Process tab page
The process page is used for specifying all parameters required for data extraction and feature pair bracketing among different LC-HRMS(/MS) files. The page is divided in the following three sections which also correspond to the general workflow of the data processing pipeline:
- Individual files processing
- This part specifies the settings for MS signal and feature pair extraction for individual LC-HRMS(/MS) files. The listed parameter options do not consider comparison of feature pairs among different measurements.
- Multiple files annotation
- With these parameter options, extracted feature pairs from the individual LC-HRMS(/MS) files are matched to their pendants in other data files also specified. This section is not avaiable in the FragExtract module.
- Run tasks
- In this part of the process tab page the data processing may be started. MetExtract supports multi-core systems and will reserve one core per default. Before the actual data processing is started, MetExtract will ask to save the experiment as well as the associated settings. Such a settings file should be saved inside the folder of the the converted mzXML or mzML files or in a parent-folder. The path to the files will be saved relative to this compilation file. During the data processing a dialog with a progress bar will be shown.
Save and load settings
Settings may be loaded or saved (e.g. different settings for Orbitrap and Q-Tof instruments). To save a current setting layout to a file (extension .ini) goto File->Save Settings and select a settings file in the file dialog. To load previously defined settings either goto File->Load Settings and select a settings file or drag'n'drop a respective settings file anywhere onto MetExtract. A settings file may also be loaded via drag'n'drop.
Large experiments
Processing many files at the same time (e.g. more than 50 samples analyzed with an LTQ Orbitrap instrument) may cause a memory error. This is a bug in the current implementation and will be resolved in further version of the software.
If this error occurs, please consider processing the data during the steps Individual files processing and Re-integration during multiple files processing sequentially for some files (e.g. less than 50). The Bracketing of feature pairs step must, however, be performed with al LC-HRMS files together.
Alternatively, the Command line parameters may also be used for automated processing.
Individual files processing
To perform feature extraction for each file separately, the checkbox entitled "Individual files processing" has to be checked. MetExtract will process each file separately and extract corresponding 12C and 13C features and feature pairs. Also, all parameter settings used for this feature pair extraction are located in the "Individual files processing" section of the "Process tab page".
Database file
For each processed LC-HRMS(/MS) file, MetExtract generates a database file named <FileName>.identified.sqlite. This database contains all extracted feature pairs from the respective file and is used for subsequent bracketing of several measurement files as well as results visualisation.
Calculate isotope enrichment
Most elements consist of more than one isotope. For example, carbon consists of the two stable isotopes 12C and 13C. The relative abundances of 12C and 13C in natural environments are 98.93% and 1.1% respectively. The other radioactive isotopes of carbon can be neglected in this respect as their relative abundances are too low. When a molecule consisting of at least one carbon atom is recorded with LC-HRMS, its isotopolog distribution will be visible. The relative abundances of this isotopolog distribution depend on the number of carbon atoms in the respective molecule as well as the abundances of the two stable isotopes. The following formula is used to calculate the theoretical isotopolog distribution for carbon (a is the total number of carbon atoms in the substance, s is the number of 13C substitutions (i.e. M+s) and p is the relative abundance of 12C (e.g. 0.9893 for native molecules):
`f(a,s,p)= (p^(a-s)*(1-p)^s*({::}_s^a))/(p^a)`
This formula can be rewritten and used for determining the enrichment of the labeling-isotope in SIL experiments. The enrichment of 13C is calculated using a highly abundant feature pair in the recorded LC-HRMS(/MS) data. For this feature pair the number of carbon atoms (a) can directly be calculated from the m/z difference between M and M'. The ratio (r) of [M'-s]/[M'] is further required as well as the number of exchanges (s), for which usually 1 is chosen. With the rewritten formula the enrichment of 13C in the labeled isotopologs is determined:
`p= (({::}_s^a)^(1/s))/(({::}_s^a)^(1/s)+r^(1/s))`
Calculating the isotopic enrichment in the labeled metabolite form is exemplified with an uniformly 13C-labeled metabolite shown in figure 5. It has 22 carbon atoms in total and is fully 13C-labeled. Using the ratio M+1/M (rn=0.23) the native 12C enrichment is calculated to be 99.96%, which is in good agreement with the natural 12C enrichment of 98.93%. Consequently, using the ratio of M'-1/M' (rl=0.38) the 13C enrichment in the uniformly labeled metabolite is determined to be 98.3%.
