Characterization of Xenobiotic Metabolism Using Ion Mobility-Mass Spectrometry and Machine Learning
May 24, 2023
Dr. Libin Xu
Xenobiotics, including drugs and environmental chemicals, exert large structural diversity in a small mass range, mostly less than 1000 Da. Xenobiotic metabolism by Phase-I and Phase-II enzymes further increases such structural diversity. The large number of molecules in a small mass range presents a challenge for their resolution by mass spectrometry alone, but such a challenge can be addressed by ion mobility-mass spectrometry (IM-MS), which can increase the throughput and confidence in the identification of these molecules. IM rapidly separates ions based on their collision cross sections with a neutral buffer gas, a physical property that is characteristic of a particular ion. In this talk, I will discuss the application of IM-MS in the characterization of the impact of xenobiotic metabolism on the gas-phase structures of xenobiotics, the assembly of a large-scale database of xenobiotics and their metabolite CCS values using high-throughput in vitro metabolite generation and IM-MS analysis, and the training of a machine learning-based CCS prediction model using both 2D and 3D molecular descriptors, which enables the prediction of different CCS values for different protomers, conformers, and positional isomers.
Dr. Xu received his PhD in organic chemistry from the University of Illinois at Chicago. He then underwent postdoctoral training at Vanderbilt University, where his research expanded to chemistry and biology of lipid peroxidation underlying human diseases, as well as mass spectrometry-based lipidomics and imaging. He joined the Department of Medicinal Chemistry at the UW as an Assistant Professor in 2014 and was promoted to Associated Professor in 2020. Dr. Xu is a recipient of the NIH Pathway to Independence Award from NICHD in 2012 and the Young Investigator Award from the Society for Free Radical Biology and Medicine in 2011. He is also the inaugural recipient of the School of Pharmacy Faculty Innovation Fund in 2016. The Xu lab is interested in the chemistry and biology of lipid oxidation, drug metabolism, the impact of xenobiotics on the nervous system, kidney, liver, and microbiome, antibiotic resistance, as well as the development of mass spectrometry-based metabolomics, lipidomics, and imaging technologies.
Achieving Ultra-High Resolution IMS-MS/MS in a Single Experiment Using a New SLIM-Orbitrap Platform
April 26, 2023
Dr. Adam Hollerbach
Predicting the properties of gas-phase ions, such as molecular formula, typically requires accurate measurements of mass-to-charge ratios (m/z) using a high-resolution mass spectrometer (MS). However, many ions have the same molecular formula and m/z but different chemical structures. Therefore, highly precise collision cross section measurements using ion mobility spectrometry (IMS) are needed in conjunction with high-resolution m/z measurements to confidently identify an unknown molecule.
To address this need, this presentation will describe the development of a new ultra-high resolution IMS-MS platform that can perform high-resolution ion mobility separations and high-resolution mass analysis, including collision induced dissociation, in a single experiment.
Ion mobility separations are performed using a traveling wave-based structures for lossless ion manipulations ion mobility spectrometer (SLIM-IMS), and mass analysis and collision induced dissociation are performed using an Orbitrap-MS. The three simultaneous measurements provided by the SLIM-Orbitrap lay the foundation for providing unique glimpses into ion structures not presently feasible or demonstrated with other IMS-MS platforms. Furthermore, combining SLIM-Orbitrap data with advanced computing tools has the ultimate goal of being able to identify the structures of unknown ions without needing a chemical reference as the final confirmation step in an analytical workflow (i.e., reference-free chemical analysis).
Dr. Adam Hollerbach is a staff scientist in the biological sciences division at PNNL. He received his PhD from Purdue University in 2018 and joined PNNL as a postdoc the same year. His current work involves developing innovative high-resolution ion mobility spectrometers (IMS) for panomics applications. His recent work on developing a miniature multilevel SLIM was featured on the front cover of Analytical Chemistry, and this work has steered the development of some of the newest high-resolution SLIM-IMS systems available at PNNL.
Manipulating Molecular Ion Trains: New Strategies for Targeted Gas-Phase Ion Chemistry
March 13, 2023
Dr. Brian H. Clowers
Impacting quality control efforts, statistical confidence, and signal-to-noise ratio for many analytical techniques, separation science is foundational across disciplines. While the capacity to isolate target chemical species and their associated signals has direct bearing on decision-making, the efficient isolation of molecules also provides a setting to conduct detailed physical and chemical experiments free from interference. By developing a series of innovative tools for separating and isolating molecular ions, our research team is pursuing a range of techniques to probe selective gas-phase ion chemistry under conditions previously unattainable.
Specific topics for discussion include efforts by the Clowers research team at WSU to apply molecular ion storage techniques toward fundamental questions within the field of gas-phase ion chemistry. The independent implementation of the SLIM platform will feature prominently in this ion chemistry quest. Salient examples of gas-phase hydrogen-deuterium exchange and cryogenic ion spectroscopy on trapped ion populations will illustrate how these experiments offer new insights into the mechanisms of molecular solvation. The digital manipulation of ions offers an unprecedented opportunity to isolate, store, and transform molecular species with implications spanning across disciplines.
Brian H. Clowers, Ph.D., is an Associate Professor of Chemistry at Washington State University (WSU), focusing on the fundamentals, development, and application of rapid gas-phase separations techniques with an emphasis on ion mobility and mass spectrometry. Before joining WSU, he worked for Pacific Northwest National Laboratory (PNNL), as part of an interdisciplinary team focusing on forensic challenges across the threat spectrum. In 2020, Dr. Clowers was named the Boeing Distinguished Professor of Science Education at WSU. Dr. Clowers is the author of over 100 peer-reviewed articles, 7 patents, and 1 book chapter. He holds a BS in chemistry from the University of Nevada, Reno, and a Ph.D. in chemistry from WSU (2005). Brian also served as a post-doctoral researcher at UC-Davis and PNNL. He is also a member of the Editorial Advisory Committee for the Journal – American Society for Mass Spectrometry and is an instructor for the short courses on ion mobility spectrometry at conferences.
Beam Ionization: Who Gets the Charge and Why Do Some Get Broken Up About It?
