Untargeted mass spectrometry is a robust tool for biology, but it usually requires a large amount of time on data analysis, especially for system biology. A framework called Multiple-Chemical nebula (MCnebula) was developed herein to facilitate the LC–MS data analysis process by focusing on critical chemical classes and visualization in multiple dimensions. This framework consists of three vital steps as follows: (1) abundance-based classes (ABC) selection algorithm, (2) critical chemical classes to classify “features” (corresponding to compounds), and (3) visualization as multiple Child-Nebulae (network graph) with annotation, chemical classification, and structure. Notably, MCnebula can be used to explore the classification and structural characteristic of unknown compounds beyond the limit of the spectral library. Moreover, it is intuitive and convenient for pathway analysis and biomarker discovery because of its function of ABC selection and visualization. MCnebula was implemented in the R language. A series of tools in R packages were provided to facilitate downstream analysis in an MCnebula-featured way, including feature selection, homology tracing of top features, pathway enrichment analysis, heat map clustering analysis, spectral visualization analysis, chemical information query, and output analysis reports. The broad utility of MCnebula was illustrated by a human-derived serum data set for metabolomics analysis. The results indicated that “Acyl carnitines” were screened out by tracing structural classes of biomarkers, which was consistent with the reference. A plant-derived data set was investigated to achieve a rapid annotation and discovery of compounds in E. ulmoides.

Lipids were separated by a C18 reverse phase column, using an Ascentis® Express 90 Å C18 HPLC column (15 cm x 2.1 mm; 2.7 μm, Supelco®) inserted into an HPLC system (Ultimate 3000 Dionex, Thermo Fisher Scientific, Bremen, Germany) with an autosampler coupled online to a Q-Exactive™ Hybrid Quadrupole-Orbitrap™ Mass Spectrometer (Thermo Fisher Scientific, Bremen, Germany). A volume of 5 μL of each sample mixture was injected into the HPLC column, at a flow rate of 260 μL/min. The temperature of the column oven was maintained at 50 °C. Elution started with 32% of mobile phase B, and the gradient used was: 45% B (1.5 min), 52% B (4 min), 58% B (5 min), 66% B (8 min), 70% B (11 min), 85% B (14 min), 97% B (18 min, maintained for 7 min), and 32% B (25.01 min, followed by a re-equilibration period of 8 min prior next injection).The Q-Exactive™ orbitrap mass spectrometer with a heated electrospray ionization source was operated in the positive mode (electrospray voltage of 3.0 kV) and negative mode (electrospray voltage of -2.7 kV). The sheath gas flow was 35 U, the auxiliary gas was 3 U, the capillary temperature was 320 °C, the S-lenses RF was 50 U and the probe`s temperature was 300 °C. Full scans MS spectra were acquired both in positive and negative modes in an m/z range of 300-1600, with a resolution of 70.000, automatic gain control (AGC) target of 3x10E6 and maximum injection time of 100 ms. For tandem MS (MS/MS) experiments, a top-10 data-dependent method was used. The top 10 most abundant precursor ions in full MS were selected to be fragmented in the collision cell HCD, with the dynamic exclusion of 30 s and an intensity threshold of 8x10E4. The MS/MS spectra were obtained with a resolution of 17,500; an AGC target of 1x10E5; an isolation window of 1 m/z; and a maximum injection time of 100 ms. A stepped normalized collision energy™ scheme was used and ranged between 25 and 30 eV for the positive ion mode and between 20, 24 and 28 for the negative ion mode. The MS/MS spectra obtained were those combining the information obtained with the different collision energies applied to each ionization mode. Data acquisition was carried out using the Xcalibur data system (V3.3, Thermo Fisher Scientific, USA).PC, LPC and SM were analysed in the LC-MS spectra in the positive ion mode, as [M+H]+ ions. The presence of the fragment ion at m/z 184, corresponding to the phosphocholine polar head group, in the MS/MS of [M+H]+ ions allows identifying PL molecular species belonging to the PC, LPC and SM classes, which were further differentiated by the characteristic retention times. The identification of PC, LPC and SM classes was confirmed in the LC-MS spectra in the negative ion mode, as formate adducts ([M+HCOO]- ions). MS/MS spectra of [M+HCOO]- ions of these three PL classes should display the typical fragment ion at m/z 168 (phosphocholine polar head group minus a methyl moiety). Carboxylate anions of fatty acyl chains can also be seen for PC and LPC. PE and LPE classes were analysed in negative ion mode ([M−H]- ions). The fragment ion at m/z 140 (phosphoethanolamine polar head group) and the carboxylate anions of fatty acyl chains can be found in the MS/MS data from negative ion mode. PI and PS species were analysed in negative ion mode, as [M−H]- ions. The presence of the fragment ion at m/z 241, corresponding to the phosphoinositol polar head group, in the MS/MS of [M-H]- ions, allows the identification of PI molecular species. PS species were identified in the MS/MS of [M-H]- ions by the neutral loss of -87Da from the molecular ion. The identification