AllExtract - Labelling configuration
The most informative labeling element in SIL-assisted metabolomics approaches is carbon as it is a constituent of virtually any metabolite. Labelling with 13C isotopes will result in the typical SIL associated pattern shown in figure 2. Other elements (e.g. N, S, P) will result in very different isotopolog patterns (see figure 6). For this experimental setup a different extraction method may be used which deduces the feature pairs not based on the isotopic patterns of the labeling element but rather on identical natural carbon isotopic distributions. This option is only available if carbon is not used as the labeling element.
Relative isotope abundances
For several different labeling elements predefined relative isotope abundances are shown when the respective labeling isotope is specified. Ensure that these relative isotope abundances match the experimental setup
- Isotope N
- Predominant isotope in nature of the labeling-element (e.g. 12C, 14N)
- Isotopic N abundance
- Relative abundance of the predominant isotope in natural environments (e.g. 98.93% for 12C)
- Isotope L
- Isotope used for labeling. Must be of the same element as Isotope A
- Isotopic L abundance
- Isotopic enrichment in the labeled metabolites. Must be equal in all labeled metabolites.
- Use Carbon validation
- Select to test either the isotope pattern of the native and labeled metabolite (unchecked state; only recommended for 13C labeling) or if the carbon isotope patterns of the native and labeled metabolite form shall be compared (checked state; recommended for labeling applications that do not use 13C as the main labeling isotope). The second option is only available for non-13C labeling experiments.
Tracer metabolisation experiment - Labelling configuration
To define a tracer under investigation, press the button "Tracer setup" in the labeling-section. A new window will open where the used tracers and labeling associated parameters can be specified. Figure 7 shows the tracer configuration dialog with DON (deoxynivalenol) as a configured tracer.
Define tracer
Each row in the dialog represents one tracer. For each tracer the following parameters have to be specified:
- Tracer
- The name of the used tracer
- Element count
- The number of atoms of the labeling element per formula unit of the tracer substance
- Isotope N
- The major natural isotope of the labeling element (e.g., 12C or 14N)
- Isotope L
- The used labeling-isotope in the labeled tracer form
- Natural abundance isotope N (%)
- The isotopic abundance of Isotope A (e.g. 98.93% for 12C)
- Enriched abundance isotope L (%)
- The isotopic abundance of Isotope B (e.g. 99.5% for 13C).
Note: depends on the used tracer
- Amount natural
- The amount of natural tracer form used
- Amount enriched
- The amount of the labeled tracer form used
- Ratio (N:L)
- Theoretical intensity ratio of the monoisotopic 12C to 13C-labeled ion signals of the tracer. This field cannot be specified. Its value is calculated automatically from the other tracer-specific parameter settings
- Min. ratio (%)
- Minimal allowed ratio relative to 'Ratio (native:enriched)' (i.e. 'Ratio (native:enriched)'*'Min. ratio (%)'). If the specified tracer is an endogenous substance to the biological system, use 0.00000001
- Max. ratio (%)
- Maximum allowed ratio relative to 'Ratio (native:enriched)' (i.e. 'Ratio (native:enriched)'*'Max. ratio (%)'). If the specified tracer is an endogenous substance to the biological system and thus the ratio can be different for different metabolites, use 9999999
- Type
- Currently unused
Pre-selection of heavy isotopes X
To detect putative native and labeled metabolite ions and fragments, MetExtract tests a certain range of possible carbon atoms. This range needs to be specified by the user:
-
Full metabolome labeling
The range can be arbitrary but should not include a too low atom count (e.g. less than 3) or a too high atom count (e.g. more than 65 if the enrichment with 13C is approximately 98.5%)
-
Tracer-fate studies
For tracer metabolisation experiments the maximum number of labeled atoms can roughly be chosen as twice the number of labeled atoms present in the tracer. E.g. Since DON has 15 carbon atoms, the search for metabolisation products can be restricted to a maximum of 30 carbon atoms, which will include metabolites containing two DON tracer units)
Parameter settings for MS signal processing
The following parameter settings are located in the section MZ picking. These settings are related to the mass accuracy and expected isotopolog ratio accuracy of the used mass spectrometer.
