Untargeted Metabolomics Datasets Research Articles

Enhancing supervised analysis of imbalanced untargeted metabolomics datasets using a CWGAN-GP framework for data augmentation

Untargeted metabolomics is an extremely useful approach for the discrimination of biological systems and biomarker identification. However, data analysis workflows are complex and face many challenges. Two of these challenges are the demand of high sample size and the possibility of severe class imbalance, which is particularly common in clinical studies. The latter can make statistical models less generalizable, increase the risk of overfitting and skew the analysis in favour of the majority class. One possible approach to mitigate this problem is data augmentation. However, the use of artificial data requires adequate data augmentation methods and criteria for assessing the quality of the generated data.In this work, we used Conditional Wasserstein Generative Adversarial Networks with Gradient Penalty (CWGAN-GPs) for data augmentation of metabolomics data. Using a set of benchmark datasets, we applied several criteria for the evaluation of the quality of generated data and assessed the performance of supervised predictive models trained with datasets that included such data. CWGAN-GP models generated realistic data with identical characteristics to real samples, mostly avoiding mode collapse. Furthermore, in cases of class imbalance, the performance of predictive models improved by supplementing the minority class with generated samples. This is evident for high quality datasets with well separated classes. Conversely, model improvements were quite modest for high class overlap datasets. This trend was confirmed by using synthetic datasets with different class separation levels. Data augmentation is a viable procedure to alleviate class imbalance problems but is not universally beneficial in metabolomics.

Soil metabolomics - current challenges and future perspectives

Soil is an extremely complex and dynamic matrix, in part, due to the wide diversity of organisms living within it. Soil organic matter (SOM) is the fundamental substrate on which the delivery of ecosystem services depends, providing the metabolic fuel to drive soil function. As such, studying the soil metabolome (the diversity and concentration of low molecular weight metabolites), as a subset of SOM, holds the potential to greatly expand our understanding of the behaviour, fate, interaction and functional significance of small organic molecules in soil. Encompassing a wide range of chemical classes (including amino acids, peptides, lipids and carbohydrates) and a large number of individual molecules (ca. n = 105 to 106), the metabolome is a resultant (indirect) output of several layers of a biological hierarchy, namely the metagenome, metatranscriptome and metaproteome. As such, it may also provide support and validation for these “multi-omics” datasets. We present a case for the increased use of untargeted metabolomics in soil biochemistry, particularly for furthering our fundamental understanding of the functions driving SOM composition and biogeochemical cycling. Further, we discuss the scale of the challenge in terms of metabolite extraction, analysis and interpretation in complex plant-soil-microbial systems. Lastly, we highlight key knowledge gaps which currently limit our use of metabolomic approaches to better understand soil processes, including: (i) interpretation of large untargeted metabolomic datasets; (ii) the source, emission and fate of soil-derived volatile organic compounds (VOCs), (iii) assessing temporal fluxes of metabolites, and (iv) monitoring ecological interactions in the rhizosphere. While the application of metabolomics in ecosystem science is still in its relative infancy, its importance in understanding the biochemical system in relation to regulation, management and underpinning the delivery of ecosystem services is key to further elucidating the complex links between organisms, as well as the fundamental ability of the biological community to process and cycle key nutrients.

IDSL.CSA: Composite Spectra Analysis for Chemical Annotation of Untargeted Metabolomics Datasets.

Poor chemical annotation of high-resolution mass spectrometry data limits applications of untargeted metabolomics datasets. Our new software, the Integrated Data Science Laboratory for Metabolomics and Exposomics─Composite Spectra Analysis (IDSL.CSA) R package, generates composite mass spectra libraries from MS1-only data, enabling the chemical annotation of high-resolution mass spectrometry coupled with liquid chromatography peaks regardless of the availability of MS2 fragmentation spectra. We demonstrate comparable annotation rates for commonly detected endogenous metabolites in human blood samples using IDSL.CSA libraries versus MS/MS libraries in validation tests. IDSL.CSA can create and search composite spectra libraries from any untargeted metabolomics dataset generated using high-resolution mass spectrometry coupled to liquid or gas chromatography instruments. The cross-applicability of these libraries across independent studies may provide access to new biological insights that may be missed due to the lack of MS2 fragmentation data. The IDSL.CSA package is available in the R-CRAN repository at https://cran.r-project.org/package=IDSL.CSA. Detailed documentation and tutorials are provided at https://github.com/idslme/IDSL.CSA.