January 30, 2023
Dr. Elizabeth H. Denis,PNNL
We are exploring ion manipulation at atmospheric pressure to increase signal for mass spectrometry and ion mobility spectrometry instruments. Increases in signal will help to develop smaller instruments that are more sensitive and portable for field applications. One way to increase sensitivity is by increasing ion-molecule reaction time such as in an atmospheric flow tube, where increased distance between the ionization source and detector allows more time for the target molecules to collide with reactant ions. One of the challenges with increased reaction time is that ion loss (e.g., by diffusion) limits instrument sensitivity to threats. Due to collisions with air molecules, ion manipulation at higher pressures (e.g., atmospheric pressure) is more challenging than under vacuum conditions.
In laboratory experiments, we are adjusting parameters such as air flow, electric fields, and polarity of the ions present to improve instrument signal. Molecular dynamics and ion trajectory simulations (SIMION and SimELIT) are being used to understand and predict ion behavior. By linking experimental observations with modeling results we can predict ion motion at ambient conditions and improve instrument design. The goal is to create portable handheld devices with similar sensitivity to that of larger benchtop instruments. Such devices would be useful for national security applications, such as detecting chemical threats at airports and ports of entry.
Dr. Elizabeth Denis is a Chemist at PNNL leading and supporting a variety of projects, including on developing and optimizing trace chemical detection of explosives and drugs using mass spectrometry. Many of her projects integrate laboratory experimental observations with computational modeling and statistics. Elizabeth earned her Ph.D. in Geosciences and Biogeochemistry from the Pennsylvania State University in 2016 and was an NSF Graduate Research Fellow. She earned a B.S. with honors in Geology-Chemistry from Brown University. Elizabeth’s passion for pushing the limits of our understanding of chemistry has driven her toward method development-focused research and has resulted in 2 patents and 20 publications.
Increasing the Versatility of MS/MS Reference Libraries with Machine Learning
October 26, 2022
Dr. Nellie Ciesielski, PNNL
Tandem mass spectrometry (MS/MS) is a primary tool for identifying small molecules and metabolite. Resultant spectra are identified by matching against spectra in reference libraries. The customizability of MS/MS acquisition techniques creates a significant challenge for building standardized libraries like those that made gas chromatography-MS a gold standard. We are developing machine learning tools to either predict differences between reference spectra and the spectra generated by local workflows or workflow specific spectra at various collision energies for a previously uncharacterized compound.
One tool interpolates spectra at intermediate collision energies based on three reference spectra. While exact collision energies applied by commercial instruments are indiscernible, this method generates spectra at a range of collision energies to find a close match despite expected shifts between reference and experimental spectra. Another provides a fragment database to a LSTM RNN that learns to reconstruct the expected spectrum.
Our final tool is a GNN that embeds quantum chemical information into the chemical bonds to inform the neural network about the bonds. GNNs work better than neural networks with molecular fingerprints as training data, but still average a performance just under 0.5 cosine similarity. While adding QC information generated mild improvements, the use of convolutional layers with attention provided an inside look at what the GNN is learning. This insight allows us to provide the GNN with further information to improve its predictions.
Nellie Ciesielski is a data scientist at Pacific Northwest National Laboratory. She earned her PhD in applied mathematics from Montana State University in 2019. Previously, she worked as an operations research analyst with the Army Futures Command. Since joining PNNL, her work has focused on projects in biosecurity, statistical analysis of chemical spaces, and developing neural networks for learning nonlinear mathematical operators.
An Imbalance in Force—Toward Phase-Transferable Simulation Models
September 28, 2022
David Van Der Spoel, Uppsala University
Molecular dynamics simulations give a powerful way to study processes such as evaporation of water from a protein in the gas phase. Since there's little experimental information on the structure of proteins under such conditions, theoretical methods can give new insights.
However, most protein force fields have been developed for use in the condensed phase only. The same holds for small molecules, be they co-factors, ligands, or drug candidates. While substantial effort has been put into tuning force fields to reproduce hydration-free energies of small molecules, limited studies have been conducted to explore gas phase properties, such as the vibrational spectra or thermochemistry. David's research group has spent considerable time evaluating different models and simultaneously exploring new generic force field models. In this presentation, he'll first give some background into applications of interest, then present some of the benchmarks achieved, and finally give an outlook of the results of new models they're developing.
David is a professor in the Department of Cell and Molecular Biology at Uppsala University, Sweden. David obtained his doctorate in biophysical chemistry from Groningen University, The Netherlands, in 1996. His research focuses on development of physical models and software for molecular simulation. He is one of the original developers of the GROMACS software and has studied protein folding, biomolecules in the gas phase, and virus capsids. Currently, his main research objective is development of the Alexandria force field.
Current Challenges in Mass Spectrometry Characterization of Non-Covalent Interactions
August 24, 2022
Dr. Valérie Gabelica, Institut Européen de Chimie et Biologie (IECB)
Non-covalent interactions orchestrate the molecules of Life, through reversible binding and self-assembly. When operated in gentle conditions, electrospray mass spectrometry can preserve non-covalent interactions from the solution to the gas phase. With advanced characterization methods such as ion mobility spectrometry and ion spectroscopy, one can hope to characterize the structure of each assembly. The wide adoption of native MS is hampered by fundamental questions that have not been fully resolved. This talk took participants through the journey of our investigations and pointed to some unresolved questions, such as: 1) What are the ionization mechanisms at stake for nucleic acids, depending on their structure and the solvent composition? 2) How to derive correct structural information from ion mobility spectrometry and gas-phase ion spectroscopy? 3) Are the solution structures preserved in the gas phase? 4) What molecular modeling approaches should be used to interpret collision cross sections? Finally, we pointed to current challenges in data interpretation and visualization, for a broader understanding by non-MS specialists.