The time interval of the chromatogram during which MetExtract inspects mass spectra and EIC chromatograms, is specified in the input fields "Start (min)" and "End (min)". Only MS spectra and EIC peaks between these time points of the LC-HRMS run are considered during data processing.
An optional intensity threshold can be specified in the field "Intensity threshold". All centroided MS signals above this intensity threshold are assumed to be putative MS signals of monoisotopic ions with natural isotopic composition or putative MS signals of a labeled ion. Required isotopolog signals may also be lower than this threshold to be recognised by MetExtract.
The input field "Max. number of charges" specifies the max. allowed charge state of an ion to be considered for data processing by MetExtract. An ion's charge state is determined using its carbon isotope pattern.
The field "Max. mass deviation (+/- ppm)" specifies the maximum tolerated mass deviation of M' as well as M+1 and M'-1 in ppm, relative to the calculated m/z value of the observed MS signal M. E.g. in case of 13C labeling M' = M + n * 1.00335 / z.
The parameters "Isotopologs N" and "Isotopologs L" specify how many isotopologs of the natural and the labeled isotopologs must be present and are verified in the respective patterns (e.g. a value of 2 for "Isotopologs N" means that the isotopologs M and M+1 must be present for the native ion form). If "1" is used, MetExtract will only check for and use the two corresponding MS signals M and M' for confirmation.
The parameters "Intensity abundance error" for "Non-labeled ion (±)" and "Labelled ion (±)" specify the maximium relative deviation between the expected and the observed ion intensity ratios of M+y relative to M and M`-x relative to M'. A parameter setting of e.g. 0.2 corresponds to a maximum tolerated relative isotopic abundance error of 20% compared to the theoretically expected one. In case of a tracer-fate experiment, the ratios of M'-y to M' are corrected by the factor M'+y to M' to account for possible moieties conjugated to the studied tracer substance.
Note: Optimization step
MetExtract II tries to find valid pairs of M and M' signals. However, it can also happen that M+1, M+2, M'-2, M'-1 signals are incorrectly paired leading to false-positives. To reduce the number of such pairings, MetExtract II only accepts MS signals, if no M-1 or M'+1 signal is detected in the same scan (the latter verification is suspended when TracExtract is used).
Low abundant isotopolog signals
The isotope patterns of native and partly or uniformly labeled metabolites have characteristic abundances relative to their monoisotopic or consistently labeled isotopologs. Depending on the abundance of the metabolite in a particular sample, these signals of the respective isotopologs may be present in the LC-HRMS data. But in some cases (e.g. low number of carbon atoms and / or low abundance of the monoisotopic or labeled metabolite) these signals can hardly be observed and thus incorrectly not detected using the strict isotopolog filtering system of MetExtract II (parameters "Isotopologs N" and "Isotopologs L").
To circumvent this problem, the option "Consider isotopolog abundance" may be used. When checked, it will instruct the isotope matching algorithm of MetExtract II to consider the observed signal intensity of putative isotopologs. Only when the calculated signal abundance of an isotopolog is above the specified intensity threshold (parameter "Isotopologs threshold (≥)") the presence of this isotopolog is verified. If the calculated abundance is lower than the set intensity threshold, the respective isotopolog signal is not required to be present in the LC-HRMS data.
With this option active, the metabolite detection algorithm of MetExtract II is more sensitive. At the same time the number of incorrectly matched signals is also increased, which leads to a higher number of false-positives. Thus, this option should be used with caution and the generated results should be carefully reviewed before further analysis of the dataset. To counteract these problems, the following settings may be adapted:
- Min. spectra (≥)
- This option should be set to a higher value to account for the low abundant peak flanks that will also be matched with this new setting
- Min. corr (≥)
- The peak shape comparison should also be set to a higher value (e.g. 0.7 or 0.8 depending on the LC method and the expected chromatographic peaks)
- Intensity abundance error
of Non-labeled ion (±)
- If no isotopolog peaks M'+1 are detected in a TracExtract experiment, it is assumed that no exists. Thus, the ratio check of M+1 relative to M is incorrect if there are additionally, non-labeled atoms of the labeling element. Thus, this parameter should be increased to e.g. 0.75 or even higher values depending on the expected number of atoms
- Number of replicates
- When using this parameter, especially in combination with a relatvie low intensity threshold, the number a feature pair needs to be detected in the replicates of an experimental group should be increased to avoid too many false-positives
Additionally, in many LC-HRMS systems the chromatographic peak shapes of low abundant ions tend to show increased noise. This complicates the automated feature pair grouping mechanisms of MetExtract II resulting in not correctly convoluted feature groups.