Metabolomics Peak Analysis Computational Tool (MPACT): An Advanced Informatics Tool for Metabolomics and Data Visualization of Molecules from Complex Biological Samples.

Untargeted metabolomics is a powerful tool for investigating chemistry of complex biological systems, but its utility is compromised by the presence of uninformative features and the limited efficiency of currently available prioritization tools. More effective filtering and prioritization tools are required to address the challenges of large untargeted metabolomics datasets. Here, we introduce Metabolomics Peak Analysis Computational Tool (MPACT), a new mass spectrometry data analysis platform employing filtering based on multiple modalities, statistical techniques incorporating multilevel replication, and interactive data visualization. We demonstrate application of MPACT to uncover hidden effects of the rare earth element cerium on tunicate-associated bacterium Streptomyces sp. PTY087I2, culminating in characterization of two thiolated compounds including a new cysteine derivative, granaticin C, and granaticin D, recently described as mycothiogranaticin A. While we demonstrate application of MPACT to microbial natural products discovery using an elicitation approach, the platform should be readily adaptable to investigation of multipartite interactions, biomarker detection, small molecules in the environment, and a wide range of other complex sample types.

Benchmarking Outlier Detection Methods for Detecting IEM Patients in Untargeted Metabolomics Data.

Untargeted metabolomics (UM) is increasingly being deployed as a strategy for screening patients that are suspected of having an inborn error of metabolism (IEM). In this study, we examined the potential of existing outlier detection methods to detect IEM patient profiles. We benchmarked 30 different outlier detection methods when applied to three untargeted metabolomics datasets. Our results show great differences in IEM detection performances across the various methods. The methods DeepSVDD and R-graph performed most consistently across the three metabolomics datasets. For datasets with a more balanced number of samples-to-features ratio, we found that AE reconstruction error, Mahalanobis and PCA reconstruction error also performed well. Furthermore, we demonstrated the importance of a PCA transform prior to applying an outlier detection method since we observed that this increases the performance of several outlier detection methods. For only one of the three metabolomics datasets, we observed clinically satisfying performances for some outlier detection methods, where we were able to detect 90% of the IEM patient samples while detecting no false positives. These results suggest that outlier detection methods have the potential to aid the clinical investigator in routine screening for IEM using untargeted metabolomics data, but also show that further improvements are needed to ensure clinically satisfying performances.

Optimization of Imputation Strategies for High-Resolution Gas Chromatography-Mass Spectrometry (HR GC-MS) Metabolomics Data.

Gas chromatography–coupled mass spectrometry (GC–MS) has been used in biomedical research to analyze volatile, non-polar, and polar metabolites in a wide array of sample types. Despite advances in technology, missing values are still common in metabolomics datasets and must be properly handled. We evaluated the performance of ten commonly used missing value imputation methods with metabolites analyzed on an HR GC–MS instrument. By introducing missing values into the complete (i.e., data without any missing values) National Institute of Standards and Technology (NIST) plasma dataset, we demonstrate that random forest (RF), glmnet ridge regression (GRR), and Bayesian principal component analysis (BPCA) shared the lowest root mean squared error (RMSE) in technical replicate data. Further examination of these three methods in data from baboon plasma and liver samples demonstrated they all maintained high accuracy. Overall, our analysis suggests that any of the three imputation methods can be applied effectively to untargeted metabolomics datasets with high accuracy. However, it is important to note that imputation will alter the correlation structure of the dataset and bias downstream regression coefficients and p-values.

CCDB: A database for exploring inter-chemical correlations in metabolomics and exposomics datasets.

Inter-chemical correlations in metabolomics and exposomics datasets provide valuable information for studying relationships among chemicals reported for human specimens. With an increase in the number of compounds for these datasets, a network graph analysis and visualization of the correlation structure is difficult to interpret. We have developed the Chemical Correlation Database (CCDB), as a systematic catalogue of inter-chemical correlation in publicly available metabolomics and exposomics studies. The database has been provided via an online interface to create single compound-centric views. We have demonstrated various applications of the database to explore: 1) the chemicals from a chemical class such as Per- and Polyfluoroalkyl Substances (PFAS), polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), phthalates and tobacco smoke related metabolites; 2) xenobiotic metabolites such as caffeine and acetaminophen; 3) endogenous metabolites (acyl-carnitines); and 4) unannotated peaks for PFAS. The database has a rich collection of 35 human studies, including the National Health and Nutrition Examination Survey (NHANES) and high-quality untargeted metabolomics datasets. CCDB is supported by a simple, interactive and user-friendly web-interface to retrieve and visualize the inter-chemical correlation data. The CCDB has the potential to be a key computational resource in metabolomics and exposomics facilitating the expansion of our understanding about biological and chemical relationships among metabolites and chemical exposures in the human body. The database is available at www.ccdb.idsl.me site.