Valérie obtained her PhD in 2002 at the University of Liège in Belgium with Prof. Edwin De Pauw. After a Humboldt post-doc in Frankfurt with Prof. Dr. Michael Karas, she rejoined the mass spectrometry laboratory in Liège where she secured a permanent research associate position from the national Funds for Scientific Research-FNRS in 2005. In 2013, she moved to IECB Bordeaux, France, to establish her independent group, obtained prestigious ATIP-Avenir and ERC funding and became a research director of the INSERM (National Institute for Health and Medical Research, France). She was awarded the French Academy of Sciences chemical biology prize in 2018, the Liliane Bettencourt prize for Life Sciences in 2021, and the Heinrich Emanuel Merck prize for Analytical Sciences in 2022. She currently serves as the director of IECB, and as associate editor for Analytical Chemistry.
Single Cell Analysis and Imaging for Spatial Metabolomics by New Ionization Modalities in Mass Spectrometry
May 25, 2022
Akos Vertes, George Washington University
To uncover cellular phenotypes resulting from variations in metabolism, it's crucial to determine the metabolite composition of single cells.
Akos Vertes' team has developed three sampling and ionization methods for mass spectrometry (MS) to cope with the challenges of limited analyte pool and high turnover rates: ultrasensitive nanophotonic laser desorption ionization MS, based on geometry-optimized silicon nanopost arrays; capillary microsampling followed by electrospray ionization MS; and optical fiber-based laser ablation electrospray ionization MS.
These techniques will be described and their applications in studies of a variety of systems, including single yeast cells exposed to oxidative stress, pavement/basal cells, and trichomes in A. thaliana leaf epidermis, and in combination with bimodal microscopy and image analysis that enabled the high-throughput ambient analysis of tissue-embedded single cells (n > 1000) and provided new insight into cellular heterogeneity. Advances in these technologies will be essential to developing new capabilities in spatial metabolomics.
Akos is a professor of chemistry at George Washington University in Washington, DC. His research interests encompass the development of new analytical techniques applicable in diverse fields of chemistry, biology, and medicine. His research has been presented in more than 180 peer-reviewed publications (h-index = 49) and two books. He's a coinventor on 19 patents and several pending patent applications. He was elected a fellow of the National Academy of Inventors and received the Distinguished Researcher Award at GWU, the 2012 Hillebrand Prize, and the Oscar and Shoshana Trachtenberg Prize for Scholarship. He served as visiting faculty at the Lawrence Berkeley National Laboratory, an MTA distinguished guest scientist at the Hungarian Academy of Sciences in Hungary, and a visiting professor at the Swiss Federal Institute of Technology Zurich in Switzerland.
Beam Ionization—Who Gets the Charge and Why Do Some Get Broken Up About It?
April 27, 2022
Chris Anderton, PNNL
Beam ionization by lasers and primary ions are key methods for directly probing samples for in situ chemical and spatial omics analyses.
Many factors play a role in the ability to measure any given molecule using beam-based ionization. A significant contributor to the mechanics of how molecules obtain a charge, and whether they experience in-source fragmentation (ISF)—where the molecule fragments during desorption and ionization—is the environment surrounding the molecules. This is often referred to as the matrix effect. Little is known about the role of the surrounding matrix on molecular ionization by beam-based methods.
Chris Anderton's team aims to understand the mechanisms of small-molecule desorption and ionization from different substrates and complex sample types (e.g., biological tissues) using laser and ion beams, and how these small molecules undergo ISF while in the gas phase. The team uses a combination of experimental data and molecular modeling approaches to elucidate these mechanisms. Ultimately, the team aims for this knowledge to be used to generate more confident interpretations of in situ data obtained from laser and primary ion beam approaches.
Chris leads a team of researchers in biogeochemical transformations at the Environmental Molecular Sciences Laboratory (EMSL). He has an extensive background in elucidating chemical interactions occurring across all kingdoms of life, including those within soils and the rhizosphere. While at EMSL, his focus has been on expanding the mass spectrometry imaging capability—making them valuable tools for analyzing bacteria communities, rhizosphere-related systems, and even human health-related processes.
Understanding and Controlling Atomically-Precise Materials Through Ion-Surface Interactions
March 30, 2022
Grant E. Johnson, PNNL
Challenges in efficient energy generation and storage, green manufacturing, and quantum computing may be addressed through a fundamental molecular-level understanding of material properties and processes enabled by unconventional mass spectrometry techniques. Novel materials not obtainable by conventional synthesis approaches may be prepared in the gas phase and delivered to surfaces using ion soft landing. A wide range of polyatomic ions with precisely defined composition and ionic charge may be immobilized on different supports with predetermined coverage and kinetic energy, thereby circumventing the sample heterogeneity, contamination, and aggregation that often confound experimental characterization and theoretical modeling of materials.
This presentation will illustrate how ion soft landing is being employed to understand the underlying phenomena that may be harnessed to improve the performance of advanced materials for fuel cell catalysts, supercapacitors, and molecular qubit arrays. Well-defined materials prepared by ion soft landing, combined with state-of-the-art characterization techniques and high-level theoretical modeling, are providing transformative insight into how materials may be designed and controlled at the atomic-level to address a range of energy-related challenges.
Dr. Grant E. Johnson is a Scientist and Group Lead in the Chemical Physics and Analysis group of the Physical Sciences Division at Pacific Northwest National Laboratory. He received his B.S. in Chemistry from the University of Delaware with Prof. Robert H. Wood. He earned his Ph.D. in Chemistry from the Pennsylvania State University with Prof. A.W. Castleman, Jr. and completed a Linus Pauling distinguished postdoctoral fellowship at PNNL with Laboratory Fellow Dr. Julia Laskin. Grant’s collaborative research spanning separation science, energy storage, and quantum information systems is funded by the U.S. Department of Energy (DOE). Grant received the Ronald L. Brodzinski Award for Early Career Exceptional Achievement in 2015.
Focusing Ions at Atmospheric Pressure Using Nonlinear DC Voltage Gradients
February 23, 2022
Adam Hollerbach, PNNL
Most ion mobility and mass spectrometers operate at low pressures (< 10 Torr) so that ions can be efficiently moved, manipulated, and analyzed. However, there's growing interest in performing experiments with ions at atmospheric pressure (AP) so that vacuum systems can be entirely avoided. The main problem when working with ions at AP is that commonly used high-voltage radio frequencies (RF) don't work at AP, and to date, few alternatives to RF exist.