See figure 9 for an illustration and table 1 for a comparison of the results for the provided datasets DON_in_Wheat.grp and DON_in_Wheat_LAC.grp
| Processing settings |
Total metabolites |
Unique metabolites |
False-positivly detected feature pairs |
Processing time |
| DON_in_Wheat.grp / 10_CM_DON.mzXML |
8 |
0 |
0 |
2.1 min |
| DON_in_Wheat_LAC.grp / 10_CM_DON.mzXML |
12 |
4 |
20 |
8.2 min |
Table 1: Comparison of the detected number of metabolites detected in the datasets DON_in_Wheat.grp and DON_in_Wheat_LAC.grp
Grouping of detected MS signals
Detected MS signal pairs consiting of M and M' signals in individual scans are clustered using hierarchical clustering with euclidean distance and average linkage. The parameters found in the section "MZ clustering" are used for this step.
The parameter "Clustering ppm (±)" specifies the maximum ppm deviation present within a MS signal cluster. All cluster exceeding this threshold are separated in two or more subclusters fulfilling this parameter.
The field "Min. spectra (≥)" is used to discard those subclusters that consist of less MS signals than specified with this paramter. This is especially helpful to discard low abundant Fourier transform artefacts.
Chromatographic peak picking
The next step in the AllExtract and TracExtract modules detect and separate different chromatographic peaks of metabolites having the same m/z value but a different retention time. An example is illustrated in figure 10. The following steps are carried out for each remaining subcluster of MS signal from the previous processing step:
- EIC chromatograms of the detected MS signal pairs (monoisotopic non- and uniformly labeled ions) are extracted from the data in a predefined m/z interval around the average m/z values using the processing parameter "EIC ppm".
- The MassSpecWavelet algorithm described by Du et al. (2006) is used for separating different chromatographic peaks in a single EIC.
The parameter "EIC width (±)" defined the ppm width that is used for extracting the EICs of M and M' for each remaining subcluster from the previous data processing steps.
An optional EIC smoothing step may be used. This can be activated with the field "Smoothing window". Several options are available. If used, the smoothing window size must also be specified with the parameter "Window size". Note: the option "Window size" is only available when smoothing is used.
The minimum and maximum width of chromatographic peaks may be specified with the parameters "Min. scale" and "Max. scale".
The maxmimul allowed peak center error is specified with the field "Center error" in the "Peak matching" category". Like the peak scale parameters this value also uses number of scand between two peak centers.
The minimum Pearson correlation between two chromatographic peaks of a native and labeled metabolite ion is specified with the parameter "Min. corr (≥)".
Each such detected and verified feature pair is considered a metabolite ion derived from a native and a labeled ion of the same metabolite.
Post processing of detected feature pairs
After feature pairs have been detected they are further annotated with putative hetero atoms and convoluted into feature groups each representing a metabolite.
Hetero atom annotation
After feature pairs have been detected, MetExtract searches for isotopolog mass peaks originating from other elements than the one used for labeling. Such hetero atoms must consist of at least two main isotopes. Chloride, for example, consists of the two main-isotopes 35Cl and 37Cl. 35Cl, with an relative abundance of 75%, is the more abundant isotope expected to be predominantly incorporated into molecules. However, since the less frequent 37Cl has a quite high abundance of 24% it is recorded in most mass spectra for substances with chloride atoms.
Isotopes having a positive m/z offset compared to the most abundant isotope of the respective element are searched at M’+isotope_m/z, while isotopes with a negative m/z offset (e.g. 54Fe) the search for respective mass peaks is performed at M-isotope_m/z
Several hetero isotopes show similar m/z value offsets and similar relative isotopolog abundances. Depending on the HRMS device, these may or may not be separated. MetExtract will report all possible hetero isotopologs.
