MetaboHunter is a web server application for semi-automatic assignment of 1D NMR spectra of metabolites.
MetaboHunter provides tools for metabolite identification based on spectra or peak lists with three different
search methods and with possibility for peak drift in a user defined spectral range. The assignment is
performed by comparison with two major publically available databases (HMDB, MMCD) of NMR metabolite standard measurements.
In this work we present the Web tool with detailed testing of the methods provided in the Supplementary data
MetaboHunter makes use of
MMCD NMR spectra. All spectra have been manually curated.
- Input type: The input data for MetaboHunter must contain a list of peaks or a complete 1H NMR spectrum.
If a complete spectrum is provided as input, noise filtering and peak detection is applied.
- Input file: The input file must have a list of peaks or a full spectrum.
Data must be organized in two columns, such that each line must have only two values.
The first value must be the ppm and the second value must be the height of the signal.
A sample file is provided here.
- Peaks list: The input could be also pasted as a list of peaks or a spectrum.
Data must be organized as described above.
- Type of metabolite: When available, the metabolite type can be used as a screening criterion.
The metabolite data originated from HMDB contains metabolites that can fall in one of the following
classes: drug, food additive, mammalian, microbial, plant, synthetic/industrial chemical.
- Sample pH: When available, the sample pH can be used as a screening criterion.
The metabolite data originated from HMDB and MMCD samples was measured on a broad range of pH values, from 2 to 10.
- Solvent: When available, the solvent can be used as a screening criterion.
The following solvents have been used to measure the metabolite spectra in HMDB and MMCD: water, CDCl3, CD3OD and 5% DMSO .
- NMR frequency: When available, the NMR frequency can be used as a screening criterion.
The metabolite data originated from HMDB and MMCD was measured using the following NMR frequencies: 400 MHz, 500 MHz and 600 MHz.
- Matching method: MetaboHunter allows the user to select their preferred method for metabolite identification
when a query spectrum or list of peaks is provided. The available methods are:
- MH1: Highest number of matched peaks,
- MH2: Highest number of matched peaks with shift tolerance,
- MH3: Greedy selection of metabolites with disjoint peaks and shift tolerance.
- Noise threshold: The noise threshold must be a number above which all peaks in the input file will be considered as
potential matches for metabolite spectra.
- Confidence threshold (percentage): The confidence threshold must be a number greater than 0 that represents the cut-off score
or percentage for matched spectra. The scoring function for all metabolite identification methods integrated in MetaboHunter
is the ration between the number of matching peaks and one plus the total number of peaks for any given metabolite.
- Shift tolerance: The shift tolerance must be a real number greater or equal to zero that represents the amount
of variability (measured in ppm) allowed for each metabolite peak in the sample to match a metabolite peak in the database.
The lists of curated peaks used by MetaboHunter can be downloaded here: HMDB peaks
, MMCD peaks
Please cite our paper if you use Metabohunter in your work.
Tulpan, D., Léger, S., Belliveau, L., Culf, A., Cuperlovic-Culf, M. (2011). MetaboHunter: semi-automatic identification of 1H-NMR metabolite spectra in complex mixtures. BMC Bioinformatics 2011, 12:400.