This presentation will discuss a new way to keep ions focused at AP by applying nonlinear direct current voltage gradients across a conventional drift tube. Hollerbach will describe how ion trajectory simulations were used to understand how ions move in the presence of nonlinear electric fields at AP. He will then show how experimental results largely agree with simulations in demonstrating a true AP ion focusing effect. The results provide fundamental insights into how ions can move at AP and establish a unique foundation for exploring other ways to achieve lossless ion transmission in AP-based instruments.
Adam Hollerbach is a scientist in biological sciences at PNNL. He received his doctorate from Purdue University in 2018 and joined PNNL as a postdoc the same year. His early work at PNNL involved developing high-resolution ion mobility spectrometers for panomics applications. More recently he's developed new ways to move and manipulate ions at atmospheric pressure using static and dynamic electric fields.
Can Vibrational Spectroscopy Finally Find its Place in the World of Analytical Mass Spectrometry?
January 24, 2022
Dr. Thomas Rizzo, Ecole Polytechnique Federale De Lausanne
Infrared (IR) spectra of gas-phase ions provide detailed fingerprints that are sensitive to the minutest differences in molecular structure and hence can easily distinguish between isomeric species. Although practiced in many academic research laboratories, IR spectroscopy has not yet found its way into the world of analytical mass spectrometry. There are at least two obvious reasons for this: (1) the addition of a spectroscopic dimension to an analytical measurement has typically taken tens of minutes for each species, making it poorly suited to high-throughput analysis; and (2) the complex, expensive lasers required have made spectroscopic measurements impractical for biomedical research.
We have overcome these problems in an approach that combines ultrahigh-resolution SLIM-based ion mobility, cryogenic IR spectroscopy, and mass spectrometry in a single instrument. By increasing sensitivity and implementing a multiplexing approach to spectral measurement, we can measure an IR fingerprint spectrum of a molecule in as little as 15 seconds. Moreover, we do this using a simple, user-friendly, fiber-pumped IR laser no larger than a shoebox.
After demonstrating the capabilities of our technique, this talk will focus on its application in distinguishing isomeric glycans and glycan-related metabolites. We also have developed schemes to generate IR reference spectra starting from a relatively small number of simple, readily available standards from which we can grow a database for more complex species.
Dr. Thomas Rizzo is Professor of Chemistry at the Ecole Polytechnique Fédérale de Lausanne (EPFL) and has served as the Head of the Department of Chemistry (1997-2004) and the Dean of the School of Basic Sciences (2004-2017). His early research focused on fundamental studies of vibrational energy redistribution in gas phase molecules using a variety of laser spectroscopic techniques. More recently, he has been combining laser spectroscopy, ion mobility, and mass spectrometry for applications in biomolecular analysis.
Some recent distinctions for his work include the Bourke Award of the Faraday Society (2009), the Ron Hites Award from the American Society for Mass Spectrometry (2017), and an ERC Advanced Grant for his work in biomolecular analysis (2018). He was elected Fellow of the American Physical Society in 1998 and Fellow of the American Association for the Advancement of Sciences (AAAS) in 2011. He has also won teaching awards both at the University of Rochester (1992) and at EPFL (2021).
A Comprehensive Workflow for Lipid Double Bond Localization and Isomer Separation Using Ozone-Induced Dissociation and Ion Mobility Spectrometry
November 10, 2021
Xueyun Zheng, PNNL
Lipids play essential roles in many biological processes and disease pathology, yet it is still challenging to unambiguously identify lipids and distinguish numerous isomeric species which can result from different fatty acyl chain lengths, fatty acyl positions and carbon-carbon (C=C) double bond orientations and locations. Therefore, developing a lipidomics workflow that enables comprehensive isomer separation and structural elucidation is crucial for confident identification of lipids and for better understanding their roles in biology. In this work, I will present a comprehensive workflow integrating ozone-induced dissociation and ion mobility spectrometry-mass spectrometry (OzID-IMS-MS) approaches and robust bioinformatics tools developed to identify lipid double bond locations and distinguish isomers. Novel in-house bioinformatics tools are being developed to analyze OzID data and assign double bond positions using a machine learning approach. These tools will be integrated to allow robust analysis for complex lipidomic data, and our progress will be presented.
Xueyun Zheng is a scientist at the Biological Sciences Division. She received her PhD in Physical Chemistry from University of California Santa Barbara in 2015 and joined PNNL as a postdoc afterward. She has been developing ion mobility spectrometry in conjunction with mass spectrometry (IMS-MS) approaches to study biological and environmental systems and has applied these in the study of protein structures. More recently, she has focused on IMS-MS metabolomic and lipidomic workflows and development. Her research involves the development and evaluation of high resolution and high-throughput IMS-MS analyses to quickly study numerous samples in a short time period without losing valuable biological information.
New Frontiers of Ultraviolet Photodissociation Mass Spectrometry for Characterization of Proteins and Protein Complexes
October 13, 2021
Dr. Jennifer S. Brodbelt, University of Texas at Austin
Advances in mass spectrometry instrumentation and experimental design have led to significant inroads in the characterization of intact proteins and protein complexes, thus translating to new application in the field of proteomics and structural biology. Ultraviolet photodissociation (UVPD) is a new ion activation mode that results in broad sequence coverage of intact proteins via more extensive backbone fragmentation than obtained from other MS/MS methods, and ion activation/dissociation can be accomplished using a single 5 ns laser pulse. UVPD offers a frontier MS/MS technology for characterization of intact proteins, including mapping post-translational modification and ligand binding sites. There has been growing interest in employing MS/MS strategies to examine native protein structures by disassembling the complexes and sequencing the constituent proteins in the gas phase. In the context of protein-ligand complexes, the relative abundances of fragment ions generated by UVPD correlate with variations in the intramolecular and intermolecular interactions that stabilize particular regions of the proteins. Owing to the fast, high-energy deposition of UV photoactivation, products retaining non-covalently bound ligands are formed and afford binding site information. For multimeric protein complexes, UVPD disassembles the complex to reflect sub-unit architecture as well as generating sequence ions that identify the proteins.