Artifactual FA dimers mimic FAHFA signals in untargeted metabolomics pipelines

FA esters of hydroxy FAs (FAHFAs) are lipokines with extensive structural and regional isomeric diversity that impact multiple physiological functions, including insulin sensitivity and glucose homeostasis. Because of their low molar abundance, FAHFAs are typically quantified using highly sensitive LC-MS/MS methods. Numerous relevant MS databases house in silico-spectra that allow identification and speciation of FAHFAs. These provisional chemical feature assignments provide a useful starting point but could lead to misidentification. To address this possibility, we analyzed human serum with a commonly applied high-resolution LC-MS untargeted metabolomics platform. We found that many chemical features are putatively assigned to the FAHFA lipid class based on exact mass and fragmentation patterns matching spectral databases. Careful validation using authentic standards revealed that many investigated signals provisionally assigned as FAHFAs are in fact FA dimers formed in the LC-MS pipeline. These isobaric FA dimers differ structurally only by the presence of an olefinic bond. Furthermore, stable isotope-labeled oleic acid spiked into human serum at subphysiological concentrations showed concentration-dependent formation of a diverse repertoire of FA dimers that analytically mimicked FAHFAs. Conversely, validated FAHFA species did not form spontaneously in the LC-MS pipeline. Together, these findings underscore that FAHFAs are endogenous lipid species. However, nonbiological FA dimers forming in the setting of high concentrations of FFAs can be misidentified as FAHFAs. Based on these results, we assembled a FA dimer database to identify nonbiological FA dimers in untargeted metabolomics datasets.

Nitrogen use efficiency, rhizosphere bacterial community, and root metabolome reprogramming due to maize seed treatment with microbial biostimulants

Seed inoculation with beneficial microorganisms has gained importance as it has been proven to show biostimulant activity in plants, especially in terms of abiotic/biotic stress tolerance and plant growth promotion, representing a sustainable way to ensure yield stability under low input sustainable agriculture. Nevertheless, limited knowledge is available concerning the molecular and physiological processes underlying the root‐inoculant symbiosis or plant response at the root system level. Our work aimed to integrate the interrelationship between agronomic traits, rhizosphere microbial population and metabolic processes in roots, following seed treatment with either arbuscular mycorrhizal fungi (AMF) or Plant Growth‐Promoting Rhizobacteria (PGPR). To this aim, maize was grown under open field conditions with either optimal or reduced nitrogen availability. Both seed treatments increased nitrogen uptake efficiency under reduced nitrogen supply revealed some microbial community changes among treatments at root microbiome level and limited yield increases, while significant changes could be observed at metabolome level. Amino acid, lipid, flavone, lignan, and phenylpropanoid concentrations were mostly modulated. Integrative analysis of multi‐omics datasets (Multiple Co‐Inertia Analysis) highlighted a strong correlation between the metagenomics and the untargeted metabolomics datasets, suggesting a coordinate modulation of root physiological traits.

Exploring the Contribution of (Poly)phenols to the Dietary Exposome Using High Resolution Mass Spectrometry Untargeted Metabolomics.

This study presents a workflow to construct a Dietary Exposome Library (DEL) comprised of phytochemicals and their metabolites derived from host and gut microbiome metabolism for use in peak identification/annotation of untargeted metabolomics datasets. An evidence mapping initiative established target analytes related to the consumption of phytochemical-rich foods. Analytes were confirmed by ultra-performance liquid chromatography-mass spectrometry (UPLC-MS(n)) analysis of human biospecimens from dietary intervention studies of (poly)phenol-rich diets. One hundred and sixty six verified compounds were subsequently analyzed on an untargeted metabolomics platform to acquire chromatographic and high-resolution mass spectral data for construction of a DEL. The DEL facilitated identification/annotation of 123 metabolites associate with exposure to (poly)phenol enriched diets, which included aromatic ketones, benzoic acids, ellagic acids, caffeoylquinic acids, catecholamines, coumarins, hippuric acid, hydroxytoluenes, phenylamines, stilbenes, urolithins, valerolactones, and xanthonoids, in untargeted metabolomics datasets acquire from human plasma and urine reference materials. The DEL focusing on (poly)phenols and their metabolites of dietary exposure facilitated identification/annotation of ingested food components and their associated pathways in untargeted metabolomics datasets acquired from human biospecimens. The DEL continues to expand with the aim to provide evidence-based data for dietary metabolites in exposome research and inform the development of dietary intervention strategies.

