5 Untargeted sample analysis¶

General¶

emzed3 heavily features untargeted analysis. To do so emzed3 provides tools from OpenMS and MZmine2. Whereas open ms features are integrated in the emzed3 core mzmine2 is an extension that has to be installed first. To get familiar with those tools we will exercise the basic workflow depicted above which comprises the core peak processing steps detection, alignment, and grouping. Moreover, we will perform some basic sample comparison applying data base matching with emzed integrated pubchem library.

Feature detection¶

Let’s start with a more general look at LC-MS data structure. LC-MS level 1 data acquisition can be regarded as a 2 dimensional separation procedure. In the first dimension, LC separates compounds by physico-chemical interactions over time (retention time). In the second dimension, compound ions are separated by their mass to charge ratios m/z whereby the continuous LC output is scanned with an MS instrument specific frequency. Hence, we can define LC-MS peaks as sequences of spectra with m/z specific intensity series of with shapes typical for applied LC separation. Accordingly, we can subdivide the task of peak detection algorithm into two principal steps:

1. EIC detection: Search the peak map for all series of consecutive m/z peaks with the same m/z values (EIC).

2. Peak detection: Find all chromatographic (LC) peaks within each EIC.

• To find an EIC we have to define the the meaning of m/z peaks (signal), what means a series of consecutive m/z peaks, and what means m/z peaks with same value. To define a signal we have to distinguish between signal (S) and noise (N). Remember, it is commonn practice to the define presence of a signal by a \(S/N >=3\) and a quantifiable signal by \(S/N >=10\). The number of (consecutive) m/z peaks or the time range considered to define a m/z trace directly depends of the applied LC-method and on the scan speed of the instrument. Whereas HPLC columns produce peaks with peak widths in the range of 20-30s, UPLC peaks are up to 10x narrower. Independent on the applied column type, compound peaks of interest can strongly differ from the typical shape and hence the range must be chosen with care. Finally, the variation of measured m/z values depends on the mass resolution \(R = mz/\Delta mz\) (measured at full width half maximum, FWHM) and is not only instrument dependent but also depends on the m/z itself for most common MS instruments (TOF, Orbitrap). The width of a m/z peak can be easily seen when data were acquired in the profile mode. However, most algorithms work with centroided data, where each spectral peak is represented by its centroided value. In that case the R value of the peak or the spectrum is required. A more practical approach is to determine the observed m/z width (max(m/z) - min(m/z)) directly from the raw data: Figure 3: Measuring m/z width with PeakMap explorer using the plot of the summed spectra. For the given EIC, the m/z difference of the minimal and maximal measured m/z value is dmz = 0.000946.

  Figure 3 shows a typical Orbitrap m/z trace. In this example the absolute trace width is around 1 mmU and the relative about 4 ppm (m/z = 238.154). If the algorithm requires relative values the acquisition m/z range is crucial to provide reasonable values (1 mmu at m/z 75.0 corresponds to ~13 ppm).

• Once EICs are available, LC-peaks can be defined. Similar to mass trace detection, the values start, apex and end of LC peaks must be determined. Since those parameters are much more algorithm dependent, we will discuss peak detection in more details directly with provided peak detectors.

We will introduce untargeted feature detection using LC-MS data acquired with an Orbitrap MS instrument and hence, some aspects will be instrument specific.

[2]:

kwargs = dict(
    common_chrom_peak_snr=10.0,
    common_chrom_fwhm=3.0,
    mtd_noise_threshold_int=7000.0,
    mtd_mass_error_ppm=20.0,
    mtd_reestimate_mt_sd="true",
    mtd_trace_termination_criterion="outlier",
    mtd_trace_termination_outliers=4,
    mtd_min_sample_rate=0.5,
    mtd_min_trace_length=5.0,
    mtd_max_trace_length=-1.0,
    epdet_width_filtering="auto",
    epdet_masstrace_snr_filtering="false",
    ffm_local_rt_range=2.0,
    ffm_local_mz_range=5.0,
    ffm_charge_lower_bound=0,
    ffm_charge_upper_bound=3,
    ffm_report_summed_ints="false",
    ffm_isotope_filtering_model="none",
    ffm_use_smoothed_intensities="false",
)

