2 detecting 3,493,496,303 individual operational taxonomic units (OTUs) spanning 4,272 unique taxonomic levels. This dataset was rarefied at the 10th percentile, representing approximately a third of the original dataset with 1,652,408,590 OTUs. Over three-quarters (77.6%; n = 1,282,928,757) of the OTUs belonged to the Enterobacteriaceae family in the Pseudomonadota (Proteobacteria) phylum. At genus level, this family mainly consisted of Escherichia (73.4%), Klebsiella (21.8%), Shigella (1.54%), Citrobacter (1.38%) and Enterobacter (1.34%). In the Bacillota (Firmicutes) phylum ( n = 217,106,676), the Lactobacillales and Bacillales comprised 89.6% ( n = 194,435,721) and 10.4% ( n = 22,640,084) of OTUs, respectively. The Lactobacillales order was dominated by Enterococcus spp. (99.8% relative abundance; n = 193,964,108), and the genus Staphylococcus was the most abundant (99.6% relative abundance; n = 22,550,143) in the Bacillales. Taxonomic diversity of enriched stool from HSCT patients A Shannon alpha diversity analysis of enriched stool samples at species-level did not show any significant difference (one-way ANOVA) when grouped by timepoints relative to HSCT (ED3). Beta diversity was calculated using the Bray-Curtis dissimilarity index (Fig. 2 A) and demonstrated clustering of conditioning week samples (orange), with weaker grouping of samples during pre-engraftment (blue). While differences in beta diversity within grouped timepoints were not significant, there were significant differences between stool collection timepoints (ANOSIM R statistic = 0.049; ** p = 0.004; ED4). Specifically, there was a significant difference (pairwise adonis *** p < 0.001; ED4) in beta-diversity when comparing samples collected during the conditioning week to the pre-engraftment, early-, and late post-engraftment timepoints; genera contributing to these differences are shown in ED5. No significant associations were observed when assessing beta diversity in relation to transplant outcomes or graft-versus-host disease (GvHD). However, patients who developed acute GvHD post-HSCT demonstrated significant log 2 fold-increases in Enterococcus (*** p < 0.001) and Morganella (*** p < 0.001) (Fig. 2 B) genera. The role of antimicrobial resistance genes in the HSCT cohort ABRIcate was used to detect AMR genes within the enriched microbiome samples, using the CARD database. 16,903 unique AMR genes were detected among the 252 enriched stool samples, with efflux-mediating genes representing 65.8% ( n = 11,118) of AMR genes detected within the collection. Non-efflux AMR genes (n = 5,785) constituted the remainder. The median number of AMR genes detected in the pre-conditioning phase ( n = 12) and conditioning week ( n = 78) were 62 and 68 AMR genes per sample, respectively, compared to a median of 55 AMR genes detected per sample during the pre-engraftment phase ( n = 89). Correspondingly, a median Shannon alpha diversity index of 3.23 was observed for AMR genes during both the pre-conditioning and conditioning phases (Fig. 3 A). At the pre-engraftment timepoint, the Shannon index dropped to 3.06. A decrease in the abundance and diversity of AMR genes between the conditioning week and pre-engraftment timepoints were both significant at *** p < 0.001 when tested with a pairwise t-test with Bonferroni correction (ED6). At early post-engraftment, there was a significant increase in the abundance (** p < 0.01) and diversity (*** p < 0.001) of AMR genes compared to the pre-engraftment timepoint. The median number of AMR genes detected at early post-engraftment was 69 (n = 67); alpha diversity = 3.32) and 96 at late post-engraftment (n = 6; alpha diversity = 3.46). Non-efflux AMR gene diversity mirrored the entire AMR gene set (Fig. 3 B), with a significant decrease (* p < 0.05) between the conditioning week (median Shannon index = 2.73) and pre-engraftment (median Shannon index = 2.66), and a significant increase (*** p < 0.001) between pre-engraftment and early post-engraftment (median = 2.89). The abundance of non-efflux AMR genes was significantly higher at early (** p < 0.01) and late (* p < 0.05) post-engraftment compared to pre-engraftment, which suggests that organisms carrying non-efflux AMR genes remain in the HSCT resistome post-transplantation. This observation may indicate preferential selection through robust tolerance and survival mechanisms among organisms carrying non-efflux AMR genes, albeit only six samples were collected at late post-engraftment. However, the increase in non-efflux AMR genes between pre-engraftment and late post-engraftment remained significant (* p < 0.05) when we subsampled the resistome data to only include the six participants for whom we had data. The abundance of five MDR-conferring ESBL/carbapenemase genes ( n = 695) – blaOXA , blaNDM , blaCTX-M , blaSHV and blaTEM was then plotted longitudinally against the HSCT timeline in Fig. 3 C. At least 1/5 MDR genes were detected in 216/252 stool samples. Overall, we observed a significant successive increase in the abundance of MDR genes when comparing samples collected during the early post engraftment timepoint to MDR genes detected in samples during the two preceding conditioning week (*** p < 0.001) and the pre-engraftment (** p < 0.01) timepoints. Notably, the blaOXA and blaNDM carbapenemase genes exhibited progressive increases in their mean abundance per sample between the conditioning week (0.60 and 0.23, respectively) and late post-engraftment (1.40 and 0.8, respectively). Using Hi-C metagenomics, we found that blaOXA and blaNDM were often co-localised on plasmids harboured by E.coli and Klebsiella pneumoniae . Such plasmids were detected in half of all patients who developed culture positive BSIs (Fig. 3 D, ED7). These plasmids also carried genes inducing resistance against tetracyclines, trimethoprim, sulphonamides, and macrolides. AST data for select enteric MDR bacteria ( E. coli , Klebsiella pneumoniae and Enterococcus faecium ) isolated from stool ( n = 90) and enriched stool resistomes are depicted in ED8. MDR isolates were detected in 57 samples from 41 patients, signifying that > 50% of patients had at least one MDR organism isolated from their stool during the HSCT timeline. Spatiotemporal isolate tracking within and between participants Plate-sweep metagenomics offers a factor-increase in the resolution of MAGs recovered from stool samples. This approach enabled us to probe microdiversity within enriched taxa to identify organisms shared across multiple patients within the cohort, and match bloodstream isolates from clinical infection episodes to gut microbiome samples. Figure 4 shows a microbial network which depicts (at isolate level) bacterial taxa detected across multiple stool samples. Consensus bins (> 70% complete, < 5% contaminated) were extracted from the enriched metagenomic sequences using the metaWRAP pipeline and indexed, following which raw sequencing reads from enriched stool metagenomes were mapped against the indexed reference bins. To detect the same isolate among multiple participants, or across multiple timepoints within the same participant, we used a stringent average nucleotide identity (ANI) threshold of 99.999% at the population-level (popANI), where at least 25% of the reads were compared against the reference bins. The resulting microbial network consisted of 98 nodes and 174 connections; the degree annotation represents the number of edges that connect to each node (Fig. 4 ). The network depicts clusters where metagenomic reads from the enriched gut microbiome samples mapped with \(\:\ge\:\) 99.999% popANI to reconstructed MAG reference bins, indicating highly similar isolates detected across multiple samples. There were several episodes (participants 3, 11, 46, 54, 57, and 61) where isolates could be linked longitudinally within the same participant, and one participant (participant 36) who had MAGs from Escherichia coli and Providencia alcalifaciens (> 85% complete, <5% contaminated) that could be linked across timepoints. Transmission links where two edges were shared between samples indicate episodes where metagenomic reads from both samples mapped to reconstructed MAGs from both samples at the \(\:\ge\:\) 99.999% ANI and \(\:\ge\:\) 25% ‘genome compared’ thresholds. Cases where these bidirectional relationships appeared across multiple participants likely indicate localised transmission clusters. Antimicrobial resistant bacterial translocation from the gut BSI bacteria were subjected to WGS. To link bloodstream isolates to gut colonisation, we indexed > 95% complete and < 5% contaminated consensus bins from the metaWRAP pipeline, which fell under the ‘high quality’ draft category according to MIMAG standards ( 16 ). For translocation, the minimum threshold for isolate-level detection was set at 99.9% popANI to account for variation introduced by differences in the sampling and sequencing techniques between blood and stool samples, with a breadth threshold set at 50% to minimise false positives (Table 2 ). Table 2 Summary of bloodstream isolates from BSI episodes which mapped to the corresponding participant’s gut microbiome sample. Bloodstream isolate Genome* Completeness Contamination Species Breadth Contain popANI P02AT005BP2 P02BT0031_bin.2 95.37 0.04 Escherichia coli 0.63 0.9999 0.9999 P05AT014BA1 P05AT0063_bin.2 99.25 0.19 Enterococcus faecium 0.64 0.9997 0.9998 P14AT043BA1 P14BT0061_bin.1 97.00 0.81 Klebsiella pneumoniae 0.77 0.9998 0.9998 P44AT012BA1 P44AT0283_bin.2 99.68 0.62 Pseudomonas aeruginosa 0.99 0.9999 0.9999 P55BT045BA1 P55AT0554_bin.2 98.80 1.14 Klebsiella pneumoniae 0.54 0.9999 0.9999 *Bloodstream isolates underwent WGS where reads mapped to reconstructed MAGs (> 97% completeness, 99.9% ANI across > 50% breadth of the reconstructed MAG. The cultured organism from the blood samples was matched to gut microbiomes in five patients at isolate-level – patient 2 ( E. coli ), 5 ( Enterococcus faecium ), 14 ( Klebsiella pneumoniae ), 44 ( Pseudomonas aeruginosa ) and 55 ( Klebsiella pneumoniae ) (Table 2 ). Patients 2, 5, 14 and 44 had episodes of culture-positive BSI post-HSCT, whereas patient 55 had culture-positive BSI 45 days pre-HSCT. In patients 2, 5, and 14, the BC sample was matched to stool samples collected prior to BSI, and in patients 44 and 55, the BSI isolate was matched to stool samples collected 16 and 100 days after BSI, respectively. In patient 55, the BC and AST report identified the organism as carbapenem-resistant Klebsiella pneumoniae. After HSCT, patient 55 developed febrile neutropenia with septic shock and grade III mucositis, but empirical treatment with meropenem/colistin/teicoplanin likely contributed to subsequent sterile blood cultures. This observation suggests that while the clinical infection in patient 55 occurred 45 days before HSCT and was resolved prior to HSCT, the organism persisted within the host in the gut. While BT was likely in this case, it is difficult to ascertain in the absence of positive blood cultures being detected post-HSCT. Nonetheless, all five BSI isolates were genotypically MDR and 4/5 were phenotypically MDR (ED9). The inStrain profile data showed high ANI between the bloodstream isolates’ genomes and reconstructed MAGs from the corresponding patients’ gut microbiome samples. Discussion Here, we present a prospective longitudinal study tracking AMR-carrying organisms, the gut resistome, and clinical infections of gut-origin in HSCT patients at a single tertiary care centre in India. Contrary to previous studies which used indirect measurements of gut permeability, invasive culture-based methods, or inferred BT through relative abundance of pathogens in the gut, our methodology was designed to capture priority AMR bacteria in the HSCT resistome associated with post-transplant BSI. This approach builds on a previous retrospective investigation by Korula et al. ( 17 ), where gut translocation was inferred by matching organisms and AST results between resistant isolates recovered from blood and surveillance stool cultures. We incorporated genomic methods and used selective enrichment to interrogate these observations prospectively at the same hospital. Overall, 39.5% of all HSCTs performed were complicated by an infection. Clostridioides difficile colitis was the most common infectious complication caused by a single bacterial pathogen post-HSCT, with an incidence of 13.6% which is on the lower end of previously reported estimates ( 2 ). Over the past two decades, infectious aetiologies post-HSCT have transitioned from Gram-positive cocci such as Staphylococci and Enterococci as the primary causative pathogens, to GNB dominating in recent years ( 18 , 19 ). In agreement, we found a higher incidence of post-HSCT BSIs caused by GNB, predominantly Klebsiella pneumoniae , E. coli and Pseudomonas aeruginosa . While the 21.0% incidence of post-HSCT BSI we observed in our cohort lies within previously reported ranges of 10–38.6% ( 14 , 20 – 22 ), the mortality rate was > 40% in BSI patients. Beta diversity of enriched stool samples showed that HSCT patients’ gut microbiomes experience similar alterations across the cohort based on the timepoint relative to HSCT. These data suggest that the preparative regimen patients undergo during the conditioning week and the antimicrobials administered in the pre-engraftment period modulate the microbiota comparably, regardless of indication for HSCT or transplant type. Wong et al. also reported analogous longitudinal shifts in faecal microbiome beta-diversities, using 16S data from an Asian cohort undergoing autologous HSCT ( 23 ). Moreover, Wong et al. described that gut microbial diversity recovered to preconditioning levels within 6-months post-HSCT, which was beyond our follow-up timeframe. In 2018, Tamburini et al . ( 24 ) retrospectively confirmed the translocation of 15/32 (47%) unique organisms detected in blood cultures from 30 patients, observed at ≥ 0.1% relative abundance in the gut in a high-income setting. We introduced a novel approach to monitor translocation of gut pathogens, especially ESBL Enterobacterales, with our enrichment-based strategy supported by Jazmati et al. ( 25 ). BC positivity is low among HSCT patients, since the antimicrobials administered to counter febrile neutropenia in the pre-engraftment phase further reduce the low sensitivity of BCs when used in critically ill patients ( 26 ). Despite this, we found that over a quarter of sequenced culture-positive BSIs (5/19, 26.3%) were caused by isolates that were also in the gut, and one probable gut-origin BSI episode caused by Pseudomonas aeruginosa . All five patients who had BSIs of gut-origin also showed evidence of enteric mucositis, which has been linked with infections by Gram negative Enterobacteriaceae , Pseudomonas aeruginosa and Gram-positive Enterococci ( 27 ). We were also able to track highly similar isolates longitudinally within and between patients. Ghosh et al. recently reported that the high prevalence of MDR in an Indian HSCT cohort was primarily associated with the blaNDM and blaOXA - 48-like carbapenemase genes ( 22 ), but only with pre-transplant surveillance samples. With the longitudinal sampling strategy, we found many blaNDM and blaOXA genes in our cohort, and this abundance significantly increased in the early post-engraftment timepoint when compared to the two preceding timepoints. In a longitudinal study with eight paediatric patients in Italy, D’Amico et al. ( 28 ) also had similar observations, reporting an AMR gene profile that had diversified with MDR genes in addition to consolidating the resistome present prior to HSCT. Stecher et al. reported that Enterobacteriaceae bloom under conditions of intestinal inflammation, which also exacerbates HGT ( 9 ). This observation highlights the critical role of the gut as a reservoir for transmissible AMR genes which can transfer on mobile genetic elements in immunocompromised hosts who frequently visit