Dr. Jennifer S. Brodbelt is the Rowland Pettit Centennial Professor of Chemistry at the University of Texas at Austin and is also serving as Chairperson. She earned her B.S. degree in chemistry at the University of Virginia and her doctorate in chemistry at Purdue University under the supervision of Prof. Graham Cooks. After a post-doctoral position at the University of California at Santa Barbara with Prof. Mike Bowers, she began her academic career at the University of Texas. Her research interests focus on the development and application of photodissociation mass spectrometry for characterization of the structures and modification of biological molecules, including peptides, proteins, nucleic acids, oligosaccharides, and lipids. She served as the Director of Graduate Education in the Department of Chemistry for over 20 years and recently became Chairperson in 2019. She servers as an Associate Editor for the Journal of the American Society for Mass Spectrometry, and she served as President of the American Society for Mass Spectrometry from 2014-2016.
Development of an Improved GC-MS Spectral Similarity and Retention Index Score Using Machine Learning
September 22, 2021
Dr. Chaevien Clendinen, PNNL
Gas chromatography mass spectrometry (GC-MS) is one of the most commonly used analytical platforms for metabolomics. Currently, GC-MS methods rely on internal libraries and online databases for metabolite identification. Identification of metabolites is one of the most critical and time-consuming steps in metabolomics analysis. Incorrect identification can lead to incorrect data interpretation and conclusions. Current practice is to use one metric to determine the ‘goodness’ of a potential reference match to queried spectra. Evaluation of the various spectral similarity metrics show that no single similarity metric performs optimally for all queried spectra. We therefore use a domain informed machine learning ensemble model to leverage the strengths of multiple similarity metrics into one optimal metric. Current retention index (RI) scoring methods assume that experimental and theoretical values follow the Gaussian distribution across different chromatograms with a user defined standard deviation for search tolerance. Testing of RI distributional assumptions shows that current assumptions of normality are incorrect for the majority of metabolites tested. We are challenging current RI scoring practices and assumptions and investigating the use of machine learning to improve retention index scores.
Dr. Chaevien Clendinen is an analytical chemist with the EMSD Biomolecular Pathways team. She is an expert in metabolomics data analysis and workflow development using a variety of analytical pipelines, including liquid chromatography mass spectrometry (LC-MS), gas chromatography MS (GC-MS), and nuclear magnetic resonance (NMR) spectroscopy. She is also an expert in structural elucidation using LC-MS and NMR.
Clendinen has 10 years of research experience across multiple disciplines, including analytical and physical chemistry, microbiology and virology, cancer biology, and biotechnology. Within EMSD, Dr. Clendinen assists EMSL users in the collection and analysis of metabolomics data from a variety of sample matrices. She is the current PI of a project under the PNNL m/q Initiative where she and her team aim to improve the confidence of identifications in LC- and GC-MS datasets, as well as develop a method to estimate false discovery rate.
Guided Ion-Beam Tandem Mass Spectrometry: The Basics and the Application
August 25, 2021
Dr. Peter B. Armentrout, University of Utah
In this talk, Peter will review the instrumental aspects that make a guided ion beam tandem mass spectrometer (GIBMS) a powerful tool for examining the kinetic energy dependence of absolute cross sections for ion-molecule reactions. It can be realized that such cross-section information, σ(E), is platform independent, which means it can be transformed into energy and temperature dependent rate constants or predict the output of any experimental study. Coupled with this instrument development are the tools for analyzing the results and extracting thermodynamic information. Examples from simple biological systems and metal ion chemistry will be provided.
Professor Armentrout received a B.S. degree in 1975 from Case Western Reserve University, Cleveland, Ohio. While at Case, Prof. Armentrout conducted research with Prof. Rob Dunbar on photodissociation spectroscopy of molecular ions which led to his interest in ion-molecule chemistry. He joined Prof. Jack Beauchamp at Caltech where he was awarded the Blanche A. Mowrer Memorial Fellowship. At Caltech, he constructed an ion beam apparatus designed to study hyperthermal reactivity of atomic uranium and other metal ions. Prof. Armentrout became a postdoctoral member of staff at Bell Labs in Murray Hill, New Jersey, where he studied electron impact ionization of metastable atoms and molecules.
Prof. Armentrout became an assistant professor at the University of California at Berkeley. He initiated a program which has come to study a wide spectrum of chemistries (primarily of transition metal species) by using ion-beam mass spectrometry. He joined the faculty at the University of Utah as an associate professor and was promoted to full professor in 1989. He has been recognized for several programs he instituted there over the years. In 1998, he was promoted to Distinguished Professor of Chemistry. Professor Armentrout is an emeritus member of the editorial board of the International Journal of Mass Spectrometry, and formerly of the Journal of the American Chemical Society, Journal of the American Society of Mass Spectrometry, Organometallics, Journal of Cluster Science, Journal of Physical Chemistry, and Journal of Chemical Physics. he is a member of the American Chemical Society, American Physical Society (fellow), American Society for Mass Spectrometry, and the American Association for the Advancement of Science (fellow).
Development of Computational Software for High Precision Collisional Cross Section Measurements
July 28, 2021
Dr. Christopher Harrilal, PNNL
Ion mobility is a well-established analytical technique that separates gas phase ions based on their size to charge ratio. Recent advancements in technology, mainly pioneered by the development of Structures for Lossless Ion Manipulations (SLIM), has allowed for ultra-high-resolution separations where resolving powers approach 700-1200. Under these conditions, ions with mobility differences of a few 100 parts-per-million can be resolved.
Of interest are the mobility differences observed between isotopomers under ultra-high-resolution conditions. Isotopomers are isotopic isomers that have identical mass/structure but differ in the position of their nuclides. As such, they are not expected to have mobility differences. Experimentally, however, mobility differences are observed and correlate well to changes in the ions center-of-mass (CoM) and moments-of-inertia (MoI). This indicates changes to rotational properties, arising from differences in the distribution of mass, that may result in subtle mobility differences. To explain the origin of these mobility differences we have developed a collisional cross section (CCS) calculator that accounts for the CoM and MoI of each isotopomer by modelling collisions between a rotating ion and neutral buffer gas. In most traditional codes, rotational motion is largely ignored. Our initial results successfully reproduced the experimental mobility shifts. Here we describe the framework of the code and provide an analysis of the scattering differences that lead to the mobility shifts. We ultimately aim to provide a tool that predicts mobility differences quickly and precisely between ions of similar mobility for future ultra-high-resolution mobility separations.
