MetaboMiner

Category Metabolomics/Metabonomics>Metabolic Profiling/Analysis Systems/Tools

Abstract MetaboMiner is a freely available, easy-to-use, NMR-based metabolomics tool that facilitates automatic peak processing, rapid compound identification, and facile spectrum annotation from either 2D TOCSY or HSQC spectra.

Using comprehensive reference libraries coupled with robust algorithms for peak matching and compound identification, MetaboMiner greatly simplifies the process of metabolite identification in complex 2D NMR spectra.

It can be used to automatically or semi-automatically identify metabolites in complex biofluids from 2D NMR spectra. MetaboMiner is able to handle both 1H-1H total correlation spectroscopy (TOCSY) and 1H-13C Heteronuclear single quantum correlation (HSQC) data.

MetaboMiner identifies compounds by comparing 2D spectral patterns in the NMR spectrum of the biofluid mixture with specially constructed libraries containing reference spectra of ~500 pure compounds. Tests using a variety of synthetic and real spectra of compound mixtures showed that MetaboMiner is able to identify >80% of detectable metabolites from good quality NMR spectra.

MetaboMiner Data Collection and Curation --

Key to the development of this software package was the creation of an extensive 2D spectral library containing TOCSY and HSQC spectra of pure metabolites.

The manufacturer’s used several publicly available sources in constructing this library. The majority of the raw 1H-1H TOCSY spectra were collected from the standard compound spectral library available at the BioMagResBank (BMRB) database.

A few additional compound spectra were obtained from the Magnetic Resonance Metabolomics Database (MRMD). The 1H-13C HSQC spectral library was downloaded from the Human Metabolome Database (HMDB).

These raw spectra contained a number of spectral artifacts (noise, water bands, asymmetries, peaks from TSP or DSS, contaminants, etc.). Consequently it was necessary to convert these raw spectra into “synthetic” or “simplified” spectra corresponding to the peaks specific to the pure compounds of interest.

This conversion was done manually, with each of these simplified, noise-free spectra being examined for inconsistencies by comparing them to the original raw spectra and the compound’s known resonance assignments.

After the spectral libraries were constructed, each peak for each compound in each library was assigned a series of uniqueness values that are specific for that reference library.

MetaboMiner Peak Processing, Peak Matching, and Compound Identification --

As part of its input, MetaboMiner requires peak lists corresponding to the peaks that were identified in either the TOCSY or HSQC spectra collected from the biofluid(s) of interest. While it is possible for users to provide manually picked peak lists, MetaboMiner also supports processing of multidimensional NMR peak lists obtained from automatic peak peaking programs.

In order to accommodate small chemical shift differences between the observed NMR spectra and the reference NMR spectra, an adaptive threshold method was implemented based on the uniqueness values of each reference peak.

During the peak searching/matching process, the search threshold varies automatically based on the maximum uniqueness value of the current peak. The peak matching and adaptive thresholding employ two (2) processes: a reverse search strategy and a forward search strategy.

In the reverse search strategy, the library peaks are searched and matched against the query peaks. Typically most query peaks find their potential matches during this reverse search step. However there are usually some peaks left without any matches.

In order to assign these unmatched peaks a forward search is performed in which the unmatched query peaks are searched against the reference library with expanded thresholds. A match is identified if only a single reference peak is identified within this range.

In MetaboMiner a compound is considered to be present only if its matched pattern satisfies the requirements of what the manufacturer’s call “minimal signatures”.

A minimal signature is defined as the minimum peak set that can uniquely identify a compound from all others in a given spectral library.

MetaboMiner User Interface Description --

MetaboMiner’s graphical user interface (GUI) was implemented using Java Swing technology. The spectral visualization and manipulation tools were built using the JGraph library. There are four (4) main functional views:

1) A Processing View;

2) A Search View;

3) An Annotation View; and

4) A Library View.

All these views share the same component arrangement, with panels on the right side being used for visualizing and manipulating peaks, and the panels on the left being used for displaying parameters, compound lists, structure images, etc. Navigation to each view is readily accessible by clicking an appropriate menu item.

When the program launches, the default view is the “Processing View” where users can copy and paste the automatically picked peak list.

After processing the raw peaks, both the original and the processed spectra will be displayed on MetaboMiner’s spectral viewing panel. With this viewing panel, users can directly edit peaks on the spectrum if necessary.

By clicking the “Search” button, MetaboMiner’s “Search View” will be displayed with its initial, automated compound identification results. Users can adjust the search threshold or switch the reference library to further refine the result.

A compound is marked as identified if the matched pattern passes the authenticity checks and satisfies the minimum signature requirement. The raw matched scores are also displayed.

MetaboMiner’s interface allows users to visually inspect the matched peaks of any metabolite against the corresponding reference spectrum.

Users can further refine the automated search results by manually annotating the raw 2D spectrum.

By clicking the “Refine” button in the “Search View”, the “Annotation View” will be launched with the identified compounds being transferred as the starting point. Users can also directly enter the “Annotation View” mode by clicking the “Annotate” button from the “Console” menu.

In order to perform manual annotation, users first need to load a high resolution spectral image in PNG format and set up the spectral axes properly.

Peak searching is performed by right clicking the peak position on the spectrum to search the reference library.

All compounds that generate peaks within the search threshold will be checked. The compound with the closest peak match will be highlighted with its database reference spectrum displayed on the uploaded “raw” spectrum. Users can perform peak annotation for any currently displayed compound.

The “Library View” is intended for browsing and managing MetaboMiner’s spectral libraries.

To view all the available reference spectra in MetaboMiner’s libraries, users must click the “Browse” button in the “Library” menu. Double clicking any compound in the compound list will open a popup window for peak editing. Any changes will be reflected on the spectrum at real time.

For researchers who study other types of biological samples (e.g. plant or microbial extracts), they may either use MetaboMiner’s generic spectral reference library or create a new library customized for that particular type of biofluid.

Library creation or deletion can be easily accomplished by clicking the appropriate menu items in the “Library” menu. The compounds in the default reference library are linked to PubChem, HMDB, and the BMRB via the hyperlink under their structure icon.

The “Graphics” menu enables users to change the size, shape, or color of the synthetic peaks to suit their preferences.

It is important to note that MetaboMiner does Not support spectral processing such as phasing, baseline correction or chemical shift referencing.

There are many other high-quality NMR-processing software products available for this task, including NMRPipe, Felix (Molecular Simulations, Inc., San Diego, CA), VNMR (Varian, Inc., Palo Alto, CA), and XWin-NMR (Bruker Analytik GmbH, Karlsruhe, Germany), to name a few.

These tools should be used prior to loading spectral images into MetaboMiner.

In other words, MetaboMiner is Not a spectral processing tool, but a NMR-based metabolomics tool (as stated above...) that facilitates automatic peak processing, rapid compound identification, and facile spectrum annotation capabilities through an intuitive graphical interface.

System Requirements

Contact manufacturer.

Manufacturer

Manufacturer Web Site MetaboMiner

Price Contact manufacturer.

G6G Abstract Number 20676

G6G Manufacturer Number 104124