[3]:

t_ara = emzed.run_feature_finder_metabo(
    peakmap_arabidopsis, verbose=False, **kwargs
)
t_caulo = emzed.run_feature_finder_metabo(
    peakmap_caulobacter, verbose=False, **kwargs
)

[4]:

t_caulo.summary()

[4]:

[5]:

t_caulo.sort_by("intensity", ascending=False)[:10]

[5]:

[6]:

t_comp = t_ara.left_join(
    t_caulo,
    t_ara.mz.approx_equal(t_caulo.mz, 0.003, 0)
    & t_ara.rt.approx_equal(t_caulo.rt, 20.0, 0),
)

[7]:

t_comp.summary()

[7]:

[8]:

t_comp.sort_by("intensity")[:10]

[8]:

MZmine2¶

Similar to run_feature_finder_metabo, MzMine2 pick_peaks features peak detection in two main processing steps ion chromatogram extraction and peak detection where different peak detection algorithms can be applied:

# mzmine2 is an extension we find in emzed name space under ext. # We create a mopre dicrect access mzmine = emzed.ext.mzmine2 mzmine.pick_peaks(peakmap, adap_chromatogram_builder, peak_resolver, rsp_parameters=None, verbose=False)

The peak_picker function arguments are

• peakmap: PeakMap object

• adap_chromatogram_builder: ADAPChromatogramBuilder object

• peak_resolver: PeakDetectionBuilder object

• rsp_parameters: RemoveShoulderPeaksParameters object, optional. If provided also shoulder peaks will be removed.

• verbose: print more output from mzmine when verbose is True .

In contrast to openMS run_feature_finder_metabo , pick_peaks does not hand over directly feature detection parameters as function arguments, but objects defining processing step and each processing step object is configured separately. The optional rsp_parameters allows removing shoulder (satellite) peaks. Those are artificial peaks which can by generated by Orbitrap Instruments.

MzMine2 provides ADAP chromatogram extraction and peak detection algorithms [Meyers et al.] (https://pubs.acs.org/doi/abs/10.1021/acs.analchem.7b00947) which we will discuss further. Let us start with ADAPChromatogramBuilder description by Meyers et al.:

1. Take all the data points in a datafile, sort them by their intensities, and remove those points (mostly noise) below a certain intensity threshold.

2. Starting with the most intense data point, the first EIC is created.

3. For this EIC, establish an immutable m/z range that is the data point’s where is specified by the user.

4. The next data point, which will be the next most intense, is added to an existing EIC if its m/z value falls within its m/z range.

5. If the next data point does not fall within an EICs m/z range, a new EIC is created. New EICs are only created if the point meets the minimum start intensity requirement set by the user.

6. An m/z range for a new EIC is created the same way as in step 3 except the boundaries will be adjusted to avoid overlapping with pre-existing EICs. As an example consider an existing EIC with m/z range (100.000,100.020) for . If the new EIC is initialized with a data point having an m/z value of 100.025, then this new EIC will have am/z range set to (100.020, 100.035) rather than (100.015, 100.035).

7. Repeat steps 4−6 until all the data has been processed.

8. Finally, a post processing step is implemented. Only EICs with a user defined number of continuous points above a user defined intensity threshold are kept.

``ADAPChromatogramBuilder`` has 4 tuning parameters:

• minimum_scan_span: Minimum number of scans over which some peak in the chromatogram must have continuous points above the noise level to be recognized as a chromatogram. The optimal value depends on the chromatography system setup. The best way to set this parameter is by studying the raw data and determining what is the typical time span of chromatographic peaks.

• mz_tolerance: Maximum allowed difference between two m/z values to be considered same. The value is specified both as absolute tolerance (in m/z) and relative tolerance (in ppm). The tolerance range is calculated using maximum of the absolute and relative tolerances.

• start_intensity: Points below this intensity will not be considered in starting a new chromatogram.

• intensity_thresh2: This parameter is the intensity value for which intensities greater than this value can contribute to the minimum_scan_span count

Note, with start_intensity and intensity_thresh2 2 different threshold parameters are provided. It is useful to choose a higher start_intensity than intensity_thresh2 values since it allows avoiding chromatogram splits along the baseline due to noise. The parameter depends a lot on signal quality i.e. electro spray stability. openMS ff_metabo handles the same issue by allowing a user defined number of outliers (values below the intensity threshold) or alternatively, by defining a signal frequency (not shown). All those values are instrument and sample dependent and require specific tuning. The art is to find the right balance between peak number and quality and requires some time investment. Good practice is to check for the detection of compounds known to be present in the samples at different abundances (i.e. due targeted analysis or spiked compounds).