Dr. Christopher Harrilal is currently a Post-Doctoral research assistant in the Integrative Omics group at PNNL. His current research is directed towards the further development of SLIM technology, a platform designed for ion transportation and mobility separations via traveling wave ion mobility. Additional research efforts have been directed towards the detailed modeling of ion-neutral collisions to explain the mobility differences observed during ultra-high-resolution ion mobility separations. Chris received his PhD from Purdue University under the supervision of Professors Scott McLuckey and Tim Zwier. His graduate research involved using conformer specific ion spectroscopy for the structural elucidation peptide ions in the gas phase. The goal of this work was to understand the folding preferences of ions based on their primary sequence in a bottom-up approach. Chris has an expertise and significant background in the development of spectroscopic techniques, mass spectrometry, ion mobility, statistical mechanics, and molecular modeling.
Bioinformatics Tools for Integrative Functional Enrichment Analysis of Metabolomics Data
June 16, 2021
Dr. Alla Karnovsky, University of Michigan
Metabolomics and lipidomics generate increasingly large and complex datasets that require powerful statistical and bioinformatics tools. A well-established approach to linking alterations in metabolite levels to specific biological processes is to map experimentally measured metabolites to known biochemical pathways and to identify the pathways that are significantly enriched with those. However, traditional enrichment analysis techniques have limited utility for the analysis of untargeted metabolomics and lipidomics data.
We developed an alternative approach, which relies on extracting meaningful associations between metabolites/lipids directly from the experimental data. I will describe our Differential Network Enrichment Analysis method that is implemented in our new user-friendly tool Filigree. It uses joint structural sparsity estimation to build partial correlation networks from the data, performs consensus clustering to identify highly connected subnetworks, and uses Network-based Gene Set Analysis (NetGSA) to identify the differentially enriched subnetworks. I will discuss several applications of Filigree for the analysis of metabolomics and lipidomics data from type I and type II diabetes and other diseases. I will also describe other bioinformatics tools for untargeted metabolomics that we are developing.
Dr. Karnovsky got her Ph.D. in cell and developmental biology from the Russian Academy of Sciences. She did her postdoctoral work at the University of Colorado at Boulder, followed by nine years of bioinformatics work in the pharmaceutical industry. In 2007 she returned to academia. Currently she is an associate professor of computational medicine and bioinformatics at the University of Michigan. Her research interests involve the analysis of high throughput omics data, focusing primarily on metabolomics, and the development of computational methods and tools for the analysis and integration of metabolomics data with other types of omics data.
MQsim: A Trajectory Simulator for the Masses and Charges
May 26, 2021
Sandilya Garimella, PNNL
In MS, ions move from an ionization source (typically under atmospheric pressure conditions) to ion detection through differential pumping stages in the presence of radio-frequency (RF), static voltages (DC), traveling waves (TW) and background gas flow fields (Continuum to Free-Molecular Regime). While current state-of-art simulation platforms (e.g., SIMION, SIS Inc, Ringoes, NJ) enable ion motion in static DC and RF fields (with user programming), the motion of charged molecules in the presence of background gases has been limited largely to static gas molecules, with significant computational costs. Simulating the effect of gas dynamic flows, that are typical to an MS system, need interfacing between different commercial software with SIMION. These software (COMSOL, Ansys etc.) tend to have their respective learning curves making multi-physics modeling inherent to MS systems difficult to investigate. Furthermore, MS systems have a wide range of operational pressures, from atmospheric pressures to vacuum. Commercial gas flow solvers tend to be limited in their computing ability to a single regime from among the continuum, transitional or free molecular flow regimes. Therefore, there is a need for a simulation software that encompasses the full range of physics encountered within MS systems to enable simulation of ion trajectories across pressures scale and in varying fields. In this talk, a comprehensive ion trajectory modeling and simulation platform (MQsim) capable of simulating ion motion under RF, TW, DC fields, gas flows under non-reactive and reactive conditions will be presented. A suite of multiphysics solvers implemented thus far will be presented. MQsim simulations will be compared with and validated against theory and SIMION simulations. The development of a GUI interface to enable users to easily setup and run complex ion trajectory simulations will be presented.
Dr. Sandilya Garimella's research focuses on computational and experimental aspects of mass spectrometry technology development applied towards highly efficient chemical analysis and highly selective material synthesis. Garimella’s graduate work involved the development of ion sampling and ion transport technologies for miniaturized mass spectrometry systems. At PNNL, Garimella was involved in the design, development, and commercialization of Structures for Lossless Ion Manipulations (SLIM) technology, which is revolutionizing analytical separations applied to panomics applications and complex biological systems analysis. He was awarded the R&D 100 Award for his contributions. Additionally, he has several patents for inventions pertaining to mass spectrometry and ion mobility spectrometry. Garimella’s research interests include miniature and field deployable mass spectrometry technologies, mass and structure selective ion soft-landing and characterization, and ion trajectory simulations in multi-scale/multi-physics environments of mass spectrometry systems.
Quantum Chemistry Based Calculation of EI and CID Mass Spectra
April 21, 2021
Stefan Grimme, University of Bonn
The GFN-xTB family of efficient semi-empirical tight-binding methods combined with the Fermi-smearing technique can describe difficult electronic structures and covalent bond breaking at a reasonable accuracy level. This enables the 'first-principles' automated quantum chemistry computation of electron ionization mass spectra basically for any chemical compound including transition metal complexes. The approach originally dubbed QCEIMS and recently renamed to QCxMS (x=EI, CID) is based on high-temperature molecular dynamics simulations which are conducted in an ion-cascading mode starting from randomized initial conditions. Typically, the fragments from hundreds to thousands of trajectories are gathered and automatically transformed into a directly-to-experiment comparable spectral form. The approach considers basic elementary processes with minor empiricism, employs realistic potential energy surfaces computed 'on-the-fly', is 'black-box' and provides decent spectra accompanied by detailed information of corresponding decomposition and reaction mechanisms including rearrangements. Some example results as well as technical details of the EI and CID run types are discussed.
