Uncovering PFAS and Other Xenobiotics in the Dark Metabolome Using Ion Mobility Spectrometry, Mass Defect Analysis, and Machine Learning

Foster, MaKayla; Rainey, Markace; Watson, Chandler; Dodds, James N.; Kirkwood, Kaylie I; Fernández, Facundo M.; Baker, Erin

doi:10.1021/acs.est.2c00201

Cited by 52 publications

(51 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These chemically diverse training sets produce CCS predictions that may not be accurate enough to filter sufficient candidates. However, if the CCS prediction accuracy is improved using a training set that is structurally closer to the analyte, then CCS filtering produces more satisfactory results, as seen in prior predictions for structurally similar xenobiotics . The prediction of the unknown analyte superclass is indeed possible using IM-MS coupled to ML .…”

Section: Resultsmentioning

confidence: 99%

“…However, if the CCS prediction accuracy is improved using a training set that is structurally closer to the analyte, then CCS filtering produces more satisfactory results, as seen in prior predictions for structurally similar xenobiotics. 33 The prediction of the unknown analyte superclass is indeed possible using IM-MS coupled to ML. 34 Notably, the effectiveness of CCS filtering also depends on the mass accuracy afforded by the back-end mass spectrometer.…”

Section: ■ Introductionmentioning

confidence: 99%

See 1 more Smart Citation

CCS Predictor 2.0: An Open-Source Jupyter Notebook Tool for Filtering Out False Positives in Metabolomics

et al. 2022

Self Cite

View full text Add to dashboard Cite

Metabolite annotation continues to be the widely accepted bottleneck in nontargeted metabolomics workflows. Annotation of metabolites typically relies on a combination of high-resolution mass spectrometry (MS) with parent and tandem measurements, isotope cluster evaluations, and Kendrick mass defect (KMD) analysis. Chromatographic retention time matching with standards is often used at the later stages of the process, which can also be followed by metabolite isolation and structure confirmation utilizing nuclear magnetic resonance (NMR) spectroscopy. The measurement of gas-phase collision cross-section (CCS) values by ion mobility (IM) spectrometry also adds an important dimension to this workflow by generating an additional molecular parameter that can be used for filtering unlikely structures. The millisecond timescale of IM spectrometry allows the rapid measurement of CCS values and allows easy pairing with existing MS workflows. Here, we report on a highly accurate machine learning algorithm (CCSP 2.0) in an open-source Jupyter Notebook format to predict CCS values based on linear support vector regression models. This tool allows customization of the training set to the needs of the user, enabling the production of models for new adducts or previously unexplored molecular classes. CCSP produces predictions with accuracy equal to or greater than existing machine learning approaches such as CCSbase, DeepCCS, and AllCCS, while being better aligned with FAIR (Findable, Accessible, Interoperable, and Reusable) data principles. Another unique aspect of CCSP 2.0 is its inclusion of a large library of 1613 molecular descriptors via the Mordred Python package, further encoding the fine aspects of isomeric molecular structures. CCS prediction accuracy was tested using CCS values in the McLean CCS Compendium with median relative errors of 1.25, 1.73, and 1.87% for the 170 [M − H] − , 155 [M + H] + , and 138 [M + Na] + adducts tested. For superclass-matched data sets, CCS predictions via CCSP allowed filtering of 36.1% of incorrect structures while retaining a total of 100% of the correct annotations using a Δ CCS threshold of 2.8% and a mass error of 10 ppm.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: ■ Introductionmentioning

confidence: 99%

CCS Predictor 2.0: An Open-Source Jupyter Notebook Tool for Filtering Out False Positives in Metabolomics

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…These advances in technology have driven a shift from targeted to nontargeted analyses that can overcome sensitivity issues associated with chemical detection [ 31 ]. Furthermore, nontargeted analyses performed with high-resolution MS and separations coupled to MS, including IMS-MS have improved the characterization of metabolites following exposure to environmental contaminants [ 69 , 70 ].…”

Section: Discussionmentioning

confidence: 99%

Evaluation of Metabolism of a Defined Pesticide Mixture through Multiple In Vitro Liver Models