The maximum allowed ratio error for a hetero atom isotopolog is specified with the parameter "Intensity error (±)"
Background noise signals may also be incorrectly interpreted as MS signals originating from hetero atom isotopologs, since isotopologs are generally very low abundant MS signals (e.g. 34S). Consequently, they need to be present in several consecutive MS scans to be correctly found. The parameter "Min. scans" defines the number of times a hetero atom isotopolog must be detected to be reported.
Hetero atoms are configured by clicking on the "Hetero atoms configuration" button. The following information is required for every possible hetero isotope:
- Isotope
- Identifier of the isotope (e.g. 34S)
- Δ m/z
- The m/z ratio offset compared to the elements most abundant isotope with a charge of 1
- Rel. ab [%]
- The abundance of the isotope relative to the element's most abundant isotope. (e.g. The natural abundances of chloride are 76% for 35Cl and 24% for 37Cl. Therefore, their ratio is 24 % / 76 % = 0.32 = 32 %)
- Min. atoms
- The minimal number of hetero isotopes to be searched for
- Max. atoms
- The maximum number of hetero isotopes to be searched for
Feature convolution
Ionisation by ES may give rise to several ion species for the same metabolite (e.g. adducts, in-source fragments or dimers). To convolute all ion species of a particular metabolite into a feature group, the Pearson correlation coefficient is used for finding feature pairs with highly similar retention time and chromatographic peak shapes. The parameter "Min. corr. (±)" in the section "Feature convolution" specifies the minimun Pearson correlation the two chromatographic peaks of the monoisotopcic isotopolog of two feature pairs must have in order to be convoluted into one feature group.
Close co-eluting metabolites may be incorrectly merged into a feature group consiting of different metabolites (see figure 11a and 11b). To split a feature group consisting of incorrectly convoluted feature pairs, HCA is used. Only such subclusters, which show a large number of highly correlated feature pairs are not further split. In the example in figure 11b, 5 feature groups are incorrectly convoluted. Using the parameter 'Min. connections' (relative amount of highly connected feature pairs in a feature group) this feature group was split into 5 feature groups. As shown in figure 11c, the feature pairs in each of these new feature groups show better co-elution.
Subsequently, each feature group is annotated with commonly observed adducts and in-source fragments. This is done by comparing the m/z value of M as well as the determined number of labeled atoms between all feature pairs of a feature group. If the difference between two feature pairs cannot be explained with two adducts observed in the respective ionization modes, putative neutral losses of in-source fragments are calculated using the Seven Golden Rules algoritm (Kind and Fiehn 2007). Adducts and elements that shall be used for the generation of neutral losses can be specified in a dialog that is accessed using the button "Relationship configuration".
- Adducts used for feature pair annotations
- Adduct
- The name of the adduct (e.g. M+Na, M+H…)
- MZ offset
- The m/z shift obtained by transforming the neutral molecule M to the respective ion species. E.g. for a protonated ion ([M+H]+) the m/z offset is 1.007278
- Polarity
- Specifies the ionization polarity in which the adduct is observed. Only "+" and "-" for the positive and negative mode are allowed.
- Elements for neutral loss calculations
- Element
- The chemical symbol of the element
- Weight
- The exact mass of the elements most abundant isotope
- Valence electrons
- The number of valence electrons of the respective element
- Min count
- The minimum number of atoms at which this element must occur per molecular formula unit
- Max count
- The maximum number of atoms at which this element can occur per molecular formula unit. Note: too high numbers here may increase overall calculation time
Conjugated moieties in tracer experiments
Biotransformation products of a tracer may have conjugated moieties that have additional atoms of the labeling element. A neutral loss may not reduce the number of labeled atoms for two features but the neutral loss may still contain the heavier isotope of the labeling element. Thus, the number of labeled atoms is used only in AllExtract for the generation of putative neutral losses. In TracExtract the number of labeled atoms is ignored.
Save results
Results of the individual files/measurements may be saved in various formats. Each output file is saved as <FileName>.<extension>
TSV list
The detected feature pairs and convoluted feature groups are stored in a matrix format. Each row represents a feature pair having the following information:
- Num
- ID of the feature pair. Note: It's named Num to prevent Excel from opening it in the wrong format.
