MetaboLights

application/xml

ftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481/m_MTBLS12481_LC-MS_negative_reverse-phase_metabolite_profiling_v2_maf.tsvftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481/a_MTBLS12481_LC-MS_negative_reverse-phase_metabolite_profiling.txtftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481/s_MTBLS12481.txtftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481/i_Investigation.txtftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481/FILES/milkHydro.FA.rawdata.noOuter.0toNA.rawftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481/FILES/brainHydro.FA.rawdata.noOuter.0toNA.rawprimary200OKftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS12481Mass spectrometry peak annotation included the following steps. First, we constructed a custom database of theoretical [M–H]- m/z values for fatty acids, varying chain lengths (C6–C30) and degrees of unsaturation (0–6 double bonds). Next, the measured peaks were matched against this database using a 10 ppm threshold. Annotation was therefore based on accurate mass only, with no confirmation by MS/MS spectra or authentic standards; thus, identifications correspond to MSI Level 3 (putative class identification). To validate annotations, we examined the systematic relationship between chain length, degree of unsaturation, m/z, and retention time: longer chains produced higher m/z and longer retention, while increasing double bonds produced lower m/z and shorter retention. This behavior was visualized in m/z–RT scatter plots to confirm the concordance with the overall pattern within the FA class. Two fatty acids in milk, C6:0 and C8:0, showed mass errors above 10 ppm (14.4 and 10.2, respectively), but were retained because they were consistently detected across replicates and species and aligned with the expected m/z–RT pattern on the plot. The high errors likely reflect calibration limitations at low m/z, where relative ppm deviation is more pronounced. NOTE: SMILES and InChI strings provided in 'Metabolites' table represent generic straight-chain fatty acid structures. For unsaturated fatty acids, double-bond positions were assigned to representative configurations (Δ9; Δ9,12; Δ9,12,15; Δ5,8,11,14; etc.) according to common metabolomics conventions. These assignments are illustrative only and do not resolve true double-bond positions or isomeric forms. All identifications are therefore reported at Metabolomics Standards Initiative (MSI) Level 3 (putative class identification, isomeric mixture).MetaboLightsPublicLiquid Chromatography MS - negative - reverse phaseSamples (3 µL) were separated on a Waters Acquity UPLC HSS T3 reverse-phase column (100 mm x 2.1 mm, 1.8 µm) with a matching Vanguard pre-column. The mobile phases used for the chromatographic separation were: Buffer A (H2O, 10 mM ammonium acetate, 0.1% formic acid) and Buffer B (acetonitrile:isopropanol 7:3, 10 mM ammonium acetate, 0.1% formic acid). The gradient separation was as follows: 1 min 55 % B, 3 min linear gradient from 55 % to 80 % B, 8 min linear gradient from 80 % to 85 % B, and 3 min linear gradient from 85 % to 100 % B. After a 4.5 min wash with 100 % B the column was re-equilibrated with 55 % B. Flow rate was 400 µL/min, column temperature 40 °C.Role of breast milk lipid composition in postnatal brain development.The Hospital for Sick ChildrenAleksandra Mitinamilkbrainmass spectrometryExtractionLipids were extracted using a modified two-phase MTBE/MeOH protocol (Sarafian, 2014), with all steps performed on ice. For milk samples, 750 µL of MeOH:MTBE (1:3, v/v) containing 1 mg/L TAG 15:0–18:1-d7–15:0 (Avanti 791648C) was added prior to extraction. Samples were vortexed for 1 min, sonicated for 15 min, and incubated at 4 °C for 30 min, followed by addition of MeOH:H2O (1:3, v/v), vortexing and centrifugation (10 min, 14.000 × g, 4 °C). The upper organic phase (400 µL) was collected, dried under vacuum (30 °C, 1 h), and reconstituted in acetonitrile:isopropanol (1:3) for hydrolysis.For brain samples, 500 µL of MeOH:MTBE (1:3, v/v) containing 1 mg/L 18:1-d7 lyso-PC (Avanti 791643C) was added prior to extraction. Extraction was performed using the same procedure as for milk samples.HydrolysisDried lipid extracts were hydrolyzed to free fatty acids (FFAs) in 100 µL MeOH:6% KOH (4:1, v/v), incubated for 2 h at 60 °C with shaking, cooled, 100 µL of saturated NaCl was added, neutralized with 50 µL 29% HCl followed by 1 min vortexing, and extracted twice with 200 µL chloroform:heptane (1:4, v/v) (Bromke, 2015). The organic phases were washed and dried under a high-speed vacuum for 1 hour. Dried lipids were resuspended in 100 µL of acetonitrile:isopropanol (1:3, v/v), vortexed for 10 sec, shaken for 10 min at 4 °C, and sonicated in an ice bath. 50 µL of each sample was transferred into glass autosampler vial for LC–MS analysis. Alkaline methanolic hydrolysis cleaves ester bonds (e.g., triacylglycerols, phospholipids, cholesteryl esters), whereas ether- and amide-linked species (plasmalogens, sphingolipids) are less efficiently hydrolyzed and may therefore be under-represented. Hydrolysis efficiency was monitored by the release of deuterated fatty acids from internal standards (TAG 15:0–18:1(d7)–15:0 in milk, LPC 18:1(d7) in brain), which showed consistent recovery across batches; any contribution of unlabeled fatty acids from standards was subtracted from the corresponding sample peaks.Capra hircusBaby formulaBos grunniensMacaca fascicularisHomo sapiensMacaca mulattaPan troglodytesBos taurusSus scrofahttps://www.ebi.ac.uk/metabolights/MTBLS12481Aleksandra Mitina. SickKids Research Institute. sasha.mitina@sickkids.ca.Philipp Khaitovich. Skolkovo Institute of Science and Technology. p.khaitovich@skoltech.ru. Vendor files were converted to .mzML format with ProteoWizard (55) (v3.0.18372; peak picking: Vendor, msLevel = 1; full time range). For parameter optimization, IPO (56) (v3.5) was run on QC files (one per batch). Optimized XCMS (57) (v3.4.4) parameters were: peakwidth = c(7.5, 34), ppm = 12, noise = 1e6, snthresh = 100, minfrac = 0.1. Files were organized by batch; alignment used the center samples MS522 (milk) and MS200 (brain). CAMERA (58) (v1.33.3) was applied after XCMS. Feature tables contained 1,250 (milk) and 211 (brain) monoisotopic peaks. Features with blank intensity >10% of real sample intensity were removed as contaminants; only features with CV <30% in QC were retained. Analyses were performed in R 3.4.0.SexAgealeksandra.akhmadullina@gmail.com Human milk Breast milk was collected at two locations: from an Eastern European population (n = 64 participants, Moscow, Russia) and from an East Asian population (n = 87 participants, Shanghai, China). Each participant collected ~5 mL of milk once per week in 10 mL containers at the end of breastfeeding. Samples were stored at –20 °C for one week, then transferred to –80 °C for long-term storage. In total, 297 samples were collected in Moscow and 291 in Shanghai. Metadata included stage of lactation (days postpartum), parity, sex of the child, mode of delivery, and maternal and neonatal anthropometrics. All procedures were approved by the Institutional Research Ethics Boards of the Skolkovo Institute of Science and Technology (Moscow, Russia), and written informed consent was obtained from all participants. Animal milk Cow, goat, pig, and yak milk was collected from private farms in China and Russia under written agreements with the owners (no financial compensation). Rhesus and crab-eating macaque milk was collected from breeding facilities in China in accordance with ethical regulations. In total, 66 cow and 35 goat samples were collected in Russia, and 49 pig, 37 cow, 15 yak, 23 rhesus, and 22 crab-eating macaque samples were collected in China. For animals, metadata included offspring birth date and sampling date. Animal milk collection was carried out by trained staff in ways that minimized stress. All human and animal milk samples are available upon reasonable request. Human brain Prefrontal cortex (PFC) and cerebellum (CB) samples (n = 92) were obtained from the Chinese Brain Bank Center (CBBC), representing African American, Caucasian, Chinese, and Hispanic populations. Samples were collected postmortem from infants aged newborn to one year, with informed consent from donors or next of kin. Metadata included self-reported ethnicity, age, and biological sex. Reported causes of death for these infant donors were accidental (car accidents, falls, and other trauma) rather than neurological or metabolic disease, which reduces the likelihood of systematic bias in lipid composition, although perimortem stress cannot be entirely excluded. Non-human primate and animal brain. Chimpanzee brain samples (n = 18) were obtained from the Max Planck Institute for Evolutionary Anthropology (Leipzig, Germany), and rhesus macaque samples (n = 38) from the Yunnan Key Laboratory of Primate Biomedical Research (China). Metadata included age (in days) and sex. All non-human primates died of causes unrelated to the study and were obtained legally and were handled in accordance with the regulations of the providing institutions. Goat (n = 26) and pig (n = 20) brain samples were obtained as by-products from meat production on private farms in Russia, under written agreements with owners with no financial compensation provided, and in compliance with local and national animal welfare regulations. All brain samples are available upon reasonable request.MetabolomicsFatty Acidsbreast milkuntargeted metabolitescerebral gray matterPrefrontal CortexFatty Acidsbreast milkuntargeted metabolitescerebral gray matterPrefrontal CortexMass spectra were acquired in negative ionization mode on a QExactive Hybrid Quadrupole-Orbitrap (Thermo Scientific) with a heated ESI source. Settings: spray voltage 3 kV; S-lens RF level 70; capillary 250 °C; aux gas heater 350 °C; sheath gas 45 a.u.; aux gas 10 a.u.; sweep gas 4 a.u. Full-scan resolution was 70,000 (at m/z 200); AGC target 1x10^6; max fill time 50 ms; scan range m/z 100–1500. Fatty acids were detected as deprotonated molecules [M–H]-.QC samples were injected every 12th sample, EQC every 24th, with blanks at the start, midpoint, and end of each queue.Samples were queued in the same 96-sample batch structure as extraction (milk: 9 complete batches + 1 short; brain: 2 batches). Each queue began with 8 blanks (acetonitrile:isopropanol) and 6 QC injections to equilibrate the column. QC was injected after every 12th sample, EQC after every 24th sample; queues ended with 5 blanks and 8 washes. Finally, 12 extraction blanks were injected followed by 2 washes.falseRole of breast milk lipid composition in postnatal brain development.Lipids in the brain play a crucial role as structural and signaling molecules, accounting for approximately 60% of the brain’s dry weight. Several lipid classes exhibit concentration profile differences between humans and other species, including closely related primates, with the most pronounced differences occurring during the first year of life. The long-chain polyunsaturated fatty acids (LC-PUFAs) that accumulate specifically in the brain are largely derived from nutrition – primarily breast milk during infancy. In this study, we analyze 837 milk samples from seven mammalian species and 194 brain samples from four mammalian species across two brain regions using liquid chromatography coupled with mass spectrometry (LC-MS). We show that differences in brain fatty acid composition are reflected in milk fatty acid composition and identify human-specific lipids present in both breast milk and the brain. Our findings highlight the influence of milk composition on brain development and suggest its potential role in shaping the unique cognitive features of the human brain. 2025-09-252025-05-15MTBLS12481