KIMBLE: A Versatile Visual NMR Metabolomics Workbench in KNIME
Aswin Verhoeven, Martin Giera, Oleg Mayboroda Leiden University Medical Center, Leiden, the Netherlands
Warning: Undefined variable $headshot in /var/www/html/view_abstract/view_abstract_in_program.php on line 704
Aswin Verhoeven (Presenter) Leiden University Medical Center
Presenter Bio: Aswin Verhoeven is a senior researcher at the Center for Proteomics and Metabolomics (CPM) at the Leiden University Medical Center (LUMC) in Leiden, the Netherlands. He obtained his Masters degree in chemistry at the Radboud University in Nijmegen. He received his PhD with his work on solid-state NMR at the Laboratory of Physical Chemistry at the ETH Zurich. After his post-doctoral research at the Max-Planck Institute for Solid-State Research in Stuttgart, Germany. In Leiden, his research focus is on the data processing automation of NMR metabolomics data.
Relevant Financial Disclosures
(within past 24 months)
No relevant financial relationship(s) to disclose.
Abstract
INTRODUCTION: Processing and analysis of NMR metabolomics data typically involves a number of different software packages and that are partly controlled manually, limiting the reproducibility of the data processing workflow.
OBJECTIVES: Construct an integrated workflow that makes NMR data processing fully reproducible without sacrificing flexibility and extendability.
METHODS: A workflow was implemented in the free and open-source KNIME Analytics Platform, and distributed in the form of easy-to-install virtual machine. KNIME shows the workflow as a visually appealing graph that can be arranged by the user. With the choice for KNIME as a platform the user can extend the metabolomics workflow by making use of the large toolbox of data processing and machine learning tools that KNIME offers. R and Python scripts can be easily integrated in the workflow. Settings, algorithms and data are kept consistent; changes in the algorithms or setting lead to a reset of the results.
RESULTS: KIMBLE (the the KNIME-based Integrated MetaBoLomics Environment) offers a reproducible, flexible and extendible way of processing NMR metabolomics data. KIMBLE converts time-domain NMR data to a set of metabolite concentrations and complementary untargeted data and performs subsequent statistical analysis. Although adaptable to many kinds of mixtures, the published KIMBLE workflow is optimized for urinary metabolomics. Now KIMBLE has been optimized for serum/plasma NMR data as well, including the ability to import results from Buker IVDr analyses. Furthermore, KIMBLE is now fully integrated with the Human Metabolome Database (HMDB) and the Europe PMC scientific literature repository, opening interesting new alleys for data analysis. Examples for the use of KIMBLE for exercise-induced ketosis and polycystic kidney disease (PKD) will be shown.
CONCLUSION: KIMBLE is a comprehensive workflow for the conversion of raw NMR data into metabolic insights. By maintaining a tight connection between settings, algorithms and data, total reproducibility of the data processing is achieved.
= Discovery stage.
= Translation stage.
= Clinically available.