- MZ
- Average m/z value of M
- L_MZ
- Average m/z value of M'
- D_MZ
- Mz difference between M and M'
- RT
- Retention time in minutes of the detected chromatographic peak of M
- Xn
- Number of atoms of the labeling element
- Charge
- Charge state of the ion
- ScanEvent
- The scanevent in which this feature pair was detected
- Ionisation_Mode
- The polarity in which this feature pair was detected
- Tracer
- The configured tracer, for which this feature pair was detected. In an AllExtract experiment FML (full metabolome labeling) will be written
- Area_N
- The peak area of the chromatographic peak of M determined by the peak picking
- Area_L
- The peak area of the chromatographic peak of M' determined by the peak picking
- Fold
- The ratio of Area_N / Area_L
- PeakRatio
- The scan-wise ratio of MS signals within the chromatographic peaks of M and M'
- PeakRatioMp1
- The scan-wise ratio of MS signals within the chromatographic peaks of M+1 and M
- PeakRatioMPm1
- The scan-wise ratio of MS signals within the chromatographic peaks of M'-1 and M'
- LeftBorder_N
- The left border of the chromatographic peak of M (in scans)
- RightBorder_N
- The right border of the chromatographic peak of M (in scans)
- LeftBorder_L
- The left border of the chromatographic peak of M' (in scans)
- RightBorder_L
- The right border of the chromatographic peak of M' (in scans)
- Group_ID
- Feature groups have been numbered. This column specifies the number of the feature group, to which this feature pair has been assigned to. All feature pairs with the same number in this column belong to the same feature group
- Corr
- The Pearson correlation of the chromatographic peaks M and M'
- Adducts
- Ion species (e.g. adducts), which have been annotated for this feature pair. (May be several if not unique)
- Hetero_Elements
- This field includes isotopologs of elements not used for the labeling (e.g. 24S). Each annotation is given with the number of atoms of the respective lement as well as the observed and expected ratios
PDF
For each input file a PDF file may be created which contains the same information as the TSV output file. Each detected feature pair is additionally visualised (mass spectra, EICs of M and M') on a PDF page. Moreover, all chromatographic peaks of M for all convoluted feature groups are illustrated.
mzXML file
This option allows saving the detected feature pairs in a new mzXML file. This mzXML file will only contain MS signals of feature pairs detected with MetExtract. All other signals that do not originate from native and labeled substances and their ions are not saved. Depending on the selected options (M, M+1, M'-1, M'), isotopologs of the detected feature pairs may also be included.
Saving the results to a new mzXML file may only work for certain input data. If this option does not work with a certain instrument, please contact the authors. Currently, this export functionality has only been tested the following instrument(s):
- Orbitrap XL from Thermo Fisher Scientific™
Multiple files annotation Only avaialbe in the modules AllExtract and TracExtract
The section "Multiple files annotation" provides parameters for matching extracted feature pairs in all processed LC-HRMS data files.
Bracket results
The matching is performed using the m/z value and retention time of M as well as the determined number of labeled atoms. The field "Max. m/z width (± ppm)" defines the maximum allowed difference in m/z of M for two feature pairs detected in different LC-HRMS files to be matched. The matching is performed using HCA. Those subclusters, which have less ppm difference between their highest and lowest m/z value of M are kept. It is recommended to set this parameter to a multiple (2-3) of the parameter "Mass deviation (± ppm)". Subsequently, all chromatographic peaks detected within this m/z cluster are again clustered using their retention time. All clusters of less than "Time window (min)" belong to the same chromatographic peak.
Align chromatograms
For slight retention time shift, an optional alignment step may be performed. This alignment is realised with the R-package PTW (Polynomial time warping). By default, no alignment is performed.
Re-integrate missing feature pairs
To fill up the created data matrix, a semi-targeted re-integration step is performed after bracketing of feature pairs. This step searches for chromatographic peaks that were not detected in some of the analyzed LC-HRMS files but successfully detected in other. Most ofter, low abundant features are missed due to the even lower abundance of the required isotopolog MS signals.
Output
The output of feature pair bracketing is a two-dimensional data matrix consisting of all detected feature pairs and their abundances in the different LC-HRMS files. Moreover, each detected feature pair is annotated with:
-
For each detected feature pair the following information is saved independently of the file it was detected in:
- Num
- ID of the feature pair. Note: It's named Num to prevent Excel from opening it in the wrong format.