Stefan Grimme studied Chemistry and finished his PhD in 1991 in Physical Chemistry on a topic in laser spectroscopy. He did his habilitation in Theoretical Chemistry in the group of Sigrid Peyerimhoff. In 2000 he got the chair for Theoretical Organic Chemistry at the University of Muenster. In 2011 he accepted an offer as the head of the newly founded Mulliken Center for Theoretical Chemistry at the University of Bonn. He has published more than 540 research articles and is the recipient of the 2013 Schroedinger medal of the World Organization of Theoretically Oriented Chemists (WATOC). In 2014 he was awarded the "Gottfried Wilhelm Leibniz-Preis" from the DFG (endowed by 2.5 million Euro). His main research interests are the development and application of quantum chemical methods for large molecules, density functional theory, non-covalent interactions, and theoretical spectroscopy.
Using Computations to Probe Microsolvated Phases During Soft Ionization Processes in Mass Spectrometry
March 24, 2021
Samantha Johnson, PNNL
Identification of molecules and elucidation of their chemical structure are ubiquitous problems in chemistry. Mass spectrometry (MS) has frequently been applied for these purposes due to its sensitivity and versatility, as it can be applied to single compounds or complex mixtures. Ideally, MS patterns are chemical fingerprints that can be used to detect known compounds or characterize new molecules. However, the process is sensitive to the nature of parent medium (matrix effects) and instrument conditions, making unambiguous identification often difficult. Toward improved MS detection and analysis, we are developing atomistic computational approaches to understand and predict how the original environment and MS conditions regulate the ionization fate of analytes.
For detection to occur, analytes must be charged and must transfer from the condensed phase to the gas phase, in a process called ionization. Our current efforts focus on soft ionization processes, such as electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI), for detection of moderately polar and nonpolar analytes. Although these methods tend to retain the analyte’s structure, rather than fragmenting it as in hard ionization approaches, chemical transformations can still occur in the microsolvated phase between condensed and gas phases and may even be accelerated relative to the condensed phase. To understand these processes, we use computational methods (density functional theory, molecular dynamics, and enhanced sampling approaches) to probe ionization propensity of analytes, as well as microsolvated cluster energetics and structures. In this talk, I will discuss how computational methods provide precious insights on how the nature of parent condensed phase influences the fate of an analyte in the gas phase (the so-called matrix effects). Using amino acids in water clusters as a testbed, I will show how microsolvated environments dictate which species enter the mass spectrometer.
Dr. Samantha Johnson is a Computational Scientist at Pacific Northwest National Laboratory. She studies molecular electrocatalysts and clusters using theoretical and computational methods. Dr. Johnson has a B.S. in Chemical Engineering from University of Colorado, Boulder. She received her Ph.D. in materials science from California Institute of Technology in 2017 and was a postdoctoral researcher in the Center for Molecular Electrocatalysis at Pacific Northwest National Laboratory. She was a National Science Foundation Graduate Fellow and a Resnick Fellow. Dr. Johnson joined PNNL as staff in 2019. Here, her work has focused on ammonia oxidation, hydrogen storage on liquid carriers, and solvation effects in mass spectrometry. In particular, she is interested in the role of a molecule’s surroundings in its performance and behavior.
Panoptic Mass Spectrometry: How and Why?
February 24, 2021
Dr. Abraham Badu-Tawiah, The Ohio State University
The complexity of contemporary research requires interdisciplinary efforts, and mass spectrometry (MS) is poised to play an important role. The presentation is intended to demonstrate recent MS experiments designed to facilitate the detection of every ionic species originating from the ion source and an on-demand diagnostic strategy for underserved groups that combines new levels of simplicity, modest cost, and a centralized detection strategy for accurate disease detection. I will focus on asymptomatic malaria detection in the developing world and chemical analysis from ultra-small sample volumes collected from pediatric patients.
Abraham Badu-Tawiah obtained his Ph.D. (2012) in chemistry from Purdue University under the supervision of Graham Cooks. From 2012 to 2014, he was a postdoctoral fellow at Harvard University under the direction of George Whitesides. He joined The Ohio State University, Department of Chemistry and Biochemistry, in July 2014 as an Assistant Professor. In June 2020, Dr. Badu-Tawiah was promoted to Associate Professor with tenure. Dr. Badu-Tawiah is a recipient of the 2020 Sloan Fellowship Award, the 2019 NIH MIRA for New Investigators Award, the 2018 ACS Division of Analytical Chemistry Arthur F. Findeis Award, the 2017 Eli Lilly Young Investigator Award in Analytical Chemistry, the 2017 American Society for Mass Spectrometry Research Award, and the 2016 Department of Energy Early Career Award. His current research is focused on the development of new mass spectrometry techniques for disease detection, and the studies of novel ion chemistry in charged micro-droplets
Reaction Roulette: Utilizing Elemental MS/MS for the Characterization of Gas Phase Ion-Molecule Interactions
January 7, 2021
Dr. Khadouja Harouaka, PNNL
Tandem mass spectrometry (MS/MS) is a fairly recent addition to the analytical chemist’s toolkit for elemental analysis. Elemental MS/MS instruments have two mass filters that sandwich a collision/reaction cell that can be pressurized with a variety of reaction gases. The key analytical advantage of this feature is that it allows for the inline separation of analytes from interference ions through gas phase ion-molecule reactions, which would otherwise have to be achieved offline through lengthy and sometimes complicated ion exchange chemistry. Instrumentation capable of elemental MS/MS, such as the Agilent 8900 QQQ ICP-MS and the Thermo iCap TQ ICP-MS have only been commercially available since 2012, which leaves plenty of opportunity for fundamental research in gas phase ion-molecule reactivity that will ultimately be leveraged for method development.