hospitals. While whole-metagenomic sequencing (WMS) hinders the association of AMR genes back to their host species, using an enrichment-based approach reduces the pool of potential gene hosts, and permits screening of specific bacterial populations compared to isolation-based approaches. Inadvertent selection of pre-existing bacteria harbouring resistance genotypes can also accelerate AMR development since this mechanism is associated with a lower fitness cost compared to the establishment of de-novo mutations conferring resistance ( 29 ). Therefore, routine surveillance of gut resistome plasticity using stool samples could indicate the most likely infectious aetiologies of AMR-BSI and help optimise treatment regimens to minimise inadvertent selection for resistance among HSCT patients. The onset of febrile neutropenia in the HSCT cohort warrants the administration of empirical antimicrobials considering the immunosuppressed state of the recipient and the exposure to a high-risk hospital environment. These broad-spectrum antimicrobials should immediately eliminate organisms that may harbour but not express AMR genes ( 30 ). Upon successful engraftment and neutrophil recovery, neutropenic fever subsides among HSCT patients, and antimicrobial regimens are tapered in response. We hypothesise that the removal of this selection pressure permits residual resistant members of the microbiome to proliferate and offers a window for nosocomial AMR pathogens to occupy the niche. This could explain the significant decrease in the abundance and diversity of AMR genes between the conditioning week and pre-engraftment, and the significant increase between pre-engraftment and early post-engraftment. This study had limitations. First, this study did not consider the clinical heterogeneity of HSCT patients. Future investigations should focus on more specific cohorts based on their age, indications for HSCT or a defined outcome, e.g., gastrointestinal-GvHD. Attempting these facets was beyond the scope of our study timeframe. Secondly, we cannot directly compare enrichment data with WMS data from neat stool samples, and selective media can potentially lead to uneven enrichment based on organisms’ differential capacities to metabolise available nutrients. However, since we focused on characterising key AMR-pathogens in the HSCT microbiome, we also avoided biased total microbiota recovery from stored faeces. Finally, while the enrichment strategy outperforms WMS from a neat sample in capturing intraspecific or population-level heterogeneity for the selected organisms, binning algorithms capture the most dominant STs ( 31 ). Nonetheless, we were able to account for the population-level heterogeneity across samples because of inStrain’s unique ‘microdiversity-aware’ metrics of conANI and popANI ( 32 ). This is the first and largest study to track the longitudinal evolution of the gut resistome and AMR-BSIs of gut-origin among HSCT patients in a LMIC with a high AMR burden. We identified that the HSCT gut microbiome is modulated by the pre-transplant conditioning regimen, and that the gut resistome is significantly depleted at the pre-engraftment phase, likely in response to conditioning regimens and empirical antimicrobials. However, the resistome is promptly restored at early post-engraftment and shifts towards a sustained and significant increase in the abundance of transmissible non-efflux AMR genes into the late post-engraftment phase. Additionally, we associated over one quarter of retrieved BSI isolates with the corresponding participant’s gut microbiomes at isolate-level resolution. In conclusion, we have established a baseline for the genomic landscape of gastrointestinal AMR carriage among HSCT patients in South Asia and provide the first direct evidence to suggest pathological BT from the gut using non-invasive sampling and enrichment-based metagenomic sequencing in this high-risk transplant cohort. Declarations Funding and acknowledgements This work was supported by a grant from the Bill and Melinda Gates Foundation (OPP1159351) and a Wellcome Senior Research Fellowship (215515/Z/19/Z) to Stephen Baker. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. We would like to thank Agila Kumari Pragasam, Chaitra Shankar, and Ellen Higginson for helpful discussions on the study design. We are also grateful to the media preparation team in the department of Clinical Microbiology at CMC Vellore, and Plamena Naydenova for her assistance with quantifying DNA concentrations. We would also like to thank Jolynne Mokaya for critical feedback on the manuscript. Finally, we would like to acknowledge the clinical teams for their assistance with sample collection and thank the patients and their families for their participation and support. Ethics approval and consent to participate This study was granted ethical approval by the Institutional Review Board (IRB), Research and Ethics committee at CMC Vellore IRB Min No. 14605 dated 27.04.2022. Patients were eligible if they were admitted to the Department of Haematology at CMC Hospital, scheduled to undergo an allogeneic stem cell transplant and provided written informed consent to participate. Informed consent forms were written in English and translated into Hindi and Tamil. Eligible patients (as per clinician’s assessment) or their legal representatives, were asked for informed consent, following which a clinical case report form was completed. Consent for publication Not applicable Conflicts of Interest The authors report there are no competing interests to declare. Author contributions Conceptualisation and design: AM, SB, BV, SS, BG Collection of samples and clinical data: SS, SD, BG Processing of samples: AM, SK, YM, DM, PS, KM Analysis and interpretation of data: AM, JJJ, SB Drafting initial paper: AM, JJJ, SB Revising the final paper: AM, JJJ, SS, SK, YM DM, PS, KM, BG, BV, SB Data Availability Statement The authors confirm that the data supporting the findings of this study are available in the article and its supplementary materials. The genomic sequencing reads have been deposited in the NCBI Sequence Read Archive under BioProject accession PRJNA1072756 (https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA1072756). References Heston SM, Young RR, Hong H, Akinboyo IC, Tanaka JS, Martin PL, et al. Microbiology of Bloodstream Infections in Children After Hematopoietic Stem Cell Transplantation: A Single-Center Experience Over Two Decades (1997–2017). Open Forum Infect Dis. 2020;7(11):ofaa465. Barman P, Choudhary D, Chopra S, Thukral T. Blood stream infections in hematopoietic stem cell transplant patients: A 2-year study from India. Oncol J India. 2020;4(2):43. Cerf-Bensussan N, Gaboriau-Routhiau V. The immune system and the gut microbiota: friends or foes? Nat Rev Immunol. 2010;10(10):735–44. Brenchley JM, Douek DC. Microbial Translocation Across the GI Tract. Annu Rev Immunol. 2012;30(1):149–73. Shono Y, van den Brink MRM. Gut microbiota injury in allogeneic haematopoietic stem cell transplantation. Nat Rev Cancer. 2018. Satlin MJ, Jenkins SG, Walsh TJ. The Global Challenge of Carbapenem-Resistant Enterobacteriaceae in Transplant Recipients and Patients With Hematologic Malignancies. Clin Infect Dis. 2014;58(9):1274–83. Koh AY. The microbiome in hematopoietic stem cell transplant recipients and cancer patients: Opportunities for clinical advances that reduce infection. Sheppard DC, editor. PLOS Pathog. 2017;13(6):e1006342. Pouch SM, Satlin MJ. Carbapenem-resistant Enterobacteriaceae in special populations: Solid organ transplant recipients, stem cell transplant recipients, and patients with hematologic malignancies. Virulence. 2017. Stecher B, Denzler R, Maier L, Bernet F, Sanders MJ, Pickard DJ, et al. Gut inflammation can boost horizontal gene transfer between pathogenic and commensal Enterobacteriaceae. Proc Natl Acad Sci. 2012;109(4):1269–74. Anthony WE, Burnham CAD, Dantas G, Kwon JH. The Gut Microbiome as a Reservoir for Antimicrobial Resistance. J Infect Dis. 2021;223(Supplement3):S209–13. Sardzikova S, Andrijkova K, Svec P, Beke G, Klucar L, Minarik G, et al. Gut diversity and the resistome as biomarkers of febrile neutropenia outcome in paediatric oncology patients undergoing hematopoietic stem cell transplantation. Sci Rep. 2024;14(1):5504. Satlin MJ, Walsh TJ. Multidrug-resistant Enterobacteriaceae, Pseudomonas aeruginosa , and vancomycin-resistant Enterococcus : Three major threats to hematopoietic stem cell transplant recipients. Transpl Infect Dis. 2017;19(6):e12762. Iskandar K, Molinier L, Hallit S, Sartelli M, Hardcastle TC, Haque M, et al. Surveillance of antimicrobial resistance in low- and middle-income countries: a scattered picture. Antimicrob Resist Infect Control. 2021;10(1):63. Ghazal SS, Stevens MP, Bearman GM, Edmond MB. Utility of surveillance blood cultures in patients undergoing hematopoietic stem cell transplantation. Antimicrob Resist Infect Control. 2014;3:20. Murray CJL, Ikuta KS, Sharara F, Swetschinski L, Robles Aguilar G, Gray A, et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet. 2022;399(10325):629–55. The Genome Standards Consortium, Bowers RM, Kyrpides NC, Stepanauskas R, Harmon-Smith M, Doud D, et al. Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nat Biotechnol. 2017;35(8):725–31. Korula A, Perumalla S, Devasia AJ, Abubacker FN, Lakshmi KM, Abraham A et al. Drug-resistant organisms are common in fecal surveillance cultures, predict bacteremia and correlate with poorer outcomes in patients undergoing allogeneic stem cell transplants. Transpl Infect Dis [Internet]. 2020 Jun [cited 2022 May 23];22(3). Available from: https://onlinelibrary.wiley.com/doi/ 10.1111/tid.13273 Collin BA, Leather HL, Wingard JR, Ramphal R. Evolution, Incidence, and Susceptibility of Bacterial Bloodstream Isolates from 519 Bone Marrow Transplant Patients. Clin Infect Dis. 2001;33(7):947–53. Salas MQ, Charry P, Puerta-Alcalde P, Martínez-Cibrian N, Solano MT, Serrahima A et al. Bacterial Bloodstream Infections in Patients Undergoing Allogeneic Hematopoietic Cell Transplantation With Post-Transplantation Cyclophosphamide. Transplant Cell Ther. 2022;28(12):850.e1-850.e10. Krüger W, Rüssmann B, Kröger N, Salomon C, Ekopf N, Elsner HA, et al. Early infections in patients undergoing bone marrow or blood stem cell transplantation – a 7 year single centre investigation of 409 cases. Bone Marrow Transpl. 1999;23(6):589–97. Ninin E, Milpied N, Moreau P, André-Richet B, Morineau N, Mahé B, et al. Longitudinal Study of Bacterial, Viral, and Fungal Infections in Adult Recipients of Bone Marrow Transplants. Clin Infect Dis. 2001;33(1):41–7. Ghosh S, Bhattacharya S, Goel G, Deshmukh RA, Javed R, Roychowdhury M et al. Hematopoietic stem-cell transplantation in a zoo of multidrug‐resistant organisms: Data from a cancer center in eastern India. Transpl Infect Dis. 2023;e14072. Wong SP, Er YX, Tan SM, Lee SC, Rajasuriar R, Lim YAL. Oral and Gut Microbiota Dysbiosis is Associated with Mucositis Severity in Autologous Hematopoietic Stem Cell Transplantation: Evidence from an Asian Population. Transplant Cell Ther. 2023;29(10):633.e1-633.e13. Tamburini FB, Andermann TM, Tkachenko E, Senchyna F, Banaei N, Bhatt AS. Precision identification of diverse bloodstream pathogens in the gut microbiome. Nat Med. 2018;24(12):1809–14. Jazmati T, Hamprecht A, Jazmati N. Comparison of stool samples and rectal swabs with and without pre-enrichment for the detection of third-generation cephalosporin-resistant Enterobacterales (3GCREB). Eur J Clin Microbiol Infect Dis. 2021;40(11):2431–6. Nieman AE, Savelkoul PHM, Beishuizen A, Henrich B, Lamik B, MacKenzie CR, et al. A prospective multicenter evaluation of direct molecular detection of blood stream infection from a clinical perspective. BMC Infect Dis. 2016;16(1):314. Balletto E, Mikulska M. Bacterial infections in haematopoietic stem cell transplant patients. Mediterr J Hematol Infect Dis. 2015;7:e2015045. D’Amico F, Soverini M, Zama D, Consolandi C, Severgnini M, Prete A, et al. Gut resistome plasticity in pediatric patients undergoing hematopoietic stem cell transplantation. Sci Rep. 2019;9(1):5649. Shepherd MJ, Fu T, Harrington NE, Kottara A, Cagney K, Chalmers JD, et al. Ecological and evolutionary mechanisms driving within-patient emergence of antimicrobial resistance. Nat Rev Microbiol. 2024;22(10):650–65. Heston SM, Young RR, Jenkins K, Martin PL, Stokhuyzen A, Ward DV, et al. The effects of antibiotic exposures on the gut resistome during hematopoietic cell transplantation in children. Gut Microbes. 2024;16(1):2333748. Uritskiy GV, DiRuggiero J, Taylor J. MetaWRAP—a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome. 2018;6(1):158. Olm MR, Crits-Christoph A, Bouma-Gregson K, Firek BA, Morowitz MJ, Banfield JF. inStrain profiles population microdiversity from metagenomic data and sensitively detects shared microbial strains. Nat Biotechnol. 2021;39(6):727–36. Lewis SJ, Heaton KW. Stool Form Scale as a Useful Guide to Intestinal Transit Time. Scand J Gastroenterol. 1997;32(9):920–4. Krakau S, Straub D, Gourlé H, Gabernet G, Nahnsen S. nf-core/mag: a best-practice pipeline for metagenome hybrid assembly and binning. NAR Genomics Bioinforma. 2022;4(1):lqac007. Hyatt D, Chen GL, LoCascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics. 2010;11(1):119. Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013;29(8):1072–5. Kang DD, Li F, Kirton E, Thomas A, Egan R, An H, et al. MetaBAT 2: an adaptive binning algorithm for robust and efficient genome reconstruction from metagenome assemblies. PeerJ. 2019;7:e7359. Wu YW, Simmons BA, Singer SW. MaxBin 2.0: an automated binning algorithm to recover genomes from multiple metagenomic datasets. Bioinformatics. 2016;32(4):605–7. Alneberg J, Bjarnason BS, De Bruijn I, Schirmer M, Quick J, Ijaz UZ, et al. Binning metagenomic contigs by coverage and composition. Nat Methods. 2014;11(11):1144–6. Chaumeil PA, Mussig AJ, Hugenholtz P, Parks DH. GTDB-Tk: a toolkit to classify genomes with the Genome Taxonomy Database. Hancock J, editor. Bioinformatics. 2020;36(6):1925–7. Wood DE, Lu J, Langmead B. Improved metagenomic analysis with Kraken 2. Genome Biol. 2019;20(1):257. Lu J, Breitwieser FP, Thielen P, Salzberg SL. Bracken: estimating species abundance in metagenomics data. PeerJ Comput Sci. 2017;3:e104. Lu J, Rincon N, Wood DE, Breitwieser FP, Pockrandt C, Langmead B, et al. Metagenome analysis using the Kraken software suite. Nat Protoc. 2022;17(12):2815–39. Foster ZSL, Sharpton TJ, Grünwald NJ, Metacoder. An R package for visualization and manipulation of community taxonomic diversity data. Poisot T, editor. PLOS Comput Biol. 2017;13(2):e1005404. Dugard P, Todman J, Staines H. Approaching Multivariate Analysis: A practical introduction [Internet]. 2nd ed. London: Routledge; 2022. 275 p. Available from: https://www.taylorfrancis.com/books/9781003343097 Arbizu PM. pairwiseAdonis: pairwise multilevel comparison using adonis. 2017. 2019. Liaw A, Wiener M. Classification and Regression by randomForest. R News. 2002;2(3):18–22. Csardi MG. Package ‘igraph’. Last Accessed. 2013;3(09):2013. Pedersen TL. tidygraph: A Tidy API for Graph Manipulation [Internet]. 