Also note, the parameter mz_tolerance is defined by a tuple (absolute_tolerance, relative_tolerance in ppm) and the applied tolerance is defined as \(max(atol, rtol)\). Some emzed expressions have a very similar syntax i.e. column.approx_equal(other_column, atol, rtol) In contrast, all core emzed expressions use the common additive tolerance definition :math:`tol = atol + rtol *value`. To avoid unexpected results, we recommend to set one of the two tuple values to 0.

When comparing openMS run_feature_finder_metabo and MZmine2 peak_picker performance we should configure the same or similar parameters with similar values. Let’s configure the ADAPChromatogramBuilder:

[9]:

mzmine = emzed.ext.mzmine2
chrom = mzmine.ADAPChromatogramBuilder()
chrom.intensity_thresh2 = 7e3
chrom.minimum_scan_span = 5
chrom.mz_tolerance = (0.008, 0)
chrom.start_intensity = 1e4

Next, we will define peak detection process. MzmMine 2 provides 5 different peak detectors, each with different strengths and weakness.

• ADAPDetector

• BaselinePeakDetector

• MinimumSearchPeakDetector

• NoiseAmplitudePeakDetector

• SavitzkyGolayPeakDetector

In the following we will focus on the ``ADAPDectector`` since it performs best in terms of data quality.

ADAP detects peaks using continuous wavelet transformation (CWT). Such transformation simplifies peak recognition since peaks can be detected by varying wavelet scale (the principle is nicely demonstrated here. In principle, LC-MS peaks can now be detected by following along their ridges or ridgeline (max values) as function of the scaling parameter see also Meyers et al.. Configuration requires 6 parameters:

• peak_duration: Range of acceptable peak lengths. Tuple (min, max) in seconds.

• rt_for_cwt_scales_duration: Upper and lower bounds of retention times to be used for setting the wavelet scales. Choose a range that is similar to the range of peak widths (FWHM) in seconds expected to be found in the data.

• sn_estimators: User can choose between two signal to noise estimator objects:

  1. IntensityWindowsSNParameters was tested on LC-MS datasets and uses the peak height as the signal level and the standard deviation of intensities around the peak as the noise level

  2. WaveletCoefficientsSNParameters was tested on GC-MS datasets and uses the continuous wavelet transform coefficients to estimate the signal and noise levels.

• sn_threshold: Signal to noise ratio threshold. The minimum signal to noise ratio a peak must have to be considered a real feature. Values greater than or equal to 7 will work well and will only detect a very small number of false positive peaks.

• coef_area_threshold: This is the best coefficient found by taking the inner product of the wavelet at the best scale and the peak, and then dividing by the area under the peak. Values around 100 work well for most data. Filters out bad peaks.

• min_feat_height: Minimum height of a feature. The smallest intensity a peak can have and be considered a real feature. Should be the same, or similar to start_intensity value of ADAPChromatogramBuilder.

Additional Notes on ``rt_for_cwt_scales_duration`` parameter. The parameter can be interpreted as the width range of the transformed peak. The applied transformation function (mother wavelet) is the so called mexican hat, corresponding to the second derivative of the Gaussian distribution or bell curve. Most importantly, the width of the center peak corresponds to the standard deviation \(\sigma\) and is a good estimator of the chromatogram peak width FWHM range. You might also check the original configuration parameter description in the adap user manual.

Again we configure the parameters in a way to obtain results comparable with the settings we used above for run_feature_finder_metabo.