- MZ
- Average m/z value of M in all detected LC-HRMS files
- L_MZ
- Average m/z value of M' in all detected LC-HRMS files
- D_MZ
- Mz difference between M and M'
- RT
- Retention time in minutes of the detected chromatographic peak of M
- Xn
- Number of atoms of the labeling element
- Charge
- Charge state of the ion
- ScanEvent
- The scanevent in which this feature pair was detected
- Ionisation_Mode
- The polarity in which this feature pair was detected
- Tracer
- The configured tracer, for which this feature pair was detected. In an AllExtract experiment FML (full metabolome labeling) will be stated
- OGroup
- Feature groups have been numbered. This column specifies the number of the feature group, to which this feature pair has been assigned to. All feature pairs with the same number in this column belong to the same feature group
- Ion
- Ion species (e.g. adducts), which have been annotated for this feature pair. (May be several if not unique)
- Loss
- Neutral losses, which have been generated if no adduct was annotated for the respective feature pair. (May be several to various other feature pairs in the group)
- M
- The postulated exact mass of the neutral metabolite as determined by the annotated adduct(s). (May be several if not unique)
- For each analyzed LC-HRMS file, the following columns are created:
- <FileName>_Area_N
- The peak area of the extracted ion chromatographic peak of M determined by the peak picking in the respective LC-HRMS file
- <FileName>_Area_L
- The peak area of the extracted ion chromatographic peak of M' determined by the peak picking in the respective LC-HRMS file
- <FileName>_FID
- The feature id (column "Num") of the feature pair in the respective LC-HRMS file
- <FileName>_GroupID
- The number of the feature group, to which this feature pair has been assigned to. Note: It is only unique in the respective LC-HRMS file and cannot be referenced to other LC-HRMS files.
- Additionally, for each analyzed LC-HRMS file, the following column is created:
- <FileName>_fold
- The ratio of M:M' of the feature pair in the respective LC-HRMS file
- For each defined experimental group, the following columns are created:
- <GroupName>_Stat_fold
- The count, in how many data files extracted ion chromatographic peaks for both M and M' have been found (after re-integration of initially missed peaks)
Note
The IDs of feature pairs as well as of feature groups (<FileName>) are unique within a certain LC-HRMS analysis. They are not matched across files and thus cannot be compared
Start data processing
To start the data processing, press the "Start" button located on the right bottom of the calulcate page. MetExtract will ask to save the current file/group compilation as well as the specified settings for data processing. Once started, a progress dialog will display the status and the current operation per file.
Parallel processing
Since many operations of MetExtract can be parallelized (e.g. feature pair detection in individual LC-HRMS(/MS) data files), multi core systems and paralled computation are supported. The default number of cores MetExtract will use is one less than available on the executing PC to not block the computer for office work. To use all available cores uncheck the box "Keep one core unused"
Processing dialog
During the processing of the specified LC-HRMS files, a dialog will be shown. It will show the overall process (first label and progress bar) of all files. Then, for each CPU core, a separate label and progress bar will be shown, which illustrate the status of the currently processed files. At the botton of the dialog, a tabular overview will be displayed showing which files have already been processed successfully or are currently being processed. Figure 12 shows an example of this dialog with 4 experimental groups and 7 parallel file processings
Results Tab Page
Results of data processing can be visualized in the third tab named "Sample results". Select a processed file from the drop-down list on the left top of the page. The results of the selected file will be presented in the list underneath. Once a result is clicked on, the headers of the list will change accordingly. This list is divided into the following five categories:
- MZs
- All detected MS signal pairs in the individual scans are listed
- MZ bins
- The remaining subclusters used for chromatographic peak picking are listed
- Feature pairs
- All detected and verified feature pairs are listed
- Feature groups
- All convoluted feature groups are listed
- Parameters
- The used data processing parameters are listed
External viewer
To view the raw-mzXML file in an external viewer use the button "open mzXML file externally". It will be opened with the default application for mzXML files.
MZs
The category "MZs" shows all MS signals from all scans of the original data file that showed the mirror symmetric isotopic patterns and fulfilled all parameter settings for this processing step. The only plot available for this result category is the main plot. It shows a scatter plot of all native MS singals detected. If a specific ion signal is selected, it will be coloured.