2023. Available from: https://github.com/thomasp85/tidygraph Si B, Liang Y, Zhao J, Zhang Y, Liao X, Jin H, et al. GGraph: An Efficient Structure-Aware Approach for Iterative Graph Processing. IEEE Trans Big Data. 2022;8(5):1182–94. Additional Declarations No competing interests reported. Supplementary Files HSCTExtendeddata.docx Cite Share Download PDF Status: Posted Version 1 posted 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-7763072","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":532377800,"identity":"61d33aef-f148-4883-b5c2-e361cf6d1082","order_by":0,"name":"Archana Madhav","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Archana","middleName":"","lastName":"Madhav","suffix":""},{"id":532377802,"identity":"52d5c1de-fabd-400d-b8ee-ff82fde55c15","order_by":1,"name":"Jobin John Jacob","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Jobin","middleName":"John","lastName":"Jacob","suffix":""},{"id":532377805,"identity":"bb874f71-0158-467f-b882-12ffae2b587c","order_by":2,"name":"Sushil Selvarajan","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Sushil","middleName":"","lastName":"Selvarajan","suffix":""},{"id":532377807,"identity":"0d921712-d69e-464b-b977-f78b14cd5a45","order_by":3,"name":"Sanika Kulkarni","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Sanika","middleName":"","lastName":"Kulkarni","suffix":""},{"id":532377809,"identity":"e5472a11-2ba6-4068-be46-e94af091ba32","order_by":4,"name":"Dhivya Murugan","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Dhivya","middleName":"","lastName":"Murugan","suffix":""},{"id":532377810,"identity":"045ca87e-b276-4df2-a501-92e82b3824e2","order_by":5,"name":"Pavithra Sathiya Narayanan","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Pavithra","middleName":"Sathiya","lastName":"Narayanan","suffix":""},{"id":532377811,"identity":"a45bb660-4b59-453e-a674-95c03640536d","order_by":6,"name":"Yesudoss Muniyandi","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Yesudoss","middleName":"","lastName":"Muniyandi","suffix":""},{"id":532377812,"identity":"f18eaa97-2c58-4b33-bb52-0b536345ef9d","order_by":7,"name":"Kalaiyarasi Mohan","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Kalaiyarasi","middleName":"","lastName":"Mohan","suffix":""},{"id":532377813,"identity":"9a35441d-1925-47e7-b624-9274332d8da7","order_by":8,"name":"Sourav Das","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Sourav","middleName":"","lastName":"Das","suffix":""},{"id":532377814,"identity":"c02113d1-eca4-4e6c-9a99-9d36fc4c6896","order_by":9,"name":"Biju George","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Biju","middleName":"","lastName":"George","suffix":""},{"id":532377815,"identity":"7c33439e-30d3-41d5-a451-6e8116f3e4ad","order_by":10,"name":"Balaji Veeraraghavan","email":"","orcid":"","institution":"Christian Medical College","correspondingAuthor":false,"prefix":"","firstName":"Balaji","middleName":"","lastName":"Veeraraghavan","suffix":""},{"id":532377816,"identity":"edb30775-8e69-4763-9701-18afd8745978","order_by":11,"name":"Stephen Baker","email":"data:image/png;base64,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","orcid":"","institution":"A*STAR","correspondingAuthor":true,"prefix":"","firstName":"Stephen","middleName":"","lastName":"Baker","suffix":""}],"badges":[],"createdAt":"2025-10-02 02:08:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7763072/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7763072/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94049750,"identity":"393bae3b-9fa8-4005-a52e-730f3c6eab80","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2427950,"visible":true,"origin":"","legend":"","description":"","filename":"HSCTmanuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/720752c09821a5bc830f6c0b.docx"},{"id":94050810,"identity":"0d705382-9984-476b-a205-74ac95a8bc5e","added_by":"auto","created_at":"2025-10-21 23:49:10","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3052939,"visible":true,"origin":"","legend":"","description":"","filename":"HSCTFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/aa1b7e26978755750745e5d6.docx"},{"id":94049751,"identity":"d3844645-82c3-49bd-ad77-2d771659486f","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"json","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13044,"visible":true,"origin":"","legend":"","description":"","filename":"4c8dd4cc0b82437b960d7204720d23d8.json","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/a13bcde9457b20e4597a3637.json"},{"id":94049758,"identity":"10895d86-096f-4ab0-bad3-4865e0572f7a","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2660937,"visible":true,"origin":"","legend":"","description":"","filename":"HSCTExtendeddata.docx","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/68cd5e577afb93e3128d561b.docx"},{"id":94049760,"identity":"ae3ecd5e-5654-4e01-9927-c4cc2f1fdd8f","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"xml","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":157562,"visible":true,"origin":"","legend":"","description":"","filename":"4c8dd4cc0b82437b960d7204720d23d81enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/09ee2523862d5ca1be667a7d.xml"},{"id":94049752,"identity":"87abf47a-3212-49c7-a8e5-f3fbdfc9e7b0","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1141662,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/99bbe9f512367d620bd5b288.png"},{"id":94049767,"identity":"5180cde4-a497-4c9a-9a86-da6d72a9dd1a","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":244214,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/2ddef23857dedbae68e72d49.jpeg"},{"id":94050814,"identity":"95e609d6-9683-4059-b3fe-d580699f0a49","added_by":"auto","created_at":"2025-10-21 23:49:10","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":187101,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/f824330cc88a99d153eb290c.png"},{"id":94049754,"identity":"89bb00c3-9d87-4971-b062-2c321c188472","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":432881,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/d800e04086d613f7bbce39c5.png"},{"id":94049763,"identity":"159b5864-bc8d-4ac6-8bbf-9bb007155343","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1141662,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/7b954833e8be1c487a939b1f.png"},{"id":94049757,"identity":"a3adf43b-87c0-4c8f-bd0e-65f3271dc10e","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":423805,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/3a000210908f9610db0bef7d.jpeg"},{"id":94049769,"identity":"8f167be4-38a0-48c2-8902-da247021623c","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":897805,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/2501bc9b8debdb449b386434.png"},{"id":94050811,"identity":"4dd646ef-2d78-4a7b-9d33-49845b8d5df9","added_by":"auto","created_at":"2025-10-21 23:49:10","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":432881,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/9737b085454abd1bf3fc08d8.png"},{"id":94049762,"identity":"bb557544-a8e0-4953-b06f-76397dbb9ad2","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":173993,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/1ef15c8b2f5854cf03002e0d.png"},{"id":94049771,"identity":"f4d42f57-64ef-4a24-8be0-c5f5b61b8646","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":50497,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/d091359ee73ce28395b8144d.png"},{"id":94049766,"identity":"60c21db9-8fe9-45d0-afd2-0486d1c59823","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":67549,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/0cce4786f7f4b8978845b6d3.png"},{"id":94049774,"identity":"c674c0f6-c9db-4dc3-842f-80a04774999a","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42104,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/007bc2ae0465a3bd09afbfea.png"},{"id":94049755,"identity":"5405c85d-f63f-4a2b-86a7-9d9dd23e0824","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":173993,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/0d681539f6f8739b4af4f6ce.png"},{"id":94049776,"identity":"34b5bae4-77eb-4fd0-ae84-7ab3fbeef26f","added_by":"auto","created_at":"2025-10-21 23:41:11","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":88142,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/f5e5fcb5aa19ffdaaf79e4eb.png"},{"id":94049768,"identity":"49e1211d-1b2a-47ad-b846-2ffe5e9b0af7","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":143968,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/b8adda0197e0432a3cd8fcaa.png"},{"id":94049772,"identity":"cde4b32b-0793-4f33-a23f-d5c46c82f6d8","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42104,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/b4b0602afa374bbf71ebd9da.png"},{"id":94050812,"identity":"ff83549d-5fa1-4570-ad64-f4c910d80bf2","added_by":"auto","created_at":"2025-10-21 23:49:10","extension":"xml","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":154128,"visible":true,"origin":"","legend":"","description":"","filename":"4c8dd4cc0b82437b960d7204720d23d81structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/002464c3eca174f9e1d87feb.xml"},{"id":94049775,"identity":"717b2c17-cd72-4b24-b4d1-ea43b7be2015","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"html","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":168678,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/3465e7164be09ea242a79d2e.html"},{"id":94049747,"identity":"229984e4-7532-43bf-b9d8-6bbceb670e93","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":586693,"visible":true,"origin":"","legend":"\u003cp\u003eOverview of Gut Microbial Composition in Enriched HSCT Stool Samples\u003c/p\u003e\n\u003cp\u003eMetacoder plots showing the abundance of reads belonging to various taxa in the HSCT study.\u003c/p\u003e\n\u003cp\u003eFigure depicts an overview of the collective microbial composition of the MacConkey-enriched stool sample collection used in the study. The detected taxids are shown in the node colour with the distribution of reads belonging to each taxon shown in the branch colours. Enlarged Pseudomonadota (Proteobacteria) (A) and Bacillota (Firmicutes) (B) phyla. Plate-sweep selection for bacterial species within two main classes – Gammaproteobacteria (Pseudomonadota), and Bacilli (Bacillota), the latter mainly comprising Staphylococci of the Bacillales order and Enterococci of the Lactobacillales order.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/f63047753a627596c3a355ca.png"},{"id":94050809,"identity":"918186aa-7921-4b27-8655-ffebdf17beba","added_by":"auto","created_at":"2025-10-21 23:49:09","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":126167,"visible":true,"origin":"","legend":"\u003cp\u003eGut Microbiome Community Shifts and Taxa Linked to GI-GvHD\u003c/p\u003e\n\u003cp\u003eNMDS plot depicting beta-diversity of enriched stool samples using the Bray-Curtis dissimilarity index. (A) Plot shows indices coloured by the timepoint when the sample was collected, relative to when the patient’s transplant was performed. Plot shows clustering of samples collected during the conditioning week (orange), and a looser clustering of samples collected during the first two-weeks post stem-cell infusion (blue; pre-engraftment). (B) Volcano plot depicting log2 fold-changes in detected taxa at genus level between patients with and without GI-GvHD at the pre-engraftment timepoint. Taxa in the top right quadrant were detected in significantly higher abundance among patients who went on to develop GI-GvHD, especially the \u003cem\u003eEnterococcus\u003c/em\u003eand \u003cem\u003eMorganella\u003c/em\u003e genera. While there were other genera (e.g., \u003cem\u003eLuteimonas\u003c/em\u003e, \u003cem\u003eCandidatus\u003c/em\u003e \u003cem\u003ehamiltonella\u003c/em\u003e) showing log-fold shifts, these were not considered to be representative shifts relative to the residual microbiome content, since we did not enrich for these organisms.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/8116b254832cdc2f4a3b2073.png"},{"id":94049753,"identity":"946740fd-467b-4641-965c-e10bb2ddf746","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":166404,"visible":true,"origin":"","legend":"\u003cp\u003eShared and Persistent Microbial Strains Across Patients and Timepoints\u003c/p\u003e\n\u003cp\u003eMicrobial network depicting highly similar or identical isolates shared between patients or that could be detected within participants across multiple timepoints. Figure shows enriched gut microbiome metagenomic reads which mapped to reconstructed MAG reference bins with \u0026gt;70% completeness, \u0026lt;5% contamination at 99.999% ANI and \u0026gt;25% of the genome mapped.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/8e75c517f308a2827477cebb.png"},{"id":94049764,"identity":"2216e9c4-6d0a-45d4-a2ba-3bc232df7325","added_by":"auto","created_at":"2025-10-21 23:41:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":271312,"visible":true,"origin":"","legend":"\u003cp\u003eDynamic Shifts in AMR Gene Diversity and Abundance During Transplantation\u003c/p\u003e\n\u003cp\u003eA) Boxplot showing the Shannon alpha-diversity indexes for all AMR genes longitudinally through the transplant timeline. Overall AMR gene diversity saw a significant decrease (\u003cem\u003ep\u003c/em\u003e\u0026lt;0.001***) when comparing samples taken during the conditioning week to those taken in the pre-engraftment period, and a significant increase (\u003cem\u003ep\u003c/em\u003e\u0026lt;0.001***) when comparing pre-engraftment samples to early post-engraftment samples. B) Boxplot showing the Shannon alpha-diversity indexes for non-efflux AMR genes longitudinally through the transplant timeline. Alpha diversity of non-efflux AMR genes showed no significant differences when comparing samples taken during the conditioning week to those taken in the pre-engraftment period, but significant increases were observed when comparing pre-engraftment samples to those collected at successive early post-engraftment (**\u003cem\u003ep\u003c/em\u003e\u0026lt;0.01) and late post-engraftment (*\u003cem\u003ep\u003c/em\u003e\u0026lt;0.05) timepoints. C) Longitudinal variations in the abundance of 5 non-efflux MDR-conferring AMR genes of clinical relevance detected per sample. There was a significant increase in ESBL gene abundance at the early post-engraftment timepoint compared to the conditioning week and pre-engraftment timepoints. There were progressive increases in the number of \u003cem\u003endm\u003c/em\u003eand \u003cem\u003eoxa\u003c/em\u003e genes detected per sample when comparing conditioning week samples to pre-engraftment and early post-engraftment samples. D) Hi-C metagenomic sequencing was performed on a subset of 11 enriched stool samples from patients with clinically significant infection episodes. Hi-C helped delineate AMR genes that co-localised on plasmids. The heatmap shows a presence/absence matrix of AMR genes found in plasmids from each of the 11 enriched stool samples (one sample did not pass QC for Hi-C).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/a7216dbc4bf227db6d5860a4.png"},{"id":97141550,"identity":"69699556-6496-4507-b986-585f8c61de08","added_by":"auto","created_at":"2025-12-01 10:06:47","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2215801,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/efbbab8e-5597-46b8-929c-37aff3067143.pdf"},{"id":94049748,"identity":"8bc51fe9-e0a4-4a4a-8dc5-9b206880d941","added_by":"auto","created_at":"2025-10-21 23:41:09","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":2660937,"visible":true,"origin":"","legend":"","description":"","filename":"HSCTExtendeddata.docx","url":"https://assets-eu.researchsquare.com/files/rs-7763072/v1/b34a2b51bd448e7ad0eb1d8f.