An International Symposium of the Association of Biomolecular Resource Facilities

Proteomics Oral Presentations

Primary tabs

Session Type: Plenary

  • Chemical Proteomics Reveals the Target Space of Hundreds of Clinical Kinase Inhibitors
    Bernhard Kuster Technical University Of Munich
    Kinase inhibitors have developed into important cancer drugs because de-regulated protein kinases are often driving the disease . Efforts in biotech and pharma have resulted in more than 30 such molecules being approved for use in humans and several hundred are undergoing clinical trials. As most kinase inhibitors target the ATP binding pocket, selectivity among the 500 human kinase is a recurring question. Polypharmacology can be beneficial as well as detrimental in clinical practice, hence, knowing the full target profile of a drug is important but rarely available. We have used a chemical proteomics approach termed kinobeads to profile 240 clinical kinase inhibitors in a dose dependent fashion against a total of 320 protein kinases and some 2,000 other kinobead binding proteins. In this presentation, I will outline how this information can be used to identify molecular targets of toxicity, re-purposing existing drugs or combinations for new indications or provide starting points for new drug discovery campaigns.
    Session Type: Plenary

Session Type: Research Group

  • Workflow Interest Network Research Project Presentation
    Emily Chen Columbia University Medical Center
    A new research group, the Workflow Interest Network (WIN), was established in 2016. Our current focuses are 1) to collaborate with other ABRF members and mass spectrometry-based research groups to identify key factors that contribute to poor reproducibility and inter-laboratory variability, and 2) to propose benchmarks for MS-based proteomics analysis as well as quality control procedures to improve reproducibility. In 2016, we have launched a test study to examine the LC-MS/MS performance among 10 MS-based core laboratories, using two sets of peptide standards and complex lysates. Preliminary analysis of the test study will be presented in the concurrent workshop. We are highly encouraged by the results of our test study. We believe that it should be possible to promote scientific reproducibility by comparing different analytic platforms and providing benchmarks for instrument performance based on the observed capabilities across a large number of datasets. Bioinformatics tools that allow rapid analysis and evaluation of LC and MS performance will also be discussed. Finally, we will announce an expansion of our test study to the broader community, inviting laboratories to participate and contribute to this benchmarking study. The study is designed to require only a very reasonable time commitment, and participating laboratories will gain essential information on their instruments’ performance in the course of helping to build a valuable benchmarking and QC resource.
    Session Type: Research Group
  • The iPRG-2016 Proteome Informatics Research Group Study: Inferring Proteoforms from Bottom-up Proteomics Data
    Magnus Palmblad Leiden University
    In 2016, the ABRF Proteome Informatics Research Group (iPRG) conducted a study on proteoform inference from bottom-up proteomics data. For this study, we acquired data from samples spiked with overlapping oligopeptides, so-called Protein Epitope Signature Tags, recombinantly expressed in E. coli into a background of E. coli proteins. This is a unique dataset in that we have ground truth on the proteform composition of each sample, and each sample contains hundreds of different proteoforms. Tandem mass spectra were acquired on a Q-Exactive Orbitrap in data-dependent acquisition mode and made available in raw and mzML formats. Participants were asked to use a method of their choosing and report the false-discovery rate or posterior error probabilities for each proteoform in a list provided in the FASTA format. In general, most participants solved the task well, but some differences were observed, suggesting possible improvements and refinements. This time we also decided to try to show different ways to improve and add value to such studies. We therefore created a dedicated website on a virtual private server to contain all aspects the study, including the submission interface. We had learned from previous studies that it is difficult for some participants to submit their results in the desired format, no matter how carefully specified. This has generated substantial additional work during the evaluation phase of previous iPRG studies. In the 2016 study, we therefore built a submission parser/validator that checked whether a submission was provided in the correct format before accepting it. The validator also provided feedback to the submitter when the uploaded results were not in the correct format. This also helped in the evaluation and comparisons of submissions. Another novelty in the 2016 study was the acceptance of submissions in R Markdown or IPython notebook formats, containing both explanation of the method and an executable script to rerun and compare submissions. These methods will be anonymized and made available for reuse by researchers conducting this type of analyses. The study will therefore live on, and as new methods and software become available, these can be benchmarked against the best solutions at the time of the study. It is also possible to combine elements from several submitted methods. All code running on the VPS behind the iPRG 2016 Website, including the submission validator, will be available to other ABRF Research Groups. In the presentation, we will also discuss lessons learned from the novel technical aspects of this iPRG study, including the use of R Markdown and IPython notebooks.