[10]:

pd = mzmine.ADAPDetector()
pd.peak_duration = (5.0, 60.0)  # in seconds
pd.rt_for_cwt_scales_duration = (0.04, 3.0)
pd.sn_threshold = 10.0
pd.sn_estimators = (
    mzmine.IntensityWindowsSNParameters()
)  # since we have LC-MS data
pd.coef_area_threshold = 50
pd.min_feat_height = 5e4

Last not least, optional rsp_parameters allows removing shoulder (satellite) peaks. The corresponding RemoveShoulderPeaksParameters object has 2 configuration parameters:

• resolution: Mass resolution is the dimensionless ratio of the mass of the peak divided by its width. Peak width is taken as the full width at half maximum intensity (FWHM). default = 100’000

• peak_model: Peaks under the curve of this peak model will be removed. Allowed values: 'GAUSS', 'LORENTZ', 'LORENTZEXTENDED'; default = 'GAUSS'

Samples were acquired with a mass resolution of 30’000 at m/z 400. In case of Orbitrap instruments the resolution is \(R = R_{ref} \sqrt(\frac{{mz}_{ref}}{mz})\) and hence, the m/z acquisition range is important. Also keep in mind, the lower the resolution the broader the correction window width. We set it up as follows:

[11]:

rsp = mzmine.RemoveShoulderPeaksParameters(30000, "LORENTZ")

Finally, we can run mzmine feature detection:

[12]:

%%capture
ta_caulo = mzmine.pick_peaks(
    peakmap_caulobacter, chrom, pd, rsp_parameters=rsp
)

[13]:

ta_caulo.sort_by("parent_id")[:10]

[13]:

id	parent_id	mz	rt	mzmin	mzmax	rtmin	rtmax	area	height	width
int	int	MzType	RtType	MzType	MzType	RtType	RtType	float	float	RtType
1	-	85.075737	1.31 m	85.068855	85.083290	0.30 m	9.83 m	1.28e+07	2.01e+06	9.54 m
2	-	86.059761	1.25 m	86.059601	86.067551	0.59 m	6.18 m	1.37e+07	2.24e+06	5.59 m
3	-	86.079033	1.31 m	86.072250	86.082825	1.28 m	8.44 m	1.09e+06	8.67e+04	7.16 m
4	-	86.096146	1.13 m	86.096039	86.096184	1.10 m	2.54 m	9.79e+05	1.55e+05	1.44 m
5	-	87.063080	1.25 m	87.062820	87.070770	0.09 m	8.53 m	1.40e+06	6.27e+04	8.44 m
6	-	88.050133	1.36 m	88.042343	88.057983	1.07 m	9.67 m	1.03e+06	3.82e+04	8.60 m
7	-	88.075447	1.45 m	88.075363	88.075470	1.44 m	1.49 m	8.37e+05	4.78e+05	0.05 m
8	-	88.111633	0.97 m	88.103668	88.111786	0.93 m	8.50 m	2.16e+06	4.51e+04	7.57 m
9	-	89.059425	0.74 m	89.052620	89.060440	0.05 m	9.98 m	1.43e+06	1.65e+04	9.93 m
10	-	90.054634	4.00 m	90.047531	90.059875	0.58 m	10.00 m	1.70e+06	5.07e+04	9.42 m

Most columns of pick_peaks and run_feature_finder_metabo output Table are the same. pick_peaks provides an additional column parent_id referring to the id of the corresponding EIC. Hence, peaks originating from the same EIC have the same parent_id value and we can sort the table by parent_id to evaluate the peak detection process:

Evaluating the different chromatograms simplifies verifying peak extraction and optimization of peak detection parameters. Since we will no longer need the unsplitteded EICs for further processing, we can remove them via a simple filter command using the fact that they have an empty parent_id:

[14]:

ta_caulo = ta_caulo.filter(ta_caulo.parent_id.is_not_none())

At this state, isotopologue grouping is still missing. We can accomplish the grouping using ``mzmine.isotope_grouper()``. To this end we have to configure the ``IsotopeGrouperParameters`` object with parameters: - ``mz_tolerance``: Maximum allowed difference between two m/z values to be considered same. The value is specified both as absolute tolerance (in m/z) and relative tolerance (in ppm). The tolerance range is calculated using the maximum of the absolute and relative tolerances. - ``rt_tolerance``: Maximum allowed difference between two retention time values in seconds. Defined as a tuple: (is_abs_value, value) with is_abs_value being either True or False to define if the provided value is absolute or relative. As an example we assume an RT = 100s and rtol = (True, 0.5) rt tolerance equals 0.5s and (False, 0.5) equals 50s. - ``monotonic_shape``: If true, then monotonically decreasing height of isotope pattern is required. - ``maximum_charge``: Maximum charge to consider for detecting the isotope patterns. - ``representative_isotope``: peak, which should represent the whole isotope pattern. For low molecular weight compounds with monotonically decreasing isotope pattern, the most intense isotope should be representative. For high molecular weight molecules, the lowest m/z isotope may be the representative. Allowed values: 'Most intense', 'Lowest m/z'