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Gut translocation of antimicrobial resistant pathogens in patients undergoing haematopoietic stem cell transplantation in India","fulltext":[{"header":"Introduction","content":"\u003cp\u003eBetween 15\u0026ndash;65% of all haematopoietic stem cell transplantation (HSCT) patients develop bacterial bloodstream infections (BSI), two-thirds of which are associated with mucosal barrier injury (MBI) (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). In 2020, a study from India reported that ~\u0026thinsp;50% of all BSIs in HSCT patients were MBI-associated, termed MBI laboratory-confirmed BSIs (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). While studies often attribute these episodes to bacterial translocation (BT) from the gut, in critically ill patients, this inference is largely derived via intestinal barrier integrity. BT is a transient occurrence and plays an important role in immune modulation (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e), however, in immunosuppressed HSCT patients with intestinal barrier dysfunction, this immunomodulatory function of the gut microbiota can contribute to BT and ultimately, bacteraemia (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGut microbiomes of HSCT patients are enriched with pathobiont Pseudomonadota (Proteobacteria) and Bacillota (Firmicutes) and depleted of commensal short-chain fatty acid-producing Bacteroidota (Bacteroidetes) (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Among the Pseudomonadota, \u003cem\u003eEnterobacteriaceae\u003c/em\u003e are the most common aetiologic agents of Gram-negative bacteraemia in HSCT patients (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e), accounting for ~\u0026thinsp;75% of culture-positive BSIs post-HSCT in India (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Risk of \u003cem\u003eEnterobacteriaceae\u003c/em\u003e BSIs is amplified by their ability to accumulate antimicrobial resistance (AMR) determinants, which may be exacerbated by stress induced by intestinal inflammation (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e), with the gut microbiome acting as a resistance gene reservoir (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Sardzikova \u003cem\u003eet al.\u003c/em\u003e reported that multi-drug resistant (MDR) \u003cem\u003eE. coli\u003c/em\u003e, \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e, \u003cem\u003eEnterococcus faecium\u003c/em\u003e and \u003cem\u003eStaphylococcus aureus\u003c/em\u003e were the leading contributors to the resistome in the HSCT gut (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn high income countries (HICs), 9\u0026ndash;44% of BSIs in HSCT patients are associated with extended-spectrum beta-lactamase (ESBL)-producing \u003cem\u003eEnterobacteriaceae\u003c/em\u003e (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e), with higher mortality rates associated with carbapenem-resistant \u003cem\u003eEnterobacteriaceae\u003c/em\u003e BSIs (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). Low-middle income countries (LMICs) consistently report a higher prevalence of AMR in almost all bacteria, yet syndrome-specific data (e.g., AMR in patients with BSIs post-HSCT) are scarce (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). The cocktail of immunosuppressants and antimicrobials administered post-HSCT also reduces the sensitivity of culture-based diagnostics of BSI, thus limiting the application of antimicrobial susceptibility testing (AST) (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). The growing challenge posed by AMR in LMICs (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e), coupled with the limited HSCT patient data in this region, underscores the need for further investigation, where genomic approaches could offer new insights and solutions.\u003c/p\u003e\u003cp\u003eHere, we investigated AMR in the gut resistomes of HSCT patients at a tertiary care centre in a LMIC with a high AMR burden and identified episodes of BSI attributable to gastrointestinal BT. We utilised plate-sweep enrichment and shotgun metagenomics on stool samples for targeted characterisation of the HSCT gut resistome, focusing on priority AMR pathogens, with enhanced metagenome-assembled genome (MAG) recovery from stool samples.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStudy design\u003c/h2\u003e\u003cp\u003eThis study was a prospective, longitudinal study enrolling 81 participants undergoing allogeneic HSCT over a period of 7 months between May and December 2022 at the Department of Haematology, Christian Medical College (CMC), Vellore, Tamil Nadu, India. The study was a collaboration between the Cambridge Institute for Therapeutic Immunology and Infectious Disease (CITIID) and the Clinical Microbiology and Clinical Haematology departments at CMC Vellore. This study was granted ethical approval by the Institutional Review Board (IRB), Research and Ethics committee at CMC Vellore IRB Min No. 14605 dated 27.04.2022.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eParticipant eligibility criteria and clinical data management\u003c/h3\u003e\n\u003cp\u003ePatients were eligible if they were admitted to the Department of Haematology at CMC Hospital, scheduled to undergo an allogeneic stem cell transplant and provided informed consent to participate. Informed consent forms were written in English and translated into Hindi and Tamil. Eligible patients (as per clinician\u0026rsquo;s assessment) or their legal representatives, were asked for informed consent, following which a clinical case report form was completed. Recorded clinical metadata included age, sex, diagnosis, donor age, donor sex, HLA match, type of transplant, preparative regimen, post-HSCT infection episodes, episodes of Clostridioides difficile colitis, GvHD and mucositis, date of absolute neutrophil count (ANC) recovery, total days of hospital stay, and outcome. Additional clinical information was obtained from the clinicians or nurses in-charge, accessed through the clinical workstation, or through discharge/death summaries. Sensitive clinical information was accessed through secure hospital servers and shared only with key personnel.\u003c/p\u003e\n\u003ch3\u003eSample collection and anonymisation\u003c/h3\u003e\n\u003cp\u003eStool samples were collected were collected in sterile stool sample collection containers and transferred to the Department of Clinical Microbiology at 4\u0026ordm;C. Once received, samples were labelled to indicate that it was a study sample and processed within 72 hours. Blood samples were collected and sent to the Microbiology department for culture and AST if a BSI was suspected by the clinician in-charge, for example, if the patient developed a fever post-HSCT. Stool samples were anonymised with a code that represented the patient number, timepoint relative to HSCT, day relative to HSCT, and sample number from that patient. For example, P02AT0353 represents patient 2, after transplant (AT), collected on day\u0026thinsp;+\u0026thinsp;35, and the third sample from this patient. The blood sample IDs contained an additional code (Blood Adult \u0026ndash; BA, Blood Paediatric \u0026ndash; BP or Blood Venous \u0026ndash; BV). This code was placed between the day relative to HSCT and sample number from that patient. For example, P02AT005BP2 represents a blood sample for patient 2, after transplant, collected on day\u0026thinsp;+\u0026thinsp;5, blood paediatric, and the second sample from this patient.\u003c/p\u003e\n\u003ch3\u003eStool sample processing\u003c/h3\u003e\n\u003cp\u003eUp to 25g (or mL, depending on consistency) of stool was transferred from the stool collection container into sterile 50mL Falcon tubes using Pasteur pipettes. If stool consistency was very hard (Type 1 or type 2 according to the Bristol stool chart classification) (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e), 1g of autoclaved glass beads were added into the 50mL Falcon tube prior to transferring stool from the collection container. Once stool was transferred, the Falcon tube(s) were tightly closed and weighed. From the total weight, the weight of the Falcon tube (13.0g (Abdos) or 13.3g (Tarsons)) was subtracted from the total weight (and 1g in addition to the tube weight was subtracted from the total weight if glass beads were added). Once the stool sample was weighed, sterile PBS was added to the Falcon tubes; the volume added was equivalent to the weight of the stool, such that the stool could be resuspended in a 1:1 ratio.\u003c/p\u003e\u003cp\u003eSamples were vortexed until Falcon tube contents were homogenised, and the stool was completely resuspended in PBS. To ensure sample homogenisation, the tube was shaken from time to time to move the glass beads around during vortexing. After vortexing, samples underwent a slow centrifuge at 400\u003cem\u003eg\u003c/em\u003e for 10 mins at 4\u0026deg;C.\u003c/p\u003e\n\u003ch3\u003eSerial dilutions and plating of stool samples\u003c/h3\u003e\n\u003cp\u003eThe 50mL Falcon tubes were retrieved from the centrifuge and 100\u0026micro;l of the supernatant was added into 900\u0026micro;l of sterile PBS to produce a 10x dilution. The 10x Eppendorf was closed tightly and inverted multiple times to ensure complete resuspension of the neat supernatant in the PBS. The 10x dilution was further diluted into a 100x dilution of neat stool.\u003c/p\u003e\u003cp\u003eMacConkey agar plates were prepared without crystal violet, so they also grew Gram-positive organisms belonging to the \u003cem\u003eStaphylococcus\u003c/em\u003e and \u003cem\u003eEnterococcus\u003c/em\u003e genera. Plates were labelled with the sample ID and dilution factor and 50\u0026micro;l of 100x dilution was pipetted onto the plate and spread using a sterile L-spreader. Inoculated MacConkey plates were incubated at 37\u0026deg;C overnight.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eStool sample plate sweeps\u003c/h2\u003e\u003cp\u003eFollowing overnight incubation, 5mL of sterile PBS was added to each plate. A complete sweep of the plate was collected by resuspending any grown colonies in PBS, aspirating the resuspended culture, and transferring the plate-sweep culture aspirate into clean 15mL Falcon tubes. Then, 1mL of culture was aliquoted into each of three Eppendorf tubes \u0026ndash; destined for DNA extraction, formalin inactivation (if eligible) and glycerol stocks (2 tubes) in trypticase soy broth.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eBlood sample collection and processing\u003c/h3\u003e\n\u003cp\u003eBlood culture bottles were incubated in the BacT/ALERT\u0026reg; automated Blood Culture system. Positive samples were passaged either on blood agar or selective media chosen according to the pathogen detected to obtain pure isolates for WGS. Isolates were incubated overnight in nutrient broth at 37\u0026ordm;C, and a 1 mL aliquot of the overnight culture was used for DNA extraction and WGS.\u003c/p\u003e\n\u003ch3\u003eFormalin inactivation\u003c/h3\u003e\n\u003cp\u003eStool samples collected at the nearest timepoint to a patient\u0026rsquo;s culture-confirmed BSI (n\u0026thinsp;=\u0026thinsp;13) were formalin inactivated for Hi-C metagenomic sequencing. The bacterial suspension was pelleted by centrifugation at 10,000 rpm for 10 minutes at room temperature. Pellets were resuspended in 1X TE buffer using an equal volume to that of the original culture and centrifuged again; this wash step was repeated once more. The bacterial pellet was resuspended in TE buffer containing a final concentration of 2.5% formaldehyde (formalin solution). Samples were incubated at room temperature for 30 min, followed by an incubation at 4\u0026deg;C for 20 min. The formalin fixation reaction was quenched by adding ice-cold glycine added to a final concentration of 0.25M. The sample tubes were inverted a few times to ensure complete resuspension and incubated for 5 minutes at room temperature, followed by a 15-minute incubation on ice. The tubes were then centrifuged for 10 minutes at 10,000 rpm at 4\u0026deg;C. The bacterial pellet was washed twice in 2mL 1X TE and centrifuged in 2mL microcentrifuge tubes. After centrifugation, the supernatant was discarded, and bacterial pellets were frozen at -80\u0026deg;C.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eDNA extraction\u003c/h2\u003e\u003cp\u003eDNA extraction was done using the Wizard\u0026reg; Genomic DNA purification kit by Promega as per the manufacturer\u0026rsquo;s protocol. Briefly, 1mL overnight plate sweep aliquot was centrifuged at 13,000\u0026ndash;16,000\u003cem\u003eg\u003c/em\u003e for 2 minutes to pellet the cells. The supernatant was removed and 600\u003cem\u003e\u0026micro;\u003c/em\u003el of nuclei lysis solution was added to the pellet and mixed by pipetting. The sample tubes were incubated at 80\u0026ordm;C for 5 minutes to lyse the cells, then cooled down to room temperature. RNAse solution (3\u003cem\u003e\u0026micro;\u003c/em\u003el) was added to the cell lysates, which were then mixed by inversion. The samples were incubated at 37\u0026ordm;C for 15\u0026ndash;60 minutes and then cooled down to room temperature. 200\u003cem\u003e\u0026micro;\u003c/em\u003el of protein precipitation solution was added to the RNAse-treated cell lysates. Tubes were vortexed vigorously at a high speed for 20 seconds to mix the protein precipitation solution with the cell lysate. The samples were incubated on ice for 10 minutes, followed by centrifugation at 13,000\u0026ndash;16,000\u003cem\u003eg\u003c/em\u003e for 5 minutes.\u003c/p\u003e\u003cp\u003eAfter centrifugation, the supernatant containing the DNA was transferred to a clean 1.5mL microcentrifuge tube containing 600\u003cem\u003e\u0026micro;\u003c/em\u003el of room-temperature isopropanol. The samples were mixed gently by inversion until thread-like strands of DNA formed. The tubes were then centrifuged again at 13,000\u0026ndash;16,000\u003cem\u003eg\u003c/em\u003e for 3 minutes. Pellets were air-dried before adding 100\u003cem\u003e\u0026micro;\u003c/em\u003el of DNA rehydration solution and incubating at 65\u0026ordm;C for one hour, periodically mixing the solution by gently tapping the tube. The DNA was then stored in the laboratory fridge at 4\u0026ordm;C awaiting shipment to Cambridge.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eShipping and sequencing\u003c/h2\u003e\u003cp\u003eBlood DNA and stool DNA were shipped at 4\u0026ordm;C from CMC to CITIID. DNA concentrations were quantified on a Qubit\u0026trade; 2.0 Fluorometer, and 50\u0026micro;l of DNA from each sample was transferred into barcoded snap-cap microcentrifuge tubes and shipped to Eurofins Genomics for standard library preparation and sequencing. DNA samples were sequenced on an Illumina NovaSeq 6000 S4 PE150 instrument, with a minimum of 10\u0026nbsp;million paired-end reads requested for metagenomic sequencing of stool samples and 100x coverage for WGS for genomes up to 10Mb. The sequencing reads for both stool and blood samples from the study can be found under BioProject accession PRJNA1072756.\u003c/p\u003e\u003cp\u003eA total of 308 stool samples were collected. Of these, 22 patients were excluded from the study because they underwent autologous HSCTs or had undergone transplantation prior to the study start date. Two-hundred and ninety-eight (298) samples were sent for whole-metagenome shotgun sequencing at Eurofins Genomics. Twenty samples were not sequenced since they contained insufficient DNA to generate libraries that met the QC requirements for successful sequencing, containing less than 1nM/L of DNA. In addition, the stool sample collection sent for sequencing also contained samples belonging to patients for whom we retrieved incomplete metadata, or those who were excluded for reasons described in the patient eligibility section above. Once their corresponding stool samples were removed, the complete stool sample collection comprised a total of 252 samples from the 81 patients who were eligible for the study.\u003c/p\u003e\u003cp\u003eFor metagenomic Hi-C, 13 formalin-inactivated stool samples were sent to Phase Genomics in Seattle, USA for Hi-C library prep and sequencing according to the ProxiMeta\u003csup\u003e\u0026trade;\u003c/sup\u003e Hi-C Kit Protocol v4.5. Contigs assembled from metagenomic read data were sent to the bioinformatics team at Phase Genomics to overlay Hi-C connections and interpret the data. Two stool samples (P14AT0065 and P51AT1247) did not pass QC for Hi-C.