    Session Type: Research Group
  • MS1-based quantification of low abundance proteins that were not identified using a MS/MS database search approach
    Yan Wang University Of Maryland
    With development in instrumentation and informatics tools, it is becoming routine to identify more than 2000 proteins in whole cell lysates via a shotgun proteomics approach. A remaining challenge is to assess changes in abundance of proteins that are at the limit of detection. In the data dependent analysis (DDA) approach the same peptide is often identified in one sample but missed in another. Presumably the missing data is due to the precursor not being isolated for fragmentation. In such cases, relative quantification could be determined from the MS1 peak intensity after using features such as accurate mass and retention time to identify the correct MS1 signal. In this study, we evaluated 4 different bioinformatic approaches in their ability to perform this analysis. In the PRG 2016 study, 4 non-mammalian proteins were spiked into 25 µg of whole HeLA cell lysate at 4 different levels: 0, 20, 100, and 500 fmols. Six non-fractionated datasets encompassing 4 separate analytical runs each and including analyses on Orbitrap Fusion, Q Exactive, and Orbitrap Velos instruments were selected for further study. In each set at least 1 peptide from each of the 4 spiked-in proteins was identified in at least one sample. Peptides from spiked-in proteins were quantified after using retention time and accurate mass information for identification. Programs used were: PEAKS (Bioinformatics Solutions, Inc.); Progenesis (Waters Corp.); MaxQuant (Max Planck Institute of Biochemistry) and Skyline (University of Washington). All evaluated software packages extracted quantitative information from MS1 spectra that did not yield Peptide Spectral Matches in samples with low concentrations of spike-in proteins. False quantification of peptides in the zero spike-in sample was observed. This is attributed to carry-over between runs and mis-assignment of noise in the signal.
    Session Type: Research Group
  • sPRG-ABRF 2016-2017: Development and Characterization of a Stable-Isotope Labeled Phosphopeptide Standard
    Antonius Koller Columbia University
    The mission of the ABRF proteomics Standards Research Group (sPRG) is to design and develop standards and resources for mass-spectrometry-based proteomics experiments. Recent advances in methodology have made phosphopeptide analysis a tractable problem for core facilities. Here we report on the development of a two-year sPRG study designed to target various issues encountered in phosphopeptide experiments. We have constructed a pool of heavy-labeled phosphopeptides that will enable core facilities to rapidly develop assays. Our pool contains over 150 phosphopeptides that have previously been observed in mass spectrometry data sets. The specific peptides have been chosen to cover as many known biologically interesting phosphosites as possible, from seven different signaling pathways: AMPK signaling, death and apoptosis signaling, ErbB signaling, insulin/IGF-1 signaling, mTOR signaling, PI3K/AKT signaling, and stress (p38/SAPK/JNK) signaling. We feel this pool will enable researchers to test the effectiveness of their enrichment workflows and to provide a benchmark for a cross-lab study. Currently, the standard is being tested in the sPRG members' laboratories to establish its properties. Later this year we will invite ABRF members and non-members to participate in the second half of our study, using this controlled standard in a HeLa S3 background to evaluate their phosphoproteomic data acquisition and analysis workflows. We hope this standard is helpful in a number of ways, including enabling phosphopeptide sample workflow development, as an internal enrichment and chromatography calibrant, and as a pre-built biological assay for a wide variety of signaling pathways.
    Session Type: Research Group
  • iPRG2016: New submission interface and its ideas
    JOON-YONG LEE Pacific Northwest National Laboratory
    For the annual Proteome Informatics Research Group (iPRG) Study, the submission has traditionally been done by participants ftp uploads. Although this is simple for participants, it does not support format validation. Because results are not required to follow a specific format, method description was incomplete, which limited reproducibility of the study. To tackle these issues, we present how we set up the iPRG2016 website in this presentation. We will describe how we build a new submission interface by adopting GitHub interaction and format validator to improve the automated submission process. Moreover, we will show how python notebooks and R markdown documents under a GitHub repository help to readily and transparently share methodologies and promote continued community participation.
    Session Type: Research Group