Metabolite discovery through global annotation of untargeted metabolomics data.

Liquid chromatography-high resolution mass spectrometry (LC-MS)-based metabolomics aims to identify and quantitate all metabolites, but most LC-MS peaks remain unidentified. Here, we present a global network optimization approach, NetID, to annotate untargeted LC-MS metabolomics data. The approach aims to generate, for all experimentally observed ion peaks, annotations that match the measured masses, retention times, and (when available) MS/MS fragmentation patterns. Peaks are connected based on mass differences reflecting adducting, fragmentation, isotopes, or feasible biochemical transformations. Global optimization generates a single network linking most observed ion peaks, enhances peak assignment accuracy, and produces chemically-informative peak-peak relationships, including for peaks lacking MS/MS spectra. Applying this approach to yeast and mouse data, we identified five previously unrecognized metabolites (thiamine derivatives and N-glucosyl-taurine). Isotope tracer studies indicate active flux through these metabolites. Thus, NetID applies existing metabolomic knowledge and global optimization to substantially improve annotation coverage and accuracy in untargeted metabolomics datasets, facilitating metabolite discovery.

An application of compositional data analysis to multiomic time-series data.

Compositional data analysis (CoDA) methods have increased in popularity as a new framework for analyzing next-generation sequencing (NGS) data. CoDA methods, such as the centered log-ratio (clr) transformation, adjust for the compositional nature of NGS counts, which is not addressed by traditional normalization methods. CoDA has only been sparsely applied to NGS data generated from microbial communities or to multiple ‘omics’ datasets. In this study, we applied CoDA methods to analyze NGS and untargeted metabolomic datasets obtained from bacterial and fungal communities. Specifically, we used clr transformation to reanalyze NGS amplicon and metabolomics data from a study investigating the effects of building material type, moisture and time on microbial and metabolomic diversity. Compared to analysis of untransformed data, analysis of clr-transformed data revealed novel relationships and stronger associations between sample conditions and microbial and metabolic community profiles.

Integrating untargeted metabolomics, genetically informed causal inference, and pathway enrichment to define the obesity metabolome.

BackgroundObesity and its associated diseases are major health problems characterized by extensive metabolic disturbances. Understanding the causal connections between these phenotypes and variation in metabolite levels can uncover relevant biology and inform novel intervention strategies. Recent studies have combined metabolite profiling with genetic instrumental variable (IV) analysis (Mendelian randomization) to infer the direction of causality between metabolites and obesity, but often omitted a large portion of untargeted profiling data consisting of unknown, unidentified metabolite signals.MethodsWe expanded upon previous research by identifying body mass index (BMI)-associated metabolites in multiple untargeted metabolomics datasets, and then performing bidirectional IV analysis to classify metabolites based on their inferred causal relationships with BMI. Meta-analysis and pathway analysis of both known and unknown metabolites across datasets were enabled by our recently developed bioinformatics suite, PAIRUP-MS.ResultsWe identified 10 known metabolites that are more likely to be causes (e.g. alpha-hydroxybutyrate) or effects (e.g. valine) of BMI, or may have more complex bidirectional cause-effect relationships with BMI (e.g. glycine). Importantly, we also identified about 5 times more unknown than known metabolites in each of these three categories. Pathway analysis incorporating both known and unknown metabolites prioritized 40 enriched (p < 0.05) metabolite sets for the cause versus effect groups, providing further support that these two metabolite groups are linked to obesity via distinct biological mechanisms.ConclusionsThese findings demonstrate the potential utility of our approach to uncover causal connections with obesity from untargeted metabolomics datasets. Combining genetically informed causal inference with the ability to map unknown metabolites across datasets provides a path to jointly analyze many untargeted datasets with obesity or other phenotypes. This approach, applied to larger datasets with genotype and untargeted metabolite data, should generate sufficient power for robust discovery and replication of causal biological connections between metabolites and various human diseases.

Identification of adulteration in botanical samples with untargeted metabolomics.