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eQuality control and assembly of sequencing reads\u003c/h2\u003e\u003cp\u003eRaw sequencing reads from the enriched stool samples were run through the nf-core/mag pipeline v2.3.0 (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). Short reads underwent trimming and adapter removal to improve read quality, before removing PhiX reads using Bowtie2. All reads were then passed onto FastQC for final quality control. Metagenomic read assembly was performed using SPAdes v3.15.4, with gene prediction carried out by Prodigal (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e). All assembled contigs underwent QC using QUAST (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAlthough the nf-core/mag pipeline had downstream binning functions, we used Metawrap v1.3.0 (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e) to produce consensus bins from three popular binning pipelines MetaBAT2 v2.12.1 (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e), Maxbin2 v2.2.7 (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e) and CONCOCT v1.1.0 (\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e). The threshold for consensus bins was set to a minimum of 70% completeness and \u0026lt;\u0026thinsp;5% contamination to extract bins of high purity. After extracting consensus bins, each bin was passed through GTDB-TK (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e) for species identification.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eTaxonomic assignment\u003c/h2\u003e\u003cp\u003eTaxonomic assignment was performed directly on raw stool sample reads using Kraken v2.1.2 (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e), followed by Bayesian re-estimation at genus level using Bracken v2.7.0 (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). Since we were using enriched samples, we found negligible levels of human DNA in our samples, however, residual human OTUs were removed with the \u0026ldquo;extract_kraken_reads.py\u0026rdquo; script. The Bracken outputs were combined using the \u0026ldquo;combine_kraken_reports.py\u0026rdquo; script and converted into a .biom format using kraken-biom v1.2.0 from the KrakenTools software suite (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). OTU tables were imported into RStudio v4.2.3 and visualisations for total taxonomic composition of the stool samples in the study collection were carried out using metacoder v0.3.6 (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). OTU IDs were merged with a dataframe from NCBI containing information for that ID at each taxonomic level to produce a data frame containing OTU IDs, abundance data and taxonomy information. This table was converted into a taxmap object which is specialised for community abundance data, using metacoder\u0026rsquo;s parse_tax_data function, by specifying which columns contain taxonomy information, the order they appear in, and whether the column headers are named by the taxonomic rank.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eMicrobial diversity\u003c/h2\u003e\u003cp\u003eBeta diversity was calculated using the Bray-Curtis dissimilarity index with the vegdist function from the vegan package. Beta diversity was plotted with the non-metric multidimensional scaling (NMDS) technique, computed using the metaMDS function optimised at 999 iterations with 3 dimensions which had the lowest stress value at 0.07 (\u0026lt;\u0026thinsp;0.099 represents a \u0026lsquo;good\u0026rsquo; fit (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e)). The NMDS scores were then extracted into a new data frame, which was merged with sample metadata and plotted. Analysis of similarities (ANOSIM) was calculated using the anosim function from the vegan package and the pairwise adonis test was performed using the pairwiseAdonis v0.4.1 (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e) package on R.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eAntimicrobial resistance gene detection\u003c/h2\u003e\u003cp\u003eAMR genes were detected from whole-genome and metagenomic sequences using ABRIcate v1.0.1 using the CARD database as reference. The ABRIcate pipeline was run with default parameters under the assumption that enriched organisms would be sequenced at sufficient coverage to reliably detect AMR genes in them. AMR gene alpha/beta diversities were also calculated to probe for longitudinal variations. Antimicrobial drug classes were reclassified as \u0026lsquo;multidrug\u0026rsquo; if the AMR gene detected conferred resistance to \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\ge\\:\\)\u003c/span\u003e\u003c/span\u003e3 antimicrobial classes.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003eAntimicrobial sensitivity testing\u003c/h2\u003e\u003cp\u003ePhenotypic antimicrobial sensitivity testing was performed on bacteria isolated from blood and stool using disk diffusion tests against 8 antimicrobials for cultured Gram-negative organisms and 4 antimicrobials for Gram-positive organisms. 6mm disks of Whatmann No. 1 filter paper were saturated with antimicrobial solutions. Gram-negative isolates were tested against Cefotaxime and Ceftazidime (third generation cephalosporins), Cefoperazone-Sulbactam (third generation cephalosporin with sulphonamide), Cefepime (fourth generation cephalosporin), Amikacin and Gentamicin (aminoglycosides), Ertapenem and Meropenem (carbapenems). Gram-positive organisms were tested against Linezolid (oxazolidinone), Ampicillin (penicillin), high-level Gentamicin (HLG; aminoglycoside) and Vancomycin (glycopeptide).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003eRandom forest classification\u003c/h2\u003e\u003cp\u003eUsing the randomForest package v4.7-1.1 (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e) on RStudio, the random forest was computed by classifying 1000 bootstrapped datasets on 500 decision trees with 3 variables considered at each split. The estimated out-of-bag (OOB) error rate was 9.9%, suggesting that 90.1% of the samples were classified correctly. The OOB error rates were then plotted against the number of decision trees to identify the optimal number of trees (~\u0026thinsp;150) required to obtain the lowest OOB error rate. To select the optimal number of variables to consider at each node of the decision trees we used an iterative \u003cem\u003efor\u003c/em\u003e loop to test the random forest using a different number of variables (i.e., 1\u0026ndash;10) from the original dataset at each step in the decision trees. In this case, the lowest OOB error rate was identical when using 2, 3 or 4 variables at each step, so we selected the default value of sampling 3 variables. This random forest was then used to draw a multidimensional scaling (MDS) plot.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eInStrain analysis\u003c/h2\u003e\u003cp\u003eInStrain v1.8.0 (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e) was used in this study to detect strains shared by multiple samples. Since InStrain was developed for WMS data, the program was run with default settings so that strain-level information could be detected from the enriched sequencing data with high confidence. For this analysis, the consensus genome bins extracted from the metaWRAP pipeline were used as reference genomes. Fasta headers in the consolidated bin files were modified to include the sample ID, such that each contig would have a unique identifier when concatenating to produce a single fasta file containing all genome bins extracted from the enriched stool sample collection. Concurrently, a scaffold-to-bin (.stb) file was produced using the \u0026lsquo;parse_stb.py\u0026rsquo; script (packaged with InStrain) to associate contigs in the combined assembly file to their original genome bin files. Then, we built a bowtie index for the combined fasta file containing extracted genome bins. Next, all the quality-checked metagenomic and whole-genome short read sequencing data were mapped to the bowtie index to identify regions where reads aligned to the index. InStrain Profile was run on on each resulting .bam file, using the --database-mode flag. The InStrain profile command filters .bam files to only retain reads that map with sufficient quality and \u0026gt;\u0026thinsp;5x coverage, hence minimising mismatches and extracting unique mappings. The reads that pass these filters are considered for microdiversity analyses that calculate nucleotide diversity at each base, in addition to scaffold-level properties including coverage, and ANI between the reads and the reference genomes.\u003c/p\u003e\u003cp\u003eFinally, the inStrain compare function was used to produce outputs containing scaffold- (comparisonsTable.tsv) and genome-level (genomeWide_compare.tsv) comparisons of the InStrain profile objects created for each sample in the step above. Pairwise comparisons in the genomewide comparisons table underwent hierarchal clustering to generate strain-level clusters. This output was merged with the GTDB-TK species-level classification output to combine species information with strain-level data.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003eMicrobial network analysis\u003c/h2\u003e\u003cp\u003eThe genome-level output from the InStrain compare function was used to produce a microbial network, to illustrate shared strains between patient samples The data frame containing genome-level comparisons was filtered to include only genomes that mapped with \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\ge\\:\\)\u003c/span\u003e\u003c/span\u003e 99.999% ANI with \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\ge\\:\\)\u003c/span\u003e\u003c/span\u003e 25% of the metagenomic reads compared to the reference index. This step reduced the genome dataset from containing 44,066 to 1,353 genome-to-genome relationships. This data frame was subset to include only columns containing the two samples and the consensus bin to which they mapped. The consensus bins were then merged with species-level taxonomy information from GTDB-TK.\u003c/p\u003e\u003cp\u003eThe microbial network was created using the igraph v1.5.1 (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e), tidygraph v1.2.3 (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e) and ggraph v2.1.0 (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e) packages on RStudio to produce a directed acyclic table graph object. Each node (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;98) represented a sample (blood or stool) and edges between nodes denote connections (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;174) between samples. A column called \u0026lsquo;degree\u0026rsquo; was added to the node table, which represents the number of connections to each sample. This was then used to produce a network plot using ggraph under the \u0026lsquo;kk\u0026rsquo; layout.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\n \u003ch2\u003eCohort summary and outcomes\u003c/h2\u003e\n \u003cp\u003eA summary of the baseline clinical characteristics of the allogeneic HSCT cohort can be found in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. Overall, 32/81 patients developed a clinically significant bacterial infection (38 episodes total) post-HSCT. \u003cem\u003eClostridioides difficile\u003c/em\u003e colitis was the most common infection post-HSCT (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;11), followed by \u003cem\u003eEnterobacteriaceae\u003c/em\u003e (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;7) and coagulase negative \u003cem\u003eStaphylococcus\u003c/em\u003e (CONS) (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5). Time taken for neutrophil recovery was the strongest predictor of discharge post-HSCT (Extended data (ED) 1).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eBaseline characteristics of the HSCT study cohort\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eBaseline characteristics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e38 (46.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e18\u0026ndash;30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13 (16.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e31\u0026ndash;40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13 (16.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e41\u0026ndash;50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (13.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e51\u0026ndash;60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 (7.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex (% male)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e53 (65.4%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiagnosis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eMalignant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e39 (48.1%\u003c/strong\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAcute lymphoblastic leukaemia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13 (33.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAcute myeloid leukaemia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (23.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMyelodysplastic syndrome\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 (20.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eChronic myeloid leukaemia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 (12.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHodgkin\u0026apos;s lymphoma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 (5.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePolycythemia vera\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNK T-cell lymphoma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eNon-malignant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e42 (51.9%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eThalassemia major\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e17 (40.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSevere aplastic anaemia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (35.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFanconi anaemia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (9.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSickle cell anaemia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWiskott Aldrich syndrome\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMucopolysaccharidosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eChronic granulomatous disease\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSevere congenital neutropenia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eParoxysmal nocturnal haemoglobinuria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (2.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eType of allogeneic transplant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eFully matched\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e51 (63.0%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRelated donor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e40 (49.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnrelated donor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (13.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eHaploidentical\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e30 (37.0%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eConditioning regimen\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMyeloablative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e58 (71.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eReduced intensity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e23 (28.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eOutcome\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDischarged\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e66 (81.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDied\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (18.