Search

Untargeted metabolomics experiments provide a snapshot of cellular metabolism but remain challenging to interpret due to the computational complexity involved in data processing and analysis. Prior to any interpretation, raw data must be processed to remove noise and to align mass-spectral peaks across samples. This step requires selection of dataset-specific parameters, as erroneous parameters can result in noise inflation. While several algorithms exist to automate parameter selection, each depends on gradient descent optimization functions. In contrast, our new parameter optimization algorithm, AutoTuner, obtains parameter estimates from raw data in a single step as opposed to many iterations. Here, we tested the accuracy and the run-time of AutoTuner in comparison to isotopologue parameter optimization (IPO), the most commonly used parameter selection tool, and compared the resulting parameters' influence on the properties of feature tables after processing. We performed a Monte Carlo experiment to test the robustness of AutoTuner parameter selection and found that AutoTuner generated similar parameter estimates from random subsets of samples. We conclude that AutoTuner is a desirable alternative to existing tools, because it is scalable, highly robust, and very fast (∼100-1000× speed improvement from other algorithms going from days to minutes). AutoTuner is freely available as an R package through BioConductor.

Dissolved metabolites serve as nutrition, energy, and chemical signals for microbial systems. However, the full scope and magnitude of these processes in marine systems are unknown, largely due to insufficient methods, including poor extraction of small, polar compounds using common solid-phase extraction resins. Here, we utilized pre-extraction derivatization and ultrahigh performance liquid chromatography electrospray ionization tandem mass spectrometry (UHPLC-ESI-MS/MS) to detect and quantify targeted dissolved metabolites in seawater and saline culture media. Metabolites were derivatized with benzoyl chloride by their primary and secondary amine and alcohol functionalities and quantified using stable isotope-labeled internal standards (SIL-ISs) produced from 13C6-labeled benzoyl chloride. We optimized derivatization, extraction, and sample preparation for field and culture samples and evaluated matrix-derived biases. We have optimized this quantitative method for 73 common metabolites, of which 50 cannot be quantified without derivatization due to low extraction efficiencies. Of the 73 metabolites, 66 were identified in either culture media or seawater and 45 of those were quantified. This derivatization method is sensitive (detection limits = pM to nM), rapid (∼5 min per sample), and high throughput.

Ultrahigh resolution mass spectrometry, such as Fourier transform ion cyclotron resonance mass spectrometry (FT ICR MS), can resolve thousands of molecular ions in complex organic matrices. A Compound Identification Algorithm (CIA) was previously developed for automated elemental formula assignment for natural organic matter (NOM). In this work, we describe software Formularity with a user-friendly interface for CIA function and newly developed search function Isotopic Pattern Algorithm (IPA). While CIA assigns elemental formulas for compounds containing C, H, O, N, S, and P, IPA is capable of assigning formulas for compounds containing other elements. We used halogenated organic compounds (HOC), a chemical class that is ubiquitous in nature as well as anthropogenic systems, as an example to demonstrate the capability of Formularity with IPA. A HOC standard mix was used to evaluate the identification confidence of IPA. Tap water and HOC spike in Suwannee River NOM were used to assess HOC identification in complex environmental samples. Strategies for reconciliation of CIA and IPA assignments were discussed. Software and sample databases with documentation are freely available.

The advent of ultra-high-resolution mass spectrometry has revolutionized the ability of aquatic biogeochemists to examine molecular-level components of complex mixtures of organic matter. The ability to accurately assess the chemical composition, elemental formulas, or both of detected compounds is critical to these studies. Here we build on previous work that uses functional group relationships between compounds to extend elemental formulas of low molecular weight compounds to those of higher molecular weight. We propose an automated compound identification algorithm (CIA) for the analysis of ultra-high-resolution mass spectra of natural organic matter acquired by electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. This approach is benchmarked with synthetic data sets of compounds cited in the literature. The sensitivity of our results is examined for different sources of error, and CIA is applied to two previously published data sets. We find that CIA works well for data sets with high mass accuracy (1 ppm) and can accurately determine the elemental formulas for95% of all compounds composed of C, H, O, and N. Data with lower mass accuracy must be accompanied with additional knowledge of chemical structure, composition, or both in order to yield accurate elemental formulas.

Full structural characterization of complex mixtures such as humic acid extracts has been elusive because of insufficient compound resolution with conventional techniques. Using electrospray ionization coupled with Fourier transform ion cyclotron resonance mass spectrometry, we were able to resolve individual compounds within humic and fulvic acid mixtures (mass resolving power approximately 80000 at 300 m/z). We examined two samples in detail: (1) dissolved organic matter (primarily fulvic acids) from Suwannee River, GA, and (2) a humic acid extract from a degraded wood collected on Mt. Rainier, WA. Sample conditions (such as solvent, pH, and concentration) and instrument parameters (such as source voltages, trapping potentials, and excitation parameters) were optimized to yield the highest mass resolving power with the least mass discrimination in positive ion mode. High resolving power was achieved with low ion densities combined with coadding numerous scans. The increased resolution allowed molecular-level comparisons of the two samples which in turn could be used to estimate the relative similarity of individual compound distribution as well as an indication of the dominant diagenetic processes in the two source environments.