Adulteration remains an issue in the dietary supplement industry, including botanical supplements. While it is common to employ a targeted analysis to detect known adulterants, this is difficult when little is known about the sample set. With this study, untargeted metabolomics using liquid chromatography coupled to ultraviolet-visible spectroscopy (LC-UV) or high-resolution mass spectrometry (LC-MS) was employed to detect adulteration in botanical dietary supplements. A training set was prepared by combining Hydrastis canadensis L. with a known adulterant, Coptis chinensis Franch., in ratios ranging from 5 to 95% adulteration. The metabolomics datasets were analyzed using both unsupervised (principal component analysis and composite score) and supervised (SIMCA) techniques. Palmatine, a known H. canadensis metabolite, was quantified as a targeted analysis comparison. While the targeted analysis was the most sensitive method tested in detecting adulteration, statistical analyses of the untargeted metabolomics datasets detected adulteration of the goldenseal samples, with SIMCA providing the greatest discriminating potential. Graphical abstract.

Filtering procedures for untargeted LC-MS metabolomics data

BackgroundUntargeted metabolomics datasets contain large proportions of uninformative features that can impede subsequent statistical analysis such as biomarker discovery and metabolic pathway analysis. Thus, there is a need for versatile and data-adaptive methods for filtering data prior to investigating the underlying biological phenomena. Here, we propose a data-adaptive pipeline for filtering metabolomics data that are generated by liquid chromatography-mass spectrometry (LC-MS) platforms. Our data-adaptive pipeline includes novel methods for filtering features based on blank samples, proportions of missing values, and estimated intra-class correlation coefficients.ResultsUsing metabolomics datasets that were generated in our laboratory from samples of human blood, as well as two public LC-MS datasets, we compared our data-adaptive filtering method with traditional methods that rely on non-method specific thresholds. The data-adaptive approach outperformed traditional approaches in terms of removing noisy features and retaining high quality, biologically informative ones. The R code for running the data-adaptive filtering method is provided at https://github.com/courtneyschiffman/Metabolomics-Filtering.ConclusionsOur proposed data-adaptive filtering pipeline is intuitive and effectively removes uninformative features from untargeted metabolomics datasets. It is particularly relevant for interrogation of biological phenomena in data derived from complex matrices associated with biospecimens.

Using Expert Driven Machine Learning to Enhance Dynamic Metabolomics Data Analysis.

Data analysis for metabolomics is undergoing rapid progress thanks to the proliferation of novel tools and the standardization of existing workflows. As untargeted metabolomics datasets and experiments continue to increase in size and complexity, standardized workflows are often not sufficiently sophisticated. In addition, the ground truth for untargeted metabolomics experiments is intrinsically unknown and the performance of tools is difficult to evaluate. Here, the problem of dynamic multi-class metabolomics experiments was investigated using a simulated dataset with a known ground truth. This simulated dataset was used to evaluate the performance of tinderesting, a new and intuitive tool based on gathering expert knowledge to be used in machine learning. The results were compared to EDGE, a statistical method for time series data. This paper presents three novel outcomes. The first is a way to simulate dynamic metabolomics data with a known ground truth based on ordinary differential equations. This method is made available through the MetaboLouise R package. Second, the EDGE tool, originally developed for genomics data analysis, is highly performant in analyzing dynamic case vs. control metabolomics data. Third, the tinderesting method is introduced to analyse more complex dynamic metabolomics experiments. This tool consists of a Shiny app for collecting expert knowledge, which in turn is used to train a machine learning model to emulate the decision process of the expert. This approach does not replace traditional data analysis workflows for metabolomics, but can provide additional information, improved performance or easier interpretation of results. The advantage is that the tool is agnostic to the complexity of the experiment, and thus is easier to use in advanced setups. All code for the presented analysis, MetaboLouise and tinderesting are freely available.

Application of Stable Isotope Labels for Metabolomics in Studies in Fatty Liver Disease.