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eThere were 10 episodes of culture-positive BSI prior to HSCT among seven patients (resolved before transplantation). Excluding catheter and central-line associated BSI, 21.0% (17/81) of patients experienced bacterial BSIs post-HSCT (20 total). The median time to culture-positive BSI was 12 days post-HSCT. We retrieved 19 bacterial BSI isolates, which were whole genome sequenced (WGS; 7/10 pre- and 12/20 post-HSCT). Gram-negative bacilli (GNB) caused 65.0% (13/20) of post-transplant BSI episodes. The mortality rate among patients who developed BSI post-HSCT was 41.2% (7/17), whereas the mortality rate among patients who developed any infectious complications post-HSCT was 37.5% (12/32). ED2 contains a summary of all clinically significant organisms detected post-HSCT.\u003c/p\u003e\n \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e\n \u003ch2\u003eBacterial taxa isolated from stool via plate-sweeps\u003c/h2\u003e\n \u003cp\u003eA median of three stool samples were collected from each patient (range 1\u0026ndash;8), with at least three stool samples collected for 61/81 patients. Samples from the remaining 20 patients were included for AMR gene profiling. Stool samples (n\u0026thinsp;=\u0026thinsp;252) were categorised into five groups based on the timepoint relative to HSCT: pre-conditioning (before day \u0026minus;\u0026thinsp;8; \u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;12), conditioning week (day \u0026minus;\u0026thinsp;8 to day 0; \u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;78), pre-engraftment (day\u0026thinsp;+\u0026thinsp;1 to day\u0026thinsp;+\u0026thinsp;15; \u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;89), early post-engraftment (day\u0026thinsp;+\u0026thinsp;16 to day\u0026thinsp;+\u0026thinsp;100; \u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;67) and late post-engraftment (day\u0026thinsp;+\u0026thinsp;101 to day\u0026thinsp;+\u0026thinsp;200; \u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;6). 100x dilutions of homogenised neat stool were plated on MacConkey agar, incubated overnight; a plate sweep and DNA extraction was performed the following day.\u003c/p\u003e\n \u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e shows an overview of the bacterial taxa that were enriched in the MacConkey plate sweeps. The omission of crystal violet permitted the growth of \u003cem\u003eStaphylococcus\u003c/em\u003e spp. and \u003cem\u003eEnterococcus\u003c/em\u003e spp. and targeted Gram-negative organisms (i.e., \u003cem\u003eE. coli\u003c/em\u003e, \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e, \u003cem\u003eAcinetobacter baumannii\u003c/em\u003e and \u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e). This approach selected for specific bacterial species commonly associated with AMR, denoted by the focused colour dispersion within specific branches of Pseudomonadota (Proteobacteria; Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eA) and Bacillota (Firmicutes; Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e\n \u003cp\u003eThe 252 stool samples generated between 1,762,443\u0026ndash;33,901,474 reads per sample, with Kraken2 detecting 3,493,496,303 individual operational taxonomic units (OTUs) spanning 4,272 unique taxonomic levels. This dataset was rarefied at the 10th percentile, representing approximately a third of the original dataset with 1,652,408,590 OTUs. Over three-quarters (77.6%; \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1,282,928,757) of the OTUs belonged to the \u003cem\u003eEnterobacteriaceae\u003c/em\u003e family in the Pseudomonadota (Proteobacteria) phylum. At genus level, this family mainly consisted of \u003cem\u003eEscherichia\u003c/em\u003e (73.4%), \u003cem\u003eKlebsiella\u003c/em\u003e (21.8%), \u003cem\u003eShigella\u003c/em\u003e (1.54%), \u003cem\u003eCitrobacter\u003c/em\u003e (1.38%) and \u003cem\u003eEnterobacter\u003c/em\u003e (1.34%). In the Bacillota (Firmicutes) phylum (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;217,106,676), the Lactobacillales and Bacillales comprised 89.6% (\u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;194,435,721) and 10.4% (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;22,640,084) of OTUs, respectively. The Lactobacillales order was dominated by \u003cem\u003eEnterococcus\u003c/em\u003e spp. (99.8% relative abundance; \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;193,964,108), and the genus \u003cem\u003eStaphylococcus\u003c/em\u003e was the most abundant (99.6% relative abundance; \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;22,550,143) in the Bacillales.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e\n \u003ch2\u003eTaxonomic diversity of enriched stool from HSCT patients\u003c/h2\u003e\n \u003cp\u003eA Shannon alpha diversity analysis of enriched stool samples at species-level did not show any significant difference (one-way ANOVA) when grouped by timepoints relative to HSCT (ED3). Beta diversity was calculated using the Bray-Curtis dissimilarity index (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eA) and demonstrated clustering of conditioning week samples (orange), with weaker grouping of samples during pre-engraftment (blue). While differences in beta diversity within grouped timepoints were not significant, there were significant differences between stool collection timepoints (ANOSIM R statistic\u0026thinsp;=\u0026thinsp;0.049; **\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004; ED4). Specifically, there was a significant difference (pairwise adonis ***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; ED4) in beta-diversity when comparing samples collected during the conditioning week to the pre-engraftment, early-, and late post-engraftment timepoints; genera contributing to these differences are shown in ED5. No significant associations were observed when assessing beta diversity in relation to transplant outcomes or graft-versus-host disease (GvHD). However, patients who developed acute GvHD post-HSCT demonstrated significant log\u003csub\u003e2\u003c/sub\u003e fold-increases in \u003cem\u003eEnterococcus\u003c/em\u003e (***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and \u003cem\u003eMorganella\u003c/em\u003e (***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eB) genera.\u003c/p\u003e\n \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e\n \u003ch2\u003eThe role of antimicrobial resistance genes in the HSCT cohort\u003c/h2\u003e\n \u003cp\u003eABRIcate was used to detect AMR genes within the enriched microbiome samples, using the CARD database. 16,903 unique AMR genes were detected among the 252 enriched stool samples, with efflux-mediating genes representing 65.8% (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;11,118) of AMR genes detected within the collection. Non-efflux AMR genes (n\u0026thinsp;=\u0026thinsp;5,785) constituted the remainder.\u003c/p\u003e\n \u003cp\u003eThe median number of AMR genes detected in the pre-conditioning phase (\u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;12) and conditioning week (\u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;78) were 62 and 68 AMR genes per sample, respectively, compared to a median of 55 AMR genes detected per sample during the pre-engraftment phase (\u003cem\u003en\u0026thinsp;=\u003c/em\u003e\u0026thinsp;89). Correspondingly, a median Shannon alpha diversity index of 3.23 was observed for AMR genes during both the pre-conditioning and conditioning phases (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eA). At the pre-engraftment timepoint, the Shannon index dropped to 3.06. A decrease in the abundance and diversity of AMR genes between the conditioning week and pre-engraftment timepoints were both significant at ***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001 when tested with a pairwise t-test with Bonferroni correction (ED6).\u003c/p\u003e\n \u003cp\u003eAt early post-engraftment, there was a significant increase in the abundance (**\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) and diversity (***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) of AMR genes compared to the pre-engraftment timepoint. The median number of AMR genes detected at early post-engraftment was 69 (n\u0026thinsp;=\u0026thinsp;67); alpha diversity\u0026thinsp;=\u0026thinsp;3.32) and 96 at late post-engraftment (n\u0026thinsp;=\u0026thinsp;6; alpha diversity\u0026thinsp;=\u0026thinsp;3.46). Non-efflux AMR gene diversity mirrored the entire AMR gene set (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eB), with a significant decrease (*\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) between the conditioning week (median Shannon index\u0026thinsp;=\u0026thinsp;2.73) and pre-engraftment (median Shannon index\u0026thinsp;=\u0026thinsp;2.66), and a significant increase (***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) between pre-engraftment and early post-engraftment (median\u0026thinsp;=\u0026thinsp;2.89). The abundance of non-efflux AMR genes was significantly higher at early (**\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) and late (*\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) post-engraftment compared to pre-engraftment, which suggests that organisms carrying non-efflux AMR genes remain in the HSCT resistome post-transplantation. This observation may indicate preferential selection through robust tolerance and survival mechanisms among organisms carrying non-efflux AMR genes, albeit only six samples were collected at late post-engraftment. However, the increase in non-efflux AMR genes between pre-engraftment and late post-engraftment remained significant (*\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) when we subsampled the resistome data to only include the six participants for whom we had data.\u003c/p\u003e\n \u003cp\u003eThe abundance of five MDR-conferring ESBL/carbapenemase genes (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;695) \u0026ndash; \u003cem\u003eblaOXA\u003c/em\u003e, \u003cem\u003eblaNDM\u003c/em\u003e, \u003cem\u003eblaCTX-M\u003c/em\u003e, \u003cem\u003eblaSHV\u003c/em\u003e and \u003cem\u003eblaTEM\u003c/em\u003e was then plotted longitudinally against the HSCT timeline in Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eC. At least 1/5 MDR genes were detected in 216/252 stool samples. Overall, we observed a significant successive increase in the abundance of MDR genes when comparing samples collected during the early post engraftment timepoint to MDR genes detected in samples during the two preceding conditioning week (***\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and the pre-engraftment (**\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) timepoints. Notably, the \u003cem\u003eblaOXA\u003c/em\u003e and \u003cem\u003eblaNDM\u003c/em\u003e carbapenemase genes exhibited progressive increases in their mean abundance per sample between the conditioning week (0.60 and 0.23, respectively) and late post-engraftment (1.40 and 0.8, respectively). Using Hi-C metagenomics, we found that \u003cem\u003eblaOXA\u003c/em\u003e and \u003cem\u003eblaNDM\u003c/em\u003e were often co-localised on plasmids harboured by \u003cem\u003eE.coli\u003c/em\u003e and \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e. Such plasmids were detected in half of all patients who developed culture positive BSIs (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD, ED7). These plasmids also carried genes inducing resistance against tetracyclines, trimethoprim, sulphonamides, and macrolides. AST data for select enteric MDR bacteria (\u003cem\u003eE. coli\u003c/em\u003e, \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e and \u003cem\u003eEnterococcus faecium\u003c/em\u003e) isolated from stool (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;90) and enriched stool resistomes are depicted in ED8. MDR isolates were detected in 57 samples from 41 patients, signifying that \u0026gt;\u0026thinsp;50% of patients had at least one MDR organism isolated from their stool during the HSCT timeline.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e\n \u003ch2\u003eSpatiotemporal isolate tracking within and between participants\u003c/h2\u003e\n \u003cp\u003ePlate-sweep metagenomics offers a factor-increase in the resolution of MAGs recovered from stool samples. This approach enabled us to probe microdiversity within enriched taxa to identify organisms shared across multiple patients within the cohort, and match bloodstream isolates from clinical infection episodes to gut microbiome samples. Figure \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e shows a microbial network which depicts (at isolate level) bacterial taxa detected across multiple stool samples. Consensus bins (\u0026gt;\u0026thinsp;70% complete, \u0026lt;\u0026thinsp;5% contaminated) were extracted from the enriched metagenomic sequences using the metaWRAP pipeline and indexed, following which raw sequencing reads from enriched stool metagenomes were mapped against the indexed reference bins.\u003c/p\u003e\n \u003cp\u003eTo detect the same isolate among multiple participants, or across multiple timepoints within the same participant, we used a stringent average nucleotide identity (ANI) threshold of 99.999% at the population-level (popANI), where at least 25% of the reads were compared against the reference bins. The resulting microbial network consisted of 98 nodes and 174 connections; the degree annotation represents the number of edges that connect to each node (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). The network depicts clusters where metagenomic reads from the enriched gut microbiome samples mapped with \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\ge\\:\\)\u003c/span\u003e\u003c/span\u003e99.999% popANI to reconstructed MAG reference bins, indicating highly similar isolates detected across multiple samples. There were several episodes (participants 3, 11, 46, 54, 57, and 61) where isolates could be linked longitudinally within the same participant, and one participant (participant 36) who had MAGs from \u003cem\u003eEscherichia coli\u003c/em\u003e and \u003cem\u003eProvidencia alcalifaciens\u003c/em\u003e (\u0026gt;\u0026thinsp;85% complete, \u0026lt;5% contaminated) that could be linked across timepoints. Transmission links where two edges were shared between samples indicate episodes where metagenomic reads from both samples mapped to reconstructed MAGs from both samples at the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\ge\\:\\)\u003c/span\u003e\u003c/span\u003e99.999% ANI and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\ge\\:\\)\u003c/span\u003e\u003c/span\u003e25% \u0026lsquo;genome compared\u0026rsquo; thresholds. Cases where these bidirectional relationships appeared across multiple participants likely indicate localised transmission clusters.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e\n \u003ch2\u003eAntimicrobial resistant bacterial translocation from the gut\u003c/h2\u003e\n \u003cp\u003eBSI bacteria were subjected to WGS. To link bloodstream isolates to gut colonisation, we indexed\u0026thinsp;\u0026gt;\u0026thinsp;95% complete and \u0026lt;\u0026thinsp;5% contaminated consensus bins from the metaWRAP pipeline, which fell under the \u0026lsquo;high quality\u0026rsquo; draft category according to MIMAG standards (\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e). For translocation, the minimum threshold for isolate-level detection was set at 99.9% popANI to account for variation introduced by differences in the sampling and sequencing techniques between blood and stool samples, with a breadth threshold set at 50% to minimise false positives (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eSummary of bloodstream isolates from BSI episodes which mapped to the corresponding participant\u0026rsquo;s gut microbiome sample.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBloodstream isolate\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eGenome*\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCompleteness\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eContamination\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSpecies\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBreadth\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eContain\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003epopANI\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP02AT005BP2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP02BT0031_bin.