Putative metabolite annotation from the Human Metabolome Database (HMDB) was performed. To generate higher confidence metabolite identifications two approaches were taken: (i) experimental MS/MS data were matched to the mzCloud database, using Compound Discoverer software (ThermoFisher v4.2.18, with a match score of >60) and (ii) retention times and/or MS/MS spectra were matched to authentic chemical standards (retention times ±22 seconds, dot product cosine score >0.6). Where multiple ion forms existed for a single metabolite, we show data for the most abundant ion form with the lowest number of missing values (typically [M+H]+ or [M-H]-, indicated in the paper). The metabolite annotation grade (according to the Metabolomics Standards Initiative, MSI) is indicated for each annotated compound in the results tables.
"],"repository":["MetaboLights"],"study_status":["Public"],"ptm_modification":[""],"instrument_platform":["Liquid Chromatography MS - positive - hilic","Liquid Chromatography MS - negative - hilic"],"chromatography_protocol":["Analysis were performed using a Thermo Scientific Dionex Ultimate 3000 UHPLC system. A hydrophilic interaction chromatography (HILIC) based Ultra-High Performance Liquid Chromatography-Mass Spectrometry (UHPLC-MS) assay was applied as used previously[1], using an Accucore-150-Amide-HILIC column (100 x 2.1 mm, 2.6 μm, Thermo Fisher Scientific, MA, USA). Mobile phase A was 95% acetonitrile/water (10 mM ammonium formate, 0.1% formic acid); mobile phase B was 50% acetonitrile/water (10 mM ammonium formate, 0.1% formic acid); and the gradient was t=0.0, 1% B; t=1.0, 1% B; t=3.0, 15% B; t=6.0, 50% B; t=9.0, 95% B; t=10.0, 95% B; t=10.5, 1% B; t=14.0, 1% B. All changes were linear (curve = 5) and the flow rate was 0.50 mL/min. Column temperature was 35 °C and injection volume was 2 μL.
Ref:
[1]
"],"publication":["Regional HNSCC metabolomics reveals widespread changes to one-carbon metabolism and S-adenosylmethionine metabolism across tumour core, tumour edge and adjacent non-tumour tissues."],"submitter_name":["Ossama Edbali","Andrew Southam","Gavin Lloyd","Ralf Weber"],"submitter_affiliation":["University of Birmingham","Phenome Centre Birmingham, University of Birmingham"],"organism_part":["blank","squamous cell cancer of the head and neck"],"technology_type":["mass spectrometry assay"],"disease":[""],"extraction_protocol":["Tissue was extracted using a biphasic methanol/chloroform/water approach. Tissue sample sizes were between 10 – 41.3 mg (with one extra small sample of 6.5 mg). Frozen tissue was homogenised with ice-cold methanol (15 µL/mg wet tissue mass) and ice-cold water (6 µL/mg) using a bead-based homogeniser (Precellys24 and CK14 tubes, Stretton Scientific, UK) for 2 x 10 s bursts of 6400 rpm. The homogenate was transferred to a 2 mL glass vial and ice-cold chloroform (15 µL/mg) and water (7.5 µL/mg) added. Sample was vortexed (60 s), incubated (on ice, 10 min), centrifuged (2,500-g, 10 min, 4 °C) and set at room temperature (5 min) to allow biphase partitioning. A fixed volume of the upper biphasic layer containing the extracted polar metabolites was taken from all samples (300 µL – equivalent to 10 mg extracted tissue) except for the small 6.5 mg sample (here 150 µL was taken – equivalent to 5 mg extracted tissue). Samples were dried using a SpeedVac Concentrator (Savant SPD111V230, Thermo Fisher Scientific). Dried polar extracts were resuspended in 100 μL of 3:1 acetonitrile:water (except for the small sample, which was resuspended in 50 μL to ensure an equivalent final concentration to the other samples), vortexed (30 s), centrifuged (21,000-rcf, 20 min, 4 °C) and 50 μL of the supernatant was aliquoted into a low recovery volume HPLC vial (Chromatography Direct, UK). Vials were set at 4 °C until and during UHPLC-MS analysis.