The progression of nonalcoholic fatty liver disease (NAFLD) increases the risks of cirrhosis and cardiovascular disease. Marked alteration of both cytosolic and mitochondrial metabolism, and in combination with insulin resistance, increases hepatic glucose production. Utilization of stable isotope tracers to study liver metabolism offers deep insight into rearrangements of metabolic pathways and substrate-product relationships under the conditions leading to fatty liver and induced by diseases, drugs, toxins, or genetic manipulations. Isotope tracing untargeted metabolomics (ITUM) recently emerged as a powerful platform in which the label can be tracked in an untargeted fashion, revealing the penetration of substrates into metabolic pathways, even at low abundance. Here, we describe a protocol that can be utilized to study the changes in utilization of any labeled substrate toward a wide range of metabolites either in isolated liver cells or whole liver tissue under conditions mimicking various stages of fatty liver disease. Furthermore, a routine protocol for extraction, separation, and mass spectrometric detection of isotopically labeled metabolites in an untargeted or targeted fashion. An informatic approach to analyze stable isotope untargeted metabolomic datasets is also described.

A Protocol to Compare Methods for Untargeted Metabolomics.

There are thousands of published methods for profiling metabolites with liquid chromatography/mass spectrometry (LC/MS). While many have been evaluated and optimized for a small number of select metabolites, very few have been assessed on the basis of global metabolite coverage. Thus, when performing untargeted metabolomics, researchers often question which combination of extraction techniques, chromatographic separations, and mass spectrometers is best for global profiling. Method comparisons are complicated because thousands of LC/MS signals (so-called features) in a typical untargeted metabolomic experiment cannot be readily identified with current resources. It is therefore challenging to distinguish methods that increase signal number due to improved metabolite coverage from methods that increase signal number due to contamination and artifacts. Here, we present the credentialing protocol to remove the latter from untargeted metabolomic datasets without having to identify metabolite structures. This protocol can be used to compare or optimize methods pertaining to any step of the untargeted metabolomic workflow (e.g., extraction, chromatography, mass spectrometer, informatic software, etc.).

Best-Matched Internal Standard Normalization in Liquid Chromatography-Mass Spectrometry Metabolomics Applied to Environmental Samples.

The goal of metabolomics is to measure the entire range of small organic molecules in biological samples. In liquid chromatography-mass spectrometry-based metabolomics, formidable analytical challenges remain in removing the nonbiological factors that affect chromatographic peak areas. These factors include sample matrix-induced ion suppression, chromatographic quality, and analytical drift. The combination of these factors is referred to as obscuring variation. Some metabolomics samples can exhibit intense obscuring variation due to matrix-induced ion suppression, rendering large amounts of data unreliable and difficult to interpret. Existing normalization techniques have limited applicability to these sample types. Here we present a data normalization method to minimize the effects of obscuring variation. We normalize peak areas using a batch-specific normalization process, which matches measured metabolites with isotope-labeled internal standards that behave similarly during the analysis. This method, called best-matched internal standard (B-MIS) normalization, can be applied to targeted or untargeted metabolomics data sets and yields relative concentrations. We evaluate and demonstrate the utility of B-MIS normalization using marine environmental samples and laboratory grown cultures of phytoplankton. In untargeted analyses, B-MIS normalization allowed for inclusion of mass features in downstream analyses that would have been considered unreliable without normalization due to obscuring variation. B-MIS normalization for targeted or untargeted metabolomics is freely available at https://github.com/IngallsLabUW/B-MIS-normalization .

Mining mass spectrometry data: Using new computational tools to find novel organic compounds in complex environmental mixtures

Untargeted metabolomics datasets provide ample opportunity for discovery of novel metabolites. The major challenge is focusing data analysis on a short list of metabolites. Here, we apply a combination of computational tools that serve to reduce complex mass spectrometry data in order allow us to focus on new environmentally-relevant metabolites. In the first portion of the project, we explored mass spectrometry data from intracellular metabolites extracted from a model marine diatom, Thalassiosira pseudonana. The fragmentation data from these samples were analyzed using molecular networking, an on-line tool that clusters metabolites based on shared structural similarities. The features within each metabolite cluster were then putatively annotated using MetFrag, an in silico fragmentation tool. Using this combination of computational tools, we observed multiple lyso-sulfolipids, organic compounds not previously known to exist within cultured marine diatoms. In the second stage of the project, we searched our environmental data for these lyso-sulfolipids. The lyso-sulfolipid with a C14:0 fatty acid was found in dissolved and particulate samples from the western Atlantic Ocean, and a culture of cyanobacteria grown in our laboratory. Thus, the putative lyso-sulfolipids are present in both laboratory experiments and environmental samples. This project highlights the value of combining computational tools to detect and putatively identify organic compounds not previously recognized as important within T. pseudonana or the marine environment. Future applications of these tools to emerging metabolomics data will further open the black box of natural organic matter, identifying molecules that can be used to understand and monitor the global carbon cycle.