Tentative peak identifications were performed by comparing mass spectra against the NIST v23 mass spectral library. Only compounds with a mass spectral match factor ≥ 75 % were selected, with annotations listed as the compound name of the top match score. 3-methylbutanal and 3-methylbutanol were confirmed with authentic standards.
"],"repository":["MetaboLights"],"study_status":["Public"],"ptm_modification":[""],"instrument_platform":["Gas Chromatography MS - positive"],"chromatography_protocol":["Data acquisition
Probes were dry purged with N2 at 50 mL min-1 for 4 min to remove water residue then analysed by TD-GC-MS. Samples were randomised to counteract sampling batch effects and a gaseous internal standard (1 ppmV p-bromofluorobenzene in N2; Thames Restek, UK) was spiked onto each sample prior to desorption. VOCs were thermally desorbed at 280 °C for 10 min (TD100, Markes International, UK) then transferred to a cryofocusing trap maintained at 0 °C, which was flash heated to 300 °C for 2 min. Samples were injected into the GC with a 1:10 split ratio. VOC separation was performed on an Agilent 7890B GC (Agilent Technologies, UK) using an Agilent DB-5ms column (30 m x 0.25 mm x 0.25 μm) with constant helium flow (1 mL min−1). The GC column oven was set to a linear temperature ramp programme with an initial temperature of 35 °C, increasing at 7.5 °C min−1 to 260 °C.
"],"publication":["Leucine degradation metabolite ratio as a measure of meropenem susceptibility in carbapenem-resistant Klebsiella pneumoniae."],"submitter_name":["Breanna Dixon"],"submitter_affiliation":["University of Manchester"],"organism_part":["Reference Standards","Headspace","headspace extraction","media"],"technology_type":["mass spectrometry assay"],"disease":[""],"extraction_protocol":["Vials were crimp sealed and high capacity sorptive extraction probes (HiSorb™, Markes International, UK) with a polydimethylsiloxane (PDMS) sorbent coating inserted into the headspace through the vial septa. Headspace was passively collected over a 6 h incubation period at 37 °C with 180 rpm agitation, after which the probes were removed and inserted into cleaned empty stainless steel TD tubes for analysis.
"],"organism":["reference standard","media","Klebsiella pneumoniae"],"full_dataset_link":["https://www.ebi.ac.uk/metabolights/MTBLS13539"],"author":["Stephen Fowler.","Drupad Trivedi.","Breanna Dixon. breanna.dixon@manchester.ac.uk.","Tim Felton.","Waqar Ahmed. University of Manchester. waqar.ahmed@manchester.ac.uk.","Kamila Schmidt."],"data_transformation_protocol":["Spectral deconvolution was performed using MassHunter Quantitative Analysis software (Agilent Technologies, UK) with a retention window size factor of 100 and delta m/z tolerance of 0.3 AMU left/0.7 AMU right. Detected peaks were aligned using a retention time window of ± 0.1 min.
All subsequent data analysis was performed in R (v4.3.1). Integrated peak areas were normalised against the p-bromofluorobenzene internal standard. Peaks resulting from suspected environmental contaminants and artefacts were removed from analysis e.g. siloxanes and phthalate-derived compounds. For missing values, a nominal value of 1/5 of the minimum positive value was assigned to each variable. For statistical analysis, data were log transformed and autoscaled with package ‘mdatools’ [29]. Univariate analysis was performed using two-way t-tests. The package ‘Hmisc’ was used for correlation analysis with Pearson method, and ‘pROC’ was used for ROC analysis [30], [31]. Mean feature ratios and corresponding standard errors were calculated from individual replicate data.
"],"study_factor":["Resistance","Treatment"],"submitter_email":["breanna.dixon@manchester.ac.uk"],"sample_collection_protocol":["Bacterial isolates
A total of 16 pathogenic Klebsiella pneumoniae isolates were used in the study (CSKP; n = 7, CRKP; n = 9). All strains were clinical isolates sourced from the North Bristol NHS Trust and Manchester University NHS Foundation Trust except for five clinical reference strains obtained from the National Collection of Type Cultures (UK Health Security Agency, UK): NCTC 13438, NCTC 13442, NCTC 13443, ATCC 13883, AND ATCC 13887.