Session Type: Scientific Session

  • Chemical Isotope Labeling LC-MS for Routine Quantitative Metabolomic Profiling with High Coverage
    Liang Li University Of Alberta
    A key step in metabolomics is to perform relative quantification of metabolomic changes among different samples. High-coverage metabolomic profiling will benefit metabolomics research in systems biology and disease biomarker discovery. To increase coverage, multiple analytical tools are often used to generate a combined metabolomic data set. The objective of our research is to develop and apply an analytical platform for in-depth metabolomic profiling based on chemical isotope labeling (CIL) LC-MS. It uses differential isotope mass tags to label a metabolite in two comparative samples (e.g., 12C-labeling of an individual sample and 13C-labeling of a pooled sample or control), followed by mixing the light-labeled sample and the heavy-labeled control and LC-MS analysis of the resultant mixture. Individual metabolites are detected as peak pairs in MS. The MS or chromatographic intensity ratio of a peak pair can be used to measure the relative concentration of the same metabolite in the sample vs. the control. For a metabolomics study involving the analysis of many different samples, the same heavy-labeled control is spiked to all the light-labeled individual samples. Thus, the intensity ratios of a given peak pair from LC-MS analyses of all the light-/heavy-labeled mixtures reflect the relative concentration changes of a metabolite in these samples. CIL LC-MS can overcome the technical problems such as matrix effects, ion suppression and instrument drifts to generate more precise and accurate quantitative results, compared to conventional LC-MS. CIL LC-MS can also significantly increase the detectability of metabolites by rationally designing the labeling reagents to target a group of metabolites (e.g., all amines) to improve both LC separation and MS sensitivity. In this presentation, recent advances in CIL LC-MS for quantitative metabolomics will be described and some recent applications of the technique for disease biomarker discovery research as well as biological studies will be discussed.
    Session Type: Scientific Session
  • Characterization of the myometrial proteome in disparate states of pregnancy using SPS MS3 workflows.
    David Quilici University Of Nevada Reno
    Reliable quantitative analysis of global protein expression changes is integral to understanding mechanisms of disease. Global expression of myometrial proteins involved in the premature induction of labor compared to normal induction of labor were analyzed by isobaric labeling with a TMT 10-Plex using MultiNotch MS3. In this study we looked at technical and biological reproducibility in addition to the comparison of pre-term labor to term labor myometrial tissue samples. We found a very low level of variation in the technical (<0.01%) and biological (0.05%) replicates. Within the comparative study we identified over 4,000 protein groups with high confidence (FDR < 0.05) and ~400 of these showed a significant change between the two groups. Affected pathways were then identified using Ingenuity pathway analysis software (IPA). Further analysis was performed using the targeted TMT approach known as TOMAHAQ (Triggered by Offset, Multiplexed, Accurate-Mass, High-Resolution, and Absolute Quantification) on 38 peptides and phosphopeptides corresponding to proteins within the identified pathways affected by premature induction in an effort to determine the role of phospho-signalling.
    Session Type: Scientific Session
  • An ‘Omics Renaissance or Stuck in the Dark Ages? Monitoring and Improving Data Quality in Clinical Proteomics and Metabolomics Studies
    J. Will Thompson Duke Proteomics And Metabolomics Shared Resource
    Analysis of proteins and metabolites by mass spectrometry is currently enjoying a renaissance in many ways. Identification of unknown proteins, and even localization of post-translational modifications with high precision and on a grand scale, is routine. It is possible to qualitatively and quantitatively profile thousands of protein groups, or metabolite features, in a single sample with a single analysis. Multiplexed targeted LC-MS/MS can quantify hundreds of analytes in only a few minutes. Exciting new sample preparation, mass spectrometry, and data analysis techniques are emerging every day. However, significant challenges still exist for our field. Compared to genomic approaches, coverage is still limited. Compared to the longitudinal precision and accuracy required in the clinical diagnostic use of mass spectrometry, ‘omics techniques lag far behind. And in terms of the throughput required for addressing the Precision Medicine Initiative, most proteomics and metabolomics analyses take far too long and are far too expensive. This presentation will focus on practical techniques implemented in our laboratory and others to try and address some of these shortcomings, including efforts to improve the throughput of unbiased proteomics profiling, track precision and bias in proteomic and metabolomic experiments using the Study Pool QC, and the use of reference pools for longitudinal performance monitoring in targeted quantitative proteomics and metabolomics studies.