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e95.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eEscherichia coli\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9999\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP05AT014BA1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP05AT0063_bin.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e99.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eEnterococcus faecium\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9997\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9998\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP14AT043BA1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP14BT0061_bin.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e97.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9998\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9998\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP44AT012BA1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP44AT0283_bin.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e99.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9999\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP55BT045BA1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eP55AT0554_bin.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e98.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9999\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.9999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"8\"\u003e*Bloodstream isolates underwent WGS where reads mapped to reconstructed MAGs (\u0026gt;\u0026thinsp;97% completeness, \u0026lt;\u0026thinsp;2% contamination) from the enriched gut microbiome samples of the same participant at \u0026gt;\u0026thinsp;99.9% ANI across \u0026gt;\u0026thinsp;50% breadth of the reconstructed MAG.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eThe cultured organism from the blood samples was matched to gut microbiomes in five patients at isolate-level \u0026ndash; patient 2 (\u003cem\u003eE. coli\u003c/em\u003e), 5 (\u003cem\u003eEnterococcus faecium\u003c/em\u003e), 14 (\u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e), 44 (\u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e) and 55 (\u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e) (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). Patients 2, 5, 14 and 44 had episodes of culture-positive BSI post-HSCT, whereas patient 55 had culture-positive BSI 45 days pre-HSCT. In patients 2, 5, and 14, the BC sample was matched to stool samples collected prior to BSI, and in patients 44 and 55, the BSI isolate was matched to stool samples collected 16 and 100 days after BSI, respectively. In patient 55, the BC and AST report identified the organism as carbapenem-resistant \u003cem\u003eKlebsiella pneumoniae.\u003c/em\u003e After HSCT, patient 55 developed febrile neutropenia with septic shock and grade III mucositis, but empirical treatment with meropenem/colistin/teicoplanin likely contributed to subsequent sterile blood cultures. This observation suggests that while the clinical infection in patient 55 occurred 45 days before HSCT and was resolved prior to HSCT, the organism persisted within the host in the gut. While BT was likely in this case, it is difficult to ascertain in the absence of positive blood cultures being detected post-HSCT. Nonetheless, all five BSI isolates were genotypically MDR and 4/5 were phenotypically MDR (ED9). The inStrain profile data showed high ANI between the bloodstream isolates\u0026rsquo; genomes and reconstructed MAGs from the corresponding patients\u0026rsquo; gut microbiome samples.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eHere, we present a prospective longitudinal study tracking AMR-carrying organisms, the gut resistome, and clinical infections of gut-origin in HSCT patients at a single tertiary care centre in India. Contrary to previous studies which used indirect measurements of gut permeability, invasive culture-based methods, or inferred BT through relative abundance of pathogens in the gut, our methodology was designed to capture priority AMR bacteria in the HSCT resistome associated with post-transplant BSI. This approach builds on a previous retrospective investigation by Korula \u003cem\u003eet al.\u003c/em\u003e (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e), where gut translocation was inferred by matching organisms and AST results between resistant isolates recovered from blood and surveillance stool cultures. We incorporated genomic methods and used selective enrichment to interrogate these observations prospectively at the same hospital.\u003c/p\u003e\u003cp\u003eOverall, 39.5% of all HSCTs performed were complicated by an infection. \u003cem\u003eClostridioides difficile\u003c/em\u003e colitis was the most common infectious complication caused by a single bacterial pathogen post-HSCT, with an incidence of 13.6% which is on the lower end of previously reported estimates (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Over the past two decades, infectious aetiologies post-HSCT have transitioned from Gram-positive cocci such as Staphylococci and Enterococci as the primary causative pathogens, to GNB dominating in recent years (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). In agreement, we found a higher incidence of post-HSCT BSIs caused by GNB, predominantly \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e, \u003cem\u003eE. coli\u003c/em\u003e and \u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e. While the 21.0% incidence of post-HSCT BSI we observed in our cohort lies within previously reported ranges of 10\u0026ndash;38.6% (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan additionalcitationids=\"CR21\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e), the mortality rate was \u0026gt;\u0026thinsp;40% in BSI patients.\u003c/p\u003e\u003cp\u003eBeta diversity of enriched stool samples showed that HSCT patients\u0026rsquo; gut microbiomes experience similar alterations across the cohort based on the timepoint relative to HSCT. These data suggest that the preparative regimen patients undergo during the conditioning week and the antimicrobials administered in the pre-engraftment period modulate the microbiota comparably, regardless of indication for HSCT or transplant type. Wong \u003cem\u003eet al.\u003c/em\u003e also reported analogous longitudinal shifts in faecal microbiome beta-diversities, using 16S data from an Asian cohort undergoing autologous HSCT (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Moreover, Wong \u003cem\u003eet al.\u003c/em\u003e described that gut microbial diversity recovered to preconditioning levels within 6-months post-HSCT, which was beyond our follow-up timeframe.\u003c/p\u003e\u003cp\u003eIn 2018, Tamburini \u003cem\u003eet al\u003c/em\u003e. (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e) retrospectively confirmed the translocation of 15/32 (47%) unique organisms detected in blood cultures from 30 patients, observed at \u0026ge;\u0026thinsp;0.1% relative abundance in the gut in a high-income setting. We introduced a novel approach to monitor translocation of gut pathogens, especially ESBL Enterobacterales, with our enrichment-based strategy supported by Jazmati \u003cem\u003eet al.\u003c/em\u003e (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). BC positivity is low among HSCT patients, since the antimicrobials administered to counter febrile neutropenia in the pre-engraftment phase further reduce the low sensitivity of BCs when used in critically ill patients (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Despite this, we found that over a quarter of sequenced culture-positive BSIs (5/19, 26.3%) were caused by isolates that were also in the gut, and one probable gut-origin BSI episode caused by \u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e. All five patients who had BSIs of gut-origin also showed evidence of enteric mucositis, which has been linked with infections by Gram negative \u003cem\u003eEnterobacteriaceae\u003c/em\u003e, \u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e and Gram-positive Enterococci (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). We were also able to track highly similar isolates longitudinally within and between patients.\u003c/p\u003e\u003cp\u003eGhosh \u003cem\u003eet al.\u003c/em\u003e recently reported that the high prevalence of MDR in an Indian HSCT cohort was primarily associated with the \u003cem\u003eblaNDM\u003c/em\u003e and \u003cem\u003eblaOXA\u003c/em\u003e-\u003cem\u003e48-like\u003c/em\u003e carbapenemase genes (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e), but only with pre-transplant surveillance samples. With the longitudinal sampling strategy, we found many \u003cem\u003eblaNDM\u003c/em\u003e and \u003cem\u003eblaOXA\u003c/em\u003e genes in our cohort, and this abundance significantly increased in the early post-engraftment timepoint when compared to the two preceding timepoints. In a longitudinal study with eight paediatric patients in Italy, D\u0026rsquo;Amico \u003cem\u003eet al.\u003c/em\u003e (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e) also had similar observations, reporting an AMR gene profile that had diversified with MDR genes in addition to consolidating the resistome present prior to HSCT.\u003c/p\u003e\u003cp\u003eStecher \u003cem\u003eet al.\u003c/em\u003e reported that \u003cem\u003eEnterobacteriaceae\u003c/em\u003e bloom under conditions of intestinal inflammation, which also exacerbates HGT (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). This observation highlights the critical role of the gut as a reservoir for transmissible AMR genes which can transfer on mobile genetic elements in immunocompromised hosts who frequently visit hospitals. While whole-metagenomic sequencing (WMS) hinders the association of AMR genes back to their host species, using an enrichment-based approach reduces the pool of potential gene hosts, and permits screening of specific bacterial populations compared to isolation-based approaches. Inadvertent selection of pre-existing bacteria harbouring resistance genotypes can also accelerate AMR development since this mechanism is associated with a lower fitness cost compared to the establishment of \u003cem\u003ede-novo\u003c/em\u003e mutations conferring resistance (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). Therefore, routine surveillance of gut resistome plasticity using stool samples could indicate the most likely infectious aetiologies of AMR-BSI and help optimise treatment regimens to minimise inadvertent selection for resistance among HSCT patients.\u003c/p\u003e\u003cp\u003eThe onset of febrile neutropenia in the HSCT cohort warrants the administration of empirical antimicrobials considering the immunosuppressed state of the recipient and the exposure to a high-risk hospital environment. These broad-spectrum antimicrobials should immediately eliminate organisms that may harbour but not express AMR genes (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). Upon successful engraftment and neutrophil recovery, neutropenic fever subsides among HSCT patients, and antimicrobial regimens are tapered in response. We hypothesise that the removal of this selection pressure permits residual resistant members of the microbiome to proliferate and offers a window for nosocomial AMR pathogens to occupy the niche. This could explain the significant decrease in the abundance and diversity of AMR genes between the conditioning week and pre-engraftment, and the significant increase between pre-engraftment and early post-engraftment.\u003c/p\u003e\u003cp\u003eThis study had limitations. First, this study did not consider the clinical heterogeneity of HSCT patients. Future investigations should focus on more specific cohorts based on their age, indications for HSCT or a defined outcome, e.g., gastrointestinal-GvHD. Attempting these facets was beyond the scope of our study timeframe. Secondly, we cannot directly compare enrichment data with WMS data from neat stool samples, and selective media can potentially lead to uneven enrichment based on organisms\u0026rsquo; differential capacities to metabolise available nutrients. However, since we focused on characterising key AMR-pathogens in the HSCT microbiome, we also avoided biased total microbiota recovery from stored faeces. Finally, while the enrichment strategy outperforms WMS from a neat sample in capturing intraspecific or population-level heterogeneity for the selected organisms, binning algorithms capture the most dominant STs (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). Nonetheless, we were able to account for the population-level heterogeneity across samples because of inStrain\u0026rsquo;s unique \u0026lsquo;microdiversity-aware\u0026rsquo; metrics of conANI and popANI (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThis is the first and largest study to track the longitudinal evolution of the gut resistome and AMR-BSIs of gut-origin among HSCT patients in a LMIC with a high AMR burden. We identified that the HSCT gut microbiome is modulated by the pre-transplant conditioning regimen, and that the gut resistome is significantly depleted at the pre-engraftment phase, likely in response to conditioning regimens and empirical antimicrobials. However, the resistome is promptly restored at early post-engraftment and shifts towards a sustained and significant increase in the abundance of transmissible non-efflux AMR genes into the late post-engraftment phase. Additionally, we associated over one quarter of retrieved BSI isolates with the corresponding participant\u0026rsquo;s gut microbiomes at isolate-level resolution. In conclusion, we have established a baseline for the genomic landscape of gastrointestinal AMR carriage among HSCT patients in South Asia and provide the first direct evidence to suggest pathological BT from the gut using non-invasive sampling and enrichment-based metagenomic sequencing in this high-risk transplant cohort.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eFunding and acknowledgements\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by a grant from the Bill and Melinda Gates Foundation (OPP1159351) and a Wellcome Senior Research Fellowship (215515/Z/19/Z) to Stephen Baker. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. We would like to thank Agila Kumari Pragasam, Chaitra Shankar, and Ellen Higginson for helpful discussions on the study design. We are also grateful to the media preparation team in the department of Clinical Microbiology at CMC Vellore, and Plamena Naydenova for her assistance with quantifying DNA concentrations. We would also like to thank Jolynne Mokaya for critical feedback on the manuscript. Finally, we would like to acknowledge the clinical teams for their assistance with sample collection and thank the patients and their families for their participation and support.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eEthics approval and consent to participate\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was granted ethical approval by the Institutional Review Board (IRB), Research and Ethics committee at CMC Vellore IRB\u0026nbsp;Min No. 14605 dated 27.04.2022. Patients were eligible if they were admitted to the Department of Haematology at CMC Hospital, scheduled to undergo an allogeneic stem cell transplant and provided written informed consent to participate. Informed consent forms were written in English and translated into Hindi and Tamil. Eligible patients (as per clinician\u0026rsquo;s assessment) or their legal representatives, were asked for informed consent, following which a clinical case report form was completed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eConsent for publication\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe authors report there are no competing interests to declare.