Headspace sampling of bacterial cultures with meropenem disks
For each of the 16 isolates, a 1 µL loop of overnight culture was inoculated in 2 mL tryptic soy broth (TSB) within a 10 mL headspace vial containing a 10 µg meropenem disk (n = 3). Meropenem-free set ups were prepared in parallel (n = 3).
Meropenem susceptibility testing using VOC analysis
Overnight cultures of four isolates (KP007, KP019, KP025 and KP030) were diluted to 0.1 OD600 in TSB, and 100 µL of each dilution was combined with 900 µL meropenem-supplemented TSB media within a 10 mL headspace vial. Isolates were selected based on their MIC value. Three levels of meropenem concentration were prepared for each isolate: 0 µg mL-1, 1µg mL-1, and 16 µg mL-1. Three replicates of each isolate were grown in each concentration. Headspace was passively collected using HiSorb™ probes over a 6 h incubation period at 37 °C with 180 rpm agitation.
Specificity testing of diagnostic metric
For specificity testing, isolates KP003, KP009, KP031 and KP034 were treated with 10 µg gentamicin and 5 µg ciprofloxacin disks. Headspace analysis was performed as before.
d10-Leucine supplementation
TSB was supplemented with d10-leucine at a concentration equivalent to its native leucine content (1.43 mg mL-1). Two K. pneumoniae isolates, KP003 (CSKP) and KP007 (CRKP), were grown in TSB with and without d10-leucine. Headspace sampling of bacterial cultures exposed to 10 µg meropenem disks was conducted as described earlier.
After GC separation, VOCs were transferred to an Agilent 7010 MS with EI+ source set to 70 eV and 230 °C, and a triple quadrupole mass analyser in full scan mode across a range of 40 – 300 m/z with an acquisition rate of 5 Hz.
"],"metabolite_name":["Dimethyl trisulfide","Glycerol 1,2-diacetate","2-Pentadecanone","2-Acetylthiazole"],"additional_accession":[]},"is_claimable":false,"name":"Leucine degradation metabolite ratio as a measure of meropenem susceptibility in carbapenem-resistant Klebsiella pneumoniae","description":"Rapid detection of carbapenem-resistant Klebsiella pneumoniae (CRKP) is essential for effective clinical management and surveillance. This study investigated volatile organic compounds (VOCs) as phenotypic biomarkers of carbapenem resistance, with the goal of developing a translationally viable diagnostic metric. Sixteen clinical K. pneumoniae isolates comprising both CRKP and susceptible (CSKP) strains, were cultured for six hours in headspace vials and VOCs extracted using PDMS-coated probes, with and without meropenem. Volatiles were analysed by thermal desorption-gas chromatography-mass spectrometry. Under meropenem stress, CRKP and CSKP exhibited distinct VOC profiles, with two leucine-derived metabolites, 3-methyl-1-butanol (syn. isoamyl alcohol) and 3-methylbutanal (syn. isovaleraldehyde), emerging as key discriminatory features. The ratio of these two metabolites accurately classified resistance phenotype, achieving 100% sensitivity and 94.1% specificity. The ratio also correlated significantly with minimum inhibitory concentration and zone of inhibition values. Stable isotope tracing confirmed their origin from leucine catabolism, and metabolic supplementation experiments suggested a functional role for 3-methyl-1-butanol in promoting growth under antibiotic stress. These findings show that antimicrobial resistance is accompanied by detectable shifts in volatile metabolism and that VOCs can serve as reliable indicators of resistance phenotype. The 3-methyl-1-butanol/3-methylbutanal ratio presents a robust biomarker of meropenem resistance, detectable within six hours of culture growth, and supports the further development of volatilome-based diagnostics for clinical microbiology.
","dates":{"publication":"2026-06-09","submission":"2025-12-19"},"accession":"MTBLS13539","cross_references":{"MetaboLights":["MTBLC173474","MTBLC89254","MTBLC4614","MTBLC173246"],"ChEBI":["CHEBI:173474","CHEBI:89254","CHEBI:4614","CHEBI:173246"]}}