In this talk, we will discuss our work in exploring the reactivity of 42 elements representative of the periodic table with N2O and CO2 using the Agilent 8900 QQQ-ICP-MS. Reasonable correlations between independent DFT derived reaction enthalpies, calculated using NWChem, and the reaction data were found, confirming that reactivity can be predicted by basic thermodynamics.
Khadouja Harouaka has been a postdoc in NSD for the last two years. She received a dual title PhD in geoscience and astrobiology from Penn State in 2016 and did a postdoc in petroleum geochemistry at Rice University before joining the Ultra-Low Background detection physics group at PNNL. Here, she uses ICP-MS/MS to measure ultra-trace quantities of radio-contaminants in detector materials for high energy and nuclear physics “rare event” experiments (e.g., dark matter detection, neutrinoless double beta decay). She dedicates the “reaction roulette” project to any and all grad students and scientists who have spent far too many days doing ion exchange column chemistry.
Illuminating the Dark Lipidome by Isomer-Resolved Mass Spectrometry
December 16, 2020
Dr. Stephen J. Blanksby, School of Chemistry & Physics, Queensland University of Technology
Mass spectrometry is emerging as a powerful tool for connecting lipid identity with cellular metabolism. Conventional approaches to lipid mass spectrometry have been challenged however, by the inability to discriminate between lipid regioisomers. These lipids, differing only in the location of key functional groups, have (by definition) identical mass-to-charge (m/z) ratios; present similar, or identical, product ions upon collision-induced dissociation; and can go unresolved by conventional chromatography. In the absence of diagnostic mass spectrometric or chromatographic signatures, lipid isomers often go undetected leading to an under-reporting of biomolecular complexity and masking changes in lipid metabolism occurring within this "dark" lipidome.
To meet this challenge, we have introduced ozone-induced dissociation (OzID) as an alternative ion activation modality that is capable of discriminating between lipids regioisomers differing in site(s) of unsaturation or substitution (sn-)position on the glycerol backbone. In this presentation, we will discuss advances stemming from the implementation of OzID in the high-pressure ion-trapping regions of different mass spectrometer geometries. The resulting improvements in sensitivity have enabled integration into high (and ultra-high) resolution mass spectrometry, liquid chromatography-mass spectrometry and imaging-mass spectrometry. Examples from each of these modalities will be presented along with their implications for isomer-resolved lipidomics in illuminating the plasticity of lipid metabolism in cancer.
Stephen Blanksby completed his PhD in the field of gas phase ion chemistry at the University of Adelaide under the supervision of Professor John Bowie. He then undertook postdoctoral research in the laboratories of Professor Helmut Schwarz at the Technical University of Berlin and at the University of Colorado with Professor G. Barney Ellison and the rest of the famed “Boulder ion gang.” Stephen was appointed to the School of Chemistry at the University of Wollongong in 2002 where he built a research group focused on putting the “fun” back into fundamental ion chemistry. Discoveries made by Stephen’s group have been successfully applied to develop new analytical technologies for lipidomics and polymer coatings research. In 2014, Stephen joined QUT as Director of the Central Analytical Research Facility (CARF) and he continues an active research program exploring the unique properties of gas phase radical ions and developing new technologies for advancing mass spectrometry.
Computationally-Driven Identification of Small Molecules in Biological Systems
November 18, 2020
Dr. Ryan Renslow, PNNL
The identification of metabolites from complex biological samples often involves matching experimental mass-spectrometry data to signatures of compounds derived from chemical databases. However, misidentifications may result due to the complexity of potential chemical space that leads to databases containing compounds with nearly identical structures. Prior knowledge of compounds that may be enzymatically consumed or produced by an organism can help reduce misidentifications by restricting initial database searching to compounds that are likely to be present in a biological system. While databases such as UniProt allow for the identification of small molecules that may be consumed or generated by enzymes encoded in an organism's genome, there is no tool available that can easily identify SMILES strings of metabolites associated with protein identifiers and expand R-containing substructures to fully-defined biologically relevant chemical structures. In this talk, I will present P2M, a tool that performs these tasks using local cross-referencing and optional external database querying with a simple command-line interface. Furthermore, I will discuss an overview of our teams’ efforts to advance standards-free and library-free identification in complex biological systems.
Dr. Ryan Renslow has been at PNNL for eight years and is currently the Molecular Analytics Team Lead in the Earth and Biological Sciences Directorate and is an Associate Research Professor in the chemical engineering department at Washington State University. He has expertise in mathematical modelling, with an emphasis on biological systems, molecular modelling, quantum chemical calculations, and deep learning. His near-term career vision is to transform chemical biology through a paradigm shift in analytical chemistry by reducing or eliminating reliance on authentic chemical standards. His multi-decadal vision is to leverage the new information created by being able to identify hundreds-of-thousands of molecules, in order to simulate cellular life processes at a molecular scale, accelerating the pace of discovery and reducing the need for laboratory experimentation.
Quantum Chemistry and Machine Learning for Improved Compound Identification in Computational Mass Spectrometry
October 28, 2020
Dr. Tobias Kind, UC Davis Genome Center
Mammalian, plant, and bacterial cells contain thousands of metabolites created by a complex biological machinery. However, many of these natural products remain unknown, despite advanced analytical technologies such as high-resolution chromatography coupled to mass spectrometry. Our current research tries to tackle this obstacle by using theoretical models from quantum chemistry including Born-Oppenheimer molecular dynamics and ensemble machine learning approaches including deep-learning and boosting. In my talk I will focus on our latest advances in modelling CID-MS/MS spectra, computation of 70 eV mass spectra, generation of new in-silico libraries and retention time modelling for liquid chromatography. The importance of interdisciplinary work that requires input from many experts across different fields and therefore goes beyond classical analytical techniques will be outlined.
Tobias Kind, Ph.D., leads the cheminformatics team and computational core at the NIH West Coast Metabolomics Center for Compound Identification at UC Davis. His research projects focus on the advancement of structure elucidation techniques and databases for small molecules, metabolites and lipids, and his research interests include, machine learning and deep-learning for compound identification, quantum chemistry and in-silico spectra, and Born-Oppenheimer ab-initio molecular dynamics.