of the remaining lipid species belonging to the classes of carnitines (CAR), ceramides (Cer), cholesteryl esters (CE, fragment ion at m/z 369), diacylglycerols (DG) and triacylglycerols (TG), was made in LC-MS spectra in the positive ion mode, as [M+H]+ ions (CAR and Cer), [M+NH4]+ ions (CE) and [M+NH4]+ ions (DG and TG) respectively. LC-MS data were processed using the Lipostar software (Molecular Discovery Ltd., version 2.1.1 x64). This software was used for raw data import, peak detection, and identification. Lipid assignment and identification was made against a database created from LIPID MAPS structure database (version December 2022), that was then fragmented using the DM Manager Module in Lipostar, according to Lipostar fragmentation rules. The raw files were imported directly and aligned using the settings according to Lange et al. Briefly, automatic peak picking was performed with SDA smoothing level set to high and minimum S/N ratio 3. Automatic isotope clustering settings were set to 7 ppm with an RT tolerance of 0.2 min. The MS/MS filter was applied to keep only features with MS/MS spectra for identification. Lipid identification was made according to the following parameters: 5 ppm precursor ion mass tolerance and 10 ppm product ion mass tolerance. The automatic approval was performed to keep structures with a quality of 2-4 stars. The lists with the identified and approved species results were exported and we used MZmine software (v2.42) to perform relative quantification.Relative quantification was performed by exporting the peak area values to a computer spreadsheet. For data normalization, the peak areas of the extracted ion chromatograms (XIC) of the lipid precursor ions of each class were divided by the peak areas of the internal standards selected for the class. Missing values were replaced by 1/5 of the minimum positive values detected in the data set. Univariate and multivariate statistical analyses were performed using R version 3.5.1 in Rstudio version 1.1.4. The data sets were then normalized to the internal standard, generalized log2, and EigenMS. Principal component analysis (PCA) was performed using the R libraries FactoMineR and factoextra. Heatmaps were created using the R package heatmap using “Euclidean” as the clustering distance and “ward.D” as the clustering method. The normality of the data was tested with the Shapiro−Wilk test. To test the significance of the differences between conditions, we used either the ANOVA or Kruskal−Wallis test, followed by Tukey’s or Dunn’s test, respectively, using the R package Rstatix. A p-value < 0.05 was considered an indicator of statistical significance. All graphics and boxplots were created using the R package ggplot2.

Horses receiving antimicrobials may develop diarrhea due to changes in the gastrointestinal microbiome and metabolome. This matched, case-controlled study compared the fecal microbiome and metabolome in hospitalized horses on antibiotics that developed diarrhea (AAD), hospitalized horses on antibiotics that did not develop diarrhea (ABX) and a healthy, non-hospitalized control population (CON). Naturally-voided fecal samples were collected from AAD horses (n=17) the day that diarrhea developed and matched to ABX (n=15) and CON (n=31) horses for diet, antimicrobial agent and duration of antimicrobial therapy (< 5 days or > 5 days). Illumina sequencing of 16S rRNA genes on fecal DNA was performed. Alpha and beta diversity metrics were generated using QIIME 2.0. A Kruskal-Wallis with Dunn’s post-test and ANOSIM testing was used for statistical analysis. Microbiome composition in AAD was significantly different from CON (ANOSIM, R= 0.568, p=0.001) and ABX (ANOSIM, R=0.121, p=0.0012). Fecal samples were lyophilized and extracted using a solvent-based method. Untargeted metabolomics using gas chromatography-mass spectrometry platforms was performed. Metabolomic data was analyzed using Metaboanalyst 4.0 and Graphpad Prism v 7. Principal component analysis plots (PCA) were used to visualize the distribution of metabolites between groups. Heat maps were used to identify the relative concentrations amongst the most abundant 25 metabolites. A one-way ANOVA was used to compare differences in metabolites amongst the three groups of horses. Only named metabolites were included in the analysis. The microbiome of AAD and ABX horses had significantly decreased richness and evenness than CON horses (p<0.05). Actinobacteria (q=0.0192) and Bacteroidetes (q=0.0005) were different between AAD and CON. Verrucomicrobia was markedly decreased in AAD compared to ABX and CON horses (q=0.0005). Horses with AAD have a dysbiosis compared to CON horses, and show minor differences in bacterial community composition to ABX horses. Metabolite profiles of horses with AAD clustered separately from those with AAD or CON. Ten metabolites were found to be significantly different between groups (P<0.05) and are listed according to their metabolic pathway: amino acid metabolism (R-equol, L-tyrosine, kynurenic acid, xanthurenic acid, 5-hydroxyindole-3-acetic acid ) lipid metabolism (docosahexaenoic acid ethyl ester), biosynthesis of secondary metabolites (daidzein, isoquinoline) and two metabolites with unidentified pathways (1,3-divinyl-2-imidazolidinone, N-acetyltyramine).