Valdiviezo

Kato

Baker

et al. 2022

Toxics

View full text Add to dashboard Cite

The evaluation of exposure to multiple contaminants in a mixture presents a number of challenges. For example, the characterization of chemical metabolism in a mixture setting remains a research area with critical knowledge gaps. Studies of chemical metabolism typically utilize suspension cultures of primary human hepatocytes; however, this model is not suitable for studies of more extended exposures and donor-to-donor variability in a metabolic capacity is unavoidable. To address this issue, we utilized several in vitro models based on human-induced pluripotent stem cell (iPSC)-derived hepatocytes (iHep) to characterize the metabolism of an equimolar (1 or 5 µM) mixture of 20 pesticides. We used iHep suspensions and 2D sandwich cultures, and a microphysiological system OrganoPlate® 2-lane 96 (MimetasTM) that also included endothelial cells and THP-1 cell-derived macrophages. When cell culture media were evaluated using gas and liquid chromatography coupled to tandem mass spectrometry methods, we found that the parent molecule concentrations diminished, consistent with metabolic activity. This effect was most pronounced in iHep suspensions with a 1 µM mixture, and was lowest in OrganoPlate® 2-lane 96 for both mixtures. Additionally, we used ion mobility spectrometry–mass spectrometry (IMS-MS) to screen for metabolite formation in these cultures. These analyses revealed the presence of five primary metabolites that allowed for a more comprehensive evaluation of chemical metabolism in vitro. These findings suggest that iHep-based suspension assays maintain higher metabolic activity compared to 2D sandwich and OrganoPlate® 2-lane 96 model. Moreover, this study illustrates that IMS-MS can characterize in vitro metabolite formation following exposure to mixtures of environmental contaminants.

show abstract

“…These chemically diverse training sets produce CCS predictions that may not be accurate enough to filter sufficient candidates. However, if the CCS prediction accuracy is improved using a training set that is structurally closer to the analyte, then CCS filtering produces more satisfactory results, as seen in prior predictions for structurally-similar xenobiotics 34 . Notably, the effectiveness of CCS filtering also depends on the mass accuracy afforded by the back-end mass spectrometer.…”

Section: Ccsp Applications To Metabolite Annotationmentioning

confidence: 98%

CCS Predictor 2.0: An Open-Source Jupyter Notebook Tool for Filtering Out False Positives in Metabolomics

Rainey

Watson

Asef

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Metabolite annotation continues to be the widely accepted bottleneck in non-targeted metabolomics workflows. Annotation of metabolites typically relies on a combination of high resolution mass spectrometry (MS) with parent and tandem measurements, isotope cluster evaluations, and Kendrick mass defect (KMD) analysis. Chromatographic retention time matching with standards is often used at the later stages of the process, which can also be followed by metabolite isolation and structure confirmation utilizing nuclear magnetic resonance (NMR) spectroscopy. The measurement of gas phase collision cross section (CCS) values by ion mobility (IM) spectrometry also adds an important dimension to this workflow by generating an additional molecular parameter that can be used for filtering unlikely structures. The millisecond timescale of IM spectrometry allows the rapid measurement of CCS values and allows easy pairing with existing MS workflows. Here, we report on a highly accurate machine learning algorithm (CCSP 2.0) in an open-source Jupyter Notebook format to predict CCS values based on linear support vector regression models. This tool allows customization of the training set to the needs of the user, enabling the production of models for new adducts or previously unexplored molecular classes. CCSP produces predictions with accuracy equal to or greater than existing machine learning approaches such as CCSbase, DeepCCS and AllCCS, while being better aligned with FAIR (Findable, Accessible, Interoperable and Reusable) data principles. Another unique aspect of CCSP 2.0 its inclusion of a large library of 1613 molecular descriptors via the Mordred Python package, further encoding the fine aspects of isomeric molecular structures. CCS prediction accuracy was tested using CCS values in the McLean CCS Compendium with median relative errors of 1.25, 1.73 and 1.87% for the 170 [M-H]-, 155 [M+H]+ and 138 [M+Na]+ adducts tested. For class-matched data sets, CCS predictions via CCSP allowed filtering of 36.1% of incorrect structures while retaining a total of 100% of the correct annotations using a CCS threshold of 2.8% and a mass error of 10 ppm.

show abstract

Uncovering PFAS and Other Xenobiotics in the Dark Metabolome Using Ion Mobility Spectrometry, Mass Defect Analysis, and Machine Learning

Cited by 52 publications

References 61 publications

CCS Predictor 2.0: An Open-Source Jupyter Notebook Tool for Filtering Out False Positives in Metabolomics

CCS Predictor 2.0: An Open-Source Jupyter Notebook Tool for Filtering Out False Positives in Metabolomics

Evaluation of Metabolism of a Defined Pesticide Mixture through Multiple In Vitro Liver Models

CCS Predictor 2.0: An Open-Source Jupyter Notebook Tool for Filtering Out False Positives in Metabolomics

Contact Info

Product

Resources

About