    Session Type: Scientific Session
  • Patching Holes in Your Bottom-up Label and Label-free Quantitative Proteomic Workflows
    Tony Herren University Of California, Davis
    Reliable quantitation of label and label-free mass spectrometry (MS) data remains a significant challenge despite much progress in the field on both the hardware and software fronts. In particular, proper quantitation and control within the workflow prior to mass spectrometry data dependent acquisition (DDA) is of crucial and often overlooked importance. Here, we will explore several of these “pre-mass spec” topics in greater detail, including quality control and optimization of sample preparation and liquid chromatography. Specifically, the importance of accurate quantitation of protein and peptide inputs and outputs during sample processing steps (extraction, digestion, C18 cleanup, enrichment/depletion) for both label and label-free workflows will be addressed. Protein recovery data from our lab suggests that recoveries throughout the sample preparation process come with significant loss and must be empirically determined at each step. Missed cleavages during enzymatic digestion, poor labeling efficiency, and addition of enrichment and/or depletion steps to sample workflows are all confounding factors for label and label-free quantitation and will also be discussed. Additionally, the influence of liquid chromatography performance on DDA quantitation across multiple sample runs in label and label-free workflows will be examined, including the effects of retention time drift, ambient environmental conditions, gradient length, peak capacity, and instrument duty cycle. These issues will be discussed in the context of a typical core facility bottom up proteomics workflow with practical tools and strategies for addressing them. Finally, a comparison of quantitative sensitivity will be made between sample data acquired using label-free MS and tandem mass tag (TMT 10-plex) data acquired using different MS parameters on a Thermo Fusion Lumos.
    Session Type: Scientific Session
  • Systems Biology Guided by Metabolomics
    Gary Siuzdak Scripps Center For Metabolomics And Mass Spectrometry
    Systems-wide analysis has been designed and implemented into our cloud-based metabolomic platform (XCMSOnline.scripps.edu and METLIN.scripps.edu) to guide large scale multi-omic experiments. This data streaming autonomous approach superimposes metabolomic data directly onto metabolic pathways, which is then integrated with transcriptomic and proteomic data. To date, the utility of this platform has been demonstrated on thousands of studies and implemented within XCMS' smartphone app. Here I will focus on the technology and demonstrate its utility and the insight it has provided in examining therapeutic remyelination in multiple sclerosis and neurodegeneration in HIV.
    Session Type: Scientific Session
  • Spatial Metabolomics via MALDI Imaging Mass Spectrometry: A Case Study in Lysosomal Storage Disease, Gangliosides and Gene Therapy
    Scott Shaffer University Of Massachusetts Medical School
    Gangliosides are glycosphingolipids composed of a ceramide base and a carbohydrate chain containing one or more sialic acids. GM1 gangliosidosis is an autosomal recessive lysosomal storage disorder caused by an enzyme deficiency of β-galactosidase leading to toxic accumulation of GM1 in the central nervous system and progressive neurodegeneration. Gene therapy mediated delivery of viral vector encoding enzyme has shown great potential for the treatment of such diseases by restoring deficient enzyme levels. In this work MALDI imaging mass spectrometry is used to measure the spatial distribution of gangliosides, ganglioside metabolites and lipids in a GM1 gangliosidosis mouse brain model, including animals following adeno-associated virus (AAV) gene therapy. Data was acquired with a Nd:YAG laser at 60µm spatial resolution using a Waters Synapt G2-Si MALDI mass spectrometer with integrated travelling wave ion mobility separation. Overall, we demonstrate an approach that measures gangliosides and their metabolites with high molecular specificity while also offering the ability to detect unanticipated, off-targeted effects induced by both disease and by gene therapy. Key aspects of building a tissue imaging capability within a core facility will be discussed.
    Session Type: Scientific Session

Platinum Sponsors of the ABRF