"],"organism":["blank","Homo sapiens"],"full_dataset_link":["https://www.ebi.ac.uk/metabolights/MTBLS1905"],"author":["Lauren Cruchley-Fuge.","Matthew Smith.","Ossama Edbali. o.edbali@bham.ac.uk.","Nikolaos Batis.","Gavin Lloyd. g.r.lloyd@bham.ac.uk.","James Higginson.","Southam Andrew. Phenome Centre Birmingham & School of Biosciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK. University of Birmingham, United Kingdom. a.d.southam@bham.ac.uk.","Weber Ralf. r.j.weber@bham.ac.uk.","Rachel Spruce.","Hisham Mehanna."],"data_transformation_protocol":["Processing of the MS1 data was carried out as follows: vendor format raw data files (.RAW) were converted to the mzML file format using ProteoWizard software. Deconvolution was performed by XCMS software6 (version 1.46.0 running in the Galaxy environment) applying min peak width (6); max peak width (30); ppm (14); mzdiff (0.001); bw (0.25); mzwid (0.01); minfrac (0.5). A data matrix of peak areas for metabolite features (m/z-retention time pairs) vs. samples was constructed. Features were retained in the data matrix if they were: present in >90% of QC samples; had a peak intensity relative standard deviation (RSD)<30% across QC samples; and had a mean QC/extract blank intensity ratio of >20% and present in >50% of the samples. Samples were excluded if more than 50% of the features were missing – only one example of this existed – 1x non-tumour (N) sample was removed from the UHPLC-MS HILIC negative ion data set. All signal correction and filtering steps were applied using R version 4.1.18 and the structToolbox package. To evaluate and correct for systematic mass error across the m/z range, 80 known reference features consistently present across the study samples were used. These included sodium formate clusters and endogenous metabolites detected in >80% of samples—such as riboflavin, 9-HODE, guanosine, fructose-6-phosphate, and L-tyrosine—spanning an m/z range of 128 to 1000. A cubic smoothing spline was first fitted to the m/z error as a function of m/z using a randomly selected 50% subset of reference features, to correct for variations in instrument accuracy across the mass range. The model was then validated on the remaining 50% of reference features. The same procedure was repeated for m/z error as a function of injection order to correct for temporal mass drift. After validating both models, the corrections were applied sequentially to the entire dataset (HILIC negative ion data ONLY), including precursor m/z values from LC-MS/MS spectra used for metabolite annotation.
Refs:
[1] Lloyd, G. R.; Jankevics, A.; Weber, R. J. M. struct: an R/Bioconductor-based framework for standardized metabolomics data analysis and beyond. Bioinformatics 2021, 36 (22-23), 5551-5552. DOI: 10.1093/bioinformatics/btaa1031 From NLM.
[2] Lloyd, G. R.; Weber, R. J. M. structToolbox: Data processing & analysis tools for Metabolomics and other omics. Bioconductor 2020. DOI: 10.18129/B9.bioc.structToolbox.
"],"study_factor":["Tissue type"],"submitter_email":["r.j.weber@bham.ac.uk","o.edbali@bham.ac.uk","g.r.lloyd@bham.ac.uk","a.d.southam@bham.ac.uk"],"sample_collection_protocol":["Samples were collected from 22 patients undergoing surgical resection of head and neck cancer. Representative samples were taken from the central core of the tumour (C), from the visible tumour edge (E), and from matched normal mucosa (N) at least 5cm from the tumour site. All patients gave informed consent to participate in this study (ethical approval 16/NW/0265). Samples were immediately flash frozen and stored at -80 °C.