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eConceptualisation and design: AM, SB, BV, SS, BG\u003c/p\u003e\n\u003cp\u003eCollection of samples and clinical data: SS, SD, BG\u003c/p\u003e\n\u003cp\u003eProcessing of samples: AM, SK, YM, DM, PS, KM\u003c/p\u003e\n\u003cp\u003eAnalysis and interpretation of data: AM, JJJ, SB\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDrafting initial paper: AM, JJJ, SB\u003c/p\u003e\n\u003cp\u003eRevising the final paper: AM, JJJ, SS, SK, YM DM, PS, KM, BG, BV, SB\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe authors confirm that the data supporting the findings of this study are available in the article and its supplementary materials. The genomic sequencing reads have been deposited in the NCBI Sequence Read Archive under BioProject accession PRJNA1072756 (https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA1072756).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHeston SM, Young RR, Hong H, Akinboyo IC, Tanaka JS, Martin PL, et al. Microbiology of Bloodstream Infections in Children After Hematopoietic Stem Cell Transplantation: A Single-Center Experience Over Two Decades (1997\u0026ndash;2017). Open Forum Infect Dis. 2020;7(11):ofaa465.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBarman P, Choudhary D, Chopra S, Thukral T. Blood stream infections in hematopoietic stem cell transplant patients: A 2-year study from India. Oncol J India. 2020;4(2):43.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCerf-Bensussan N, Gaboriau-Routhiau V. The immune system and the gut microbiota: friends or foes? Nat Rev Immunol. 2010;10(10):735\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrenchley JM, Douek DC. Microbial Translocation Across the GI Tract. Annu Rev Immunol. 2012;30(1):149\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShono Y, van den Brink MRM. Gut microbiota injury in allogeneic haematopoietic stem cell transplantation. Nat Rev Cancer. 2018.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSatlin MJ, Jenkins SG, Walsh TJ. The Global Challenge of Carbapenem-Resistant Enterobacteriaceae in Transplant Recipients and Patients With Hematologic Malignancies. Clin Infect Dis. 2014;58(9):1274\u0026ndash;83.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKoh AY. The microbiome in hematopoietic stem cell transplant recipients and cancer patients: Opportunities for clinical advances that reduce infection. Sheppard DC, editor. PLOS Pathog. 2017;13(6):e1006342.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePouch SM, Satlin MJ. Carbapenem-resistant Enterobacteriaceae in special populations: Solid organ transplant recipients, stem cell transplant recipients, and patients with hematologic malignancies. Virulence. 2017.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStecher B, Denzler R, Maier L, Bernet F, Sanders MJ, Pickard DJ, et al. Gut inflammation can boost horizontal gene transfer between pathogenic and commensal Enterobacteriaceae. Proc Natl Acad Sci. 2012;109(4):1269\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAnthony WE, Burnham CAD, Dantas G, Kwon JH. The Gut Microbiome as a Reservoir for Antimicrobial Resistance. J Infect Dis. 2021;223(Supplement3):S209\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSardzikova S, Andrijkova K, Svec P, Beke G, Klucar L, Minarik G, et al. Gut diversity and the resistome as biomarkers of febrile neutropenia outcome in paediatric oncology patients undergoing hematopoietic stem cell transplantation. Sci Rep. 2024;14(1):5504.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSatlin MJ, Walsh TJ. Multidrug-resistant Enterobacteriaceae, \u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e, and vancomycin-resistant \u003cem\u003eEnterococcus\u003c/em\u003e: Three major threats to hematopoietic stem cell transplant recipients. Transpl Infect Dis. 2017;19(6):e12762.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIskandar K, Molinier L, Hallit S, Sartelli M, Hardcastle TC, Haque M, et al. Surveillance of antimicrobial resistance in low- and middle-income countries: a scattered picture. Antimicrob Resist Infect Control. 2021;10(1):63.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhazal SS, Stevens MP, Bearman GM, Edmond MB. Utility of surveillance blood cultures in patients undergoing hematopoietic stem cell transplantation. Antimicrob Resist Infect Control. 2014;3:20.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMurray CJL, Ikuta KS, Sharara F, Swetschinski L, Robles Aguilar G, Gray A, et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet. 2022;399(10325):629\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eThe Genome Standards Consortium, Bowers RM, Kyrpides NC, Stepanauskas R, Harmon-Smith M, Doud D, et al. Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nat Biotechnol. 2017;35(8):725\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKorula A, Perumalla S, Devasia AJ, Abubacker FN, Lakshmi KM, Abraham A et al. Drug-resistant organisms are common in fecal surveillance cultures, predict bacteremia and correlate with poorer outcomes in patients undergoing allogeneic stem cell transplants. Transpl Infect Dis [Internet]. 2020 Jun [cited 2022 May 23];22(3). Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://onlinelibrary.wiley.com/doi/\u003c/span\u003e\u003cspan address=\"https://onlinelibrary.wiley.com/doi/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/tid.13273\u003c/span\u003e\u003cspan address=\"10.1111/tid.13273\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCollin BA, Leather HL, Wingard JR, Ramphal R. Evolution, Incidence, and Susceptibility of Bacterial Bloodstream Isolates from 519 Bone Marrow Transplant Patients. Clin Infect Dis. 2001;33(7):947\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSalas MQ, Charry P, Puerta-Alcalde P, Mart\u0026iacute;nez-Cibrian N, Solano MT, Serrahima A et al. Bacterial Bloodstream Infections in Patients Undergoing Allogeneic Hematopoietic Cell Transplantation With Post-Transplantation Cyclophosphamide. Transplant Cell Ther. 2022;28(12):850.e1-850.e10.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKr\u0026uuml;ger W, R\u0026uuml;ssmann B, Kr\u0026ouml;ger N, Salomon C, Ekopf N, Elsner HA, et al. Early infections in patients undergoing bone marrow or blood stem cell transplantation \u0026ndash; a 7 year single centre investigation of 409 cases. Bone Marrow Transpl. 1999;23(6):589\u0026ndash;97.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNinin E, Milpied N, Moreau P, Andr\u0026eacute;-Richet B, Morineau N, Mah\u0026eacute; B, et al. Longitudinal Study of Bacterial, Viral, and Fungal Infections in Adult Recipients of Bone Marrow Transplants. Clin Infect Dis. 2001;33(1):41\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhosh S, Bhattacharya S, Goel G, Deshmukh RA, Javed R, Roychowdhury M et al. Hematopoietic stem-cell transplantation in a zoo of multidrug‐resistant organisms: Data from a cancer center in eastern India. Transpl Infect Dis. 2023;e14072.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWong SP, Er YX, Tan SM, Lee SC, Rajasuriar R, Lim YAL. Oral and Gut Microbiota Dysbiosis is Associated with Mucositis Severity in Autologous Hematopoietic Stem Cell Transplantation: Evidence from an Asian Population. Transplant Cell Ther. 2023;29(10):633.e1-633.e13.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTamburini FB, Andermann TM, Tkachenko E, Senchyna F, Banaei N, Bhatt AS. Precision identification of diverse bloodstream pathogens in the gut microbiome. Nat Med. 2018;24(12):1809\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJazmati T, Hamprecht A, Jazmati N. Comparison of stool samples and rectal swabs with and without pre-enrichment for the detection of third-generation cephalosporin-resistant Enterobacterales (3GCREB). Eur J Clin Microbiol Infect Dis. 2021;40(11):2431\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNieman AE, Savelkoul PHM, Beishuizen A, Henrich B, Lamik B, MacKenzie CR, et al. A prospective multicenter evaluation of direct molecular detection of blood stream infection from a clinical perspective. BMC Infect Dis. 2016;16(1):314.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBalletto E, Mikulska M. Bacterial infections in haematopoietic stem cell transplant patients. Mediterr J Hematol Infect Dis. 2015;7:e2015045.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eD\u0026rsquo;Amico F, Soverini M, Zama D, Consolandi C, Severgnini M, Prete A, et al. Gut resistome plasticity in pediatric patients undergoing hematopoietic stem cell transplantation. Sci Rep. 2019;9(1):5649.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShepherd MJ, Fu T, Harrington NE, Kottara A, Cagney K, Chalmers JD, et al. Ecological and evolutionary mechanisms driving within-patient emergence of antimicrobial resistance. Nat Rev Microbiol. 2024;22(10):650\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHeston SM, Young RR, Jenkins K, Martin PL, Stokhuyzen A, Ward DV, et al. The effects of antibiotic exposures on the gut resistome during hematopoietic cell transplantation in children. Gut Microbes. 2024;16(1):2333748.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUritskiy GV, DiRuggiero J, Taylor J. MetaWRAP\u0026mdash;a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome. 2018;6(1):158.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOlm MR, Crits-Christoph A, Bouma-Gregson K, Firek BA, Morowitz MJ, Banfield JF. inStrain profiles population microdiversity from metagenomic data and sensitively detects shared microbial strains. Nat Biotechnol. 2021;39(6):727\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLewis SJ, Heaton KW. Stool Form Scale as a Useful Guide to Intestinal Transit Time. Scand J Gastroenterol. 1997;32(9):920\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKrakau S, Straub D, Gourl\u0026eacute; H, Gabernet G, Nahnsen S. nf-core/mag: a best-practice pipeline for metagenome hybrid assembly and binning. NAR Genomics Bioinforma. 2022;4(1):lqac007.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHyatt D, Chen GL, LoCascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics. 2010;11(1):119.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013;29(8):1072\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKang DD, Li F, Kirton E, Thomas A, Egan R, An H, et al. MetaBAT 2: an adaptive binning algorithm for robust and efficient genome reconstruction from metagenome assemblies. PeerJ. 2019;7:e7359.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu YW, Simmons BA, Singer SW. MaxBin 2.0: an automated binning algorithm to recover genomes from multiple metagenomic datasets. Bioinformatics. 2016;32(4):605\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlneberg J, Bjarnason BS, De Bruijn I, Schirmer M, Quick J, Ijaz UZ, et al. Binning metagenomic contigs by coverage and composition. Nat Methods. 2014;11(11):1144\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChaumeil PA, Mussig AJ, Hugenholtz P, Parks DH. GTDB-Tk: a toolkit to classify genomes with the Genome Taxonomy Database. Hancock J, editor. Bioinformatics. 2020;36(6):1925\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWood DE, Lu J, Langmead B. Improved metagenomic analysis with Kraken 2. Genome Biol. 2019;20(1):257.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLu J, Breitwieser FP, Thielen P, Salzberg SL. Bracken: estimating species abundance in metagenomics data. PeerJ Comput Sci. 2017;3:e104.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLu J, Rincon N, Wood DE, Breitwieser FP, Pockrandt C, Langmead B, et al. Metagenome analysis using the Kraken software suite. Nat Protoc. 2022;17(12):2815\u0026ndash;39.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFoster ZSL, Sharpton TJ, Gr\u0026uuml;nwald NJ, Metacoder. An R package for visualization and manipulation of community taxonomic diversity data. Poisot T, editor. PLOS Comput Biol. 2017;13(2):e1005404.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDugard P, Todman J, Staines H. Approaching Multivariate Analysis: A practical introduction [Internet]. 2nd ed. London: Routledge; 2022. 275 p. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.taylorfrancis.com/books/9781003343097\u003c/span\u003e\u003cspan address=\"https://www.taylorfrancis.com/books/9781003343097\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArbizu PM. pairwiseAdonis: pairwise multilevel comparison using adonis. 2017. 2019.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiaw A, Wiener M. Classification and Regression by randomForest. R News. 2002;2(3):18\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCsardi MG. Package \u0026lsquo;igraph\u0026rsquo;. Last Accessed. 2013;3(09):2013.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePedersen TL. tidygraph: A Tidy API for Graph Manipulation [Internet]. 2023. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/thomasp85/tidygraph\u003c/span\u003e\u003cspan address=\"https://github.com/thomasp85/tidygraph\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSi B, Liang Y, Zhao J, Zhang Y, Liao X, Jin H, et al. GGraph: An Efficient Structure-Aware Approach for Iterative Graph Processing. IEEE Trans Big Data. 2022;8(5):1182\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7763072/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7763072/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003ePatients undergoing haematopoietic stem cell transplantation (HSCT) in low- and middle-income countries face high rates of antimicrobial resistant (AMR) bloodstream infections (BSIs), but the origins and dynamics of these infections remain poorly understood.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eWe prospectively studied 81 HSCT patients in India, collecting 252 longitudinal stool samples and applying an enrichment-based metagenomic approach to selectively recover priority AMR pathogens. This strategy enabled high-resolution metagenome-assembled genomes and strain tracking. We identified a marked depletion of gut resistomes during pre-engraftment, followed by a rapid rebound with engraftment, driven by expansion of plasmid-borne carbapenemase and ESBL genes (\u003cem\u003eblaNDM\u003c/em\u003e, \u003cem\u003eblaOXA\u003c/em\u003e). Using whole genome sequencing, Hi-C metagenomics, and strain-level comparisons (\u0026gt;\u0026thinsp;99.9% ANI), we directly linked gut-colonising organisms to five culture-confirmed AMR-BSI episodes, providing genomic evidence of gut translocation. Patients with BSIs had high mortality (\u0026gt;\u0026thinsp;40%), underlining the clinical impact of these events.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eWe demonstrate that stool-based enrichment metagenomics is a practical and cost-effective approach for non-invasive monitoring of gut recovery and AMR risk after HSCT. Our findings provide the first direct genomic evidence of gut-derived AMR-BSIs in an LMIC cohort, highlighting translocation as a major driver of post-transplant mortality and a critical target for surveillance and intervention.\u003c/p\u003e","manuscriptTitle":"Gut translocation of antimicrobial resistant pathogens in patients undergoing haematopoietic stem cell transplantation in India","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-21 23:41:04","doi":"10.21203/rs.3.rs-7763072/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d4e104a0-0d74-4e7c-9c22-e43d3cb514e4","owner":[],"postedDate":"October 21st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-01T07:24:16+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-21 23:41:04","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7763072","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7763072","identity":"rs-7763072","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}