MZ bins
The category "MZ bins" shows all clusters remaining after MS signal clustering. Only the main plot is available for a graphical illustration of these bins. A selected bin will be colour highlighted among all extracted MZ bins.
Feature pairs
The category "Feature pairs" shows detected feature pairs. For a feature pair two plots are available. The main plot shows the chromatographic peaks of the natural and labeled ion forms. The intensities of the labeled metabolite ions have negative intensities while the native metabolite ions are depicted with positive intensities. The second plot for each feature is its mass spectrum (see figure 13).
Feature groups
The category "Feature groups" shows convoluted feature groups. For each feature group three plots are available. The main plot shows an overlay of all feature pairs convoluted into the selected feature group. Intensities of differently abundant feature pairs are noarmalised to 1 if the field Normalise is selected. The second plot shows an illustration the correlation and similarity of all feature pairs in this feature group. The third plot depicts an annotated MS scan for this feature group (see figure 14)
Filters
MetExtract offers the possibility to filter all results for certain criteria. The filter can be entered in the text field "Filter" in the "Sample results" tab. It is applied to each result category. The entered text is first split into chunks using the space character. Each i-th chunk is a filter for the i-th column in the results tab.
EXAMPLE 1: If the user wants to search for feature pairs with a m/z value of 297.1333 for M, the text "297.1333" is entered in the field "Filter". MetExtract will then only show results having the text 297.1333 in the first column. If the search shall be further restricted to 15 carbon atoms, the user must enter "297.1333 15". This will also search in the second column for the text 15 and only show those feature pairs, which have 15 or 115 or 1500 carbon atoms.
EXAMPLE 2: If the user wants to search for 297.1333 +- 1 amu the filter must be set to "297.1333+-1". This will instruct MetExtract to search for m/z values of M within the range of 296.1333-298.1333 amu. It is also possible to search within a specific ppm window. The filter must then be set to "297.1333+-5ppm".
Results view - FragExtract
FragExtract shows analyzed LC-HRMS/MS targets and detected fragment peaks in form of a hierarchical tree. The first level of the tree are the LC-HRMS/MS targets (for example CouAgm in the green box in figure 15). It includes the name (column "Target name"), m/z value of the native precursor ion (column "MZ"), its total number of labeled carbon atoms (column "Cn"), either user defined or generated sum formulas of the parent (column "Sum formula"), the charge of the precursor ion (column "Charge"), the scan number of the MS/MS scan used for analyis of the native precursor ion (column "Native scan num"), the scan number of the MS/MS scan used for analysis of the U-13C-labeled precursor ion (column "Labelled scan num"), The scan event of the LC-HRMS full scan(s), and the scan events of the LC-HRMS/MS scans of the native and the U-13C-labeled precursor ions.
The next level lists detected MS/MS fragment peaks found for both the native and the U-13C-labeled precursor ions (blue box in figure 15). Only those peaks present as both forms with approximately the same relative intensity and a defined Δm/z value corresponding to the total number of carbon atoms in the respective fragment ion, are shown. Each such fragment peak is annotated with its assigned peak number (column "Target name", form "Peak XX"), the m/z value of the fragment peak derived from the monoisotopic precursor ion (column "MZ"), the determined number of carbon atoms for this fragment peak (column "Cn") followed by the calculated m/z value of the U-13C-labeled fragment ion pendant, a generated sum formula for the fragment peak (column "Sum formula"), the adduct of the generated sum formula (column "Charge"), the neutral loss of the fragment in respect to the used precursor ion (column "Native scan num"), and the relative intensity of the fragment peak in the LC-HRMS/MS scan of the native precursor ion (column "Labelled scan num"). If several putative sum formulas could be calculated for one fragment ion, a start ("*") is shown instead of the sum formula (orange box in figure 15, first row). All generated sum formulas are then shown one level beneath the annotated fragment peak (redundant information is not shown).
Results of FragExtract processing can easily be exported in tab-separated table format. To copy the results of all MS/MS targets in a LC-HRMS/MS file, right-click on the top-level item named "MSMS targets" and select "Copy". To copy the results of a single target, right-click on the target name and select "Copy". The copied information can then be transfered to e.g. Excel via the paste operation (Strg+v). The table contains information about the precursor-ions and the extracted MS/MS peaks as well as generated sum formulas.