"],"omics_type":["Metabolomics"],"study_design":["ultra-performance liquid chromatography-mass spectrometry","pooled quality control sample","blank","squamous cell cancer of the head and neck","untargeted analysis","Spatial tumour location","Advanced Head and Neck Squamous Cell Carcinoma","Homo sapiens","sample preparation blank","experimental sample","Thermo Scientific Dionex Ultimate 3000 UHPLC system","untargeted metabolites","Thermo Scientific Q Exactive Focus","S-adenosyl-L-methionine"],"curator_keywords":["ultra-performance liquid chromatography-mass spectrometry","pooled quality control sample","blank","squamous cell cancer of the head and neck","untargeted analysis","Spatial tumour location","Advanced Head and Neck Squamous Cell Carcinoma","Homo sapiens","sample preparation blank","experimental sample","Thermo Scientific Dionex Ultimate 3000 UHPLC system","untargeted metabolites","Thermo Scientific Q Exactive Focus","S-adenosyl-L-methionine"],"mass_spectrometry_protocol":["Analysis was performed using a Thermo Scientific Q Exactive Focus. Data were acquired separately in positive and negative ionisation modes (70 – 1050 m/z) with a resolution of 70,000 (FWHM at m/z 200). Ion source parameters: Sheath gas = 53 arbitrary units, Aux gas = 14 arbitrary units, Sweep gas = 3 arbitrary units, Spray Voltage = 3.5kV (positive ion) / 2.7kV (negative ion), Capillary temp. = 269 °C (positive ion) / 320 °C (negative ion), Aux gas heater temp. = 438°C (positive ion) / 320 °C (negative ion). Quality control (QC) samples were analysed as column equilibration samples (the first eight injections).
QC samples were recorded throughout the run to check for analytical quality: a further 2 QC samples at the beginning of the run, then a QC analysed every seventh injection and 2 QCs at the end of the analytical batch. Extract blank samples was analysed as the 3rd and last injections of the analytical batch. For the purpose of metabolite annotation, data dependent MS2 in ‘Discovery mode’ was applied to five QC samples over five mass ranges (70 - 120 m/z; 120-170 m/z; 170-220 m/z; 220-270 m/z; 270-1050 m/z) using following settings: MS1 resolution = 35,000, MS2 resolution = 17,500; Isolation width = 3.0 m/z; stepped collision energies = 25, 60, 100 eV.
"],"metabolite_name":["adenosine","carnosine","1-methyladenosine","5'-S-methyl-5'-thioadenosine","carnitine"],"additional_accession":[]},"is_claimable":false,"name":"Regional HNSCC metabolomics reveals widespread changes to one-carbon metabolism and S-adenosylmethionine metabolism across tumour core, tumour edge and adjacent non-tumour tissues","description":"Cancer, including head and neck squamous cell carcinoma (HNSCC), induces changes to metabolism that drive the disease. Regional metabolomics can help to understand metabolic variation across the tumour including changes near the tumour core, where hypoxia is likely more pronounced. We apply ultra-high performance liquid chromatography-mass spectrometry metabolomics to regionally distinct patient tissue samples: tumour edge, tumour core and adjacent non-tumour. Statistical, correlation and pathway enrichment analyses were performed. Markers of hypoxia or pseudohypoxia—lactate, succinate, fumarate, and the lactate:pyruvate ratio—were elevated in both core and edge tumour regions relative to adjacent tissue, with a trend toward stronger changes in the core. One-carbon metabolites were altered in HNSCC, including tumour-associated increases of S-adenosylmethionine (SAM) and SAM metabolites (S-adenosylhomocysteine, polyamines, methylated nucleosides, dimethylarginine, trimethylysine and 1-methylnicotinamide). Histidine, tryptophan, choline and folate appear metabolically connected to one-carbon metabolism in HNSCC: histidine, L-kynurenine (tryptophan metabolite), some purine metabolites (including deoxyguanosine, deoxyinosine) and choline were elevated in tumour tissue; while histidine/SAM, L-kynurenine/deoxyguanosine, L-kynurenine/deoxyinosine and folate/methionine were correlated in tumour tissue only. Tumour edge and core exhibited similar one-carbon metabolic changes relative to non-tumour, but the magnitude of change was generally greater in the core reflecting location dependent variation of SAM metabolism in HNSCC.
","dates":{"publication":"2026-06-03","submission":"2026-06-02"},"accession":"MTBLS1905","cross_references":{"MetaboLights":["MTBLC16027","MTBLC15676","MTBLC27470","MTBLC16020","MTBLC17509","MTBLC16335","MTBLC17126","MTBLC15727"],"ChEBI":["CHEBI:16027","CHEBI:15676","CHEBI:27470","CHEBI:16020","CHEBI:17509","CHEBI:16335","CHEBI:17126","CHEBI:15727"]}}