Note, parameter mz_tolerance refers not to the mz values of isotopologues directly, but refers to the nominal isotopologue mass shift of neighboring isotopologues of an isotopologue pattern. The pattern results from the natural isotope distributions of compound elements and can be calculated (see also link). Hence, mz_tolerance should be selected in a way that it fullfills following condition for all compounds of interest:

\(\lvert (mz_{n+1, measured} -mz_{n+1, calculated}-mz_{n, measured} + mz_{n, calculated})*z\rvert\le mz_{tolerance}\)

with n corresponding to the nominal isotopologue number. Since most biomolecules are composed out of the elements C, H, N, O, P, S isotopologue shifts are mainly driven by \(^{13}C\), \(^{18}O\), \(^{34}S\) and to a certain extend by \(^{15}N\), spanning a mass range of (0.995 and 1.005), which corresponds to a mz_tolerance of about 0.005 for single charged ions. Since \(^{13}C\) is mainly repsonsible for metabolites M1 isotopologue, isotopologue grouping algorithm assums a default isotope mass shift of 1.0033 Da corresponding to the \(^{12}C\) - \(^{13}C\) mass difference. However, high mass resolution instruments are capable to separate \(^{13}C\) and \(^{15}N\) isotopologues and if you want to include the \(^{15}N\) isotopologue peak you should increase the mass tolerance to \(\pm\) 0.008.

For most metabolites with typical elemental composition, a monotonic shape can be assumed and hence the parameter monotonic_shape should be set to True. However, keep in mind that compounds rich in S atoms or some adducts can ommit isotopologue grouping when monotopic shape is assumed, i.e. grouping of \(M_{1}\) and \(M_{2}\) of ions with a Cl adduct will fail due to the high abundance of \(^{35}Cl\).

We configure the IsotopeGrouperParameters as follows:

[15]:

ig = mzmine.IsotopeGrouperParameters()
ig.mz_tolerance = (0.005, 0)
ig.rt_tolerance = (True, 2.0)  # (is_abs_value, value)
ig.monotonic_shape = True
ig.maximum_charge = 3
ig.representative_isotope = "Lowest m/z"

And we get our feature table by:

[16]:

ta_caulo_final = mzmine.isotope_grouper(ta_caulo, ig)

shutdown send alive token

got first alive token1711363218.9265552
wait for first alive token
json file
/tmp/tmpgddzwb8p.json
Mar 25, 2024 10:40:19 AM net.sf.mzmine.modules.rawdatamethods.rawdataimport.fileformats.MzMLReadTask run
INFO: Started parsing file /tmp/tmp7agqe5yd.mzML
Mar 25, 2024 10:40:19 AM uk.ac.ebi.jmzml.MzMLElement loadProperties
WARNING: MzIdentML Configuration file: jar:file:/root/.emzed3/emzed.ext.mzmine2/mzmine2/MZmine-2.41.2/lib/jmzml-1.7.11.jar!/defaultMzMLElement.cfg.xml

got 1711363219.4278626

Mar 25, 2024 10:40:20 AM net.sf.mzmine.modules.rawdatamethods.rawdataimport.fileformats.MzMLReadTask run
INFO: Finished parsing /tmp/tmp7agqe5yd.mzML, parsed 1066 scans
/tmp/tmp3e11ksan.csv

Mar 25, 2024 10:40:21 AM net.sf.mzmine.modules.peaklistmethods.isotopes.deisotoper.IsotopeGrouperTask run
INFO: Running isotopic peak grouper on

Mar 25, 2024 10:40:21 AM net.sf.mzmine.modules.peaklistmethods.isotopes.deisotoper.IsotopeGrouperTask run
INFO: Finished isotopic peak grouper on
write output file to /tmp/tmpvegf08d6.txt

extracted 1115 deconvolved peaks
extracted 792 deconvolved peaks

!!!DONE

[17]:

ta_caulo_final.sort_by("mz", ascending=False)[:10]

[17]:

id	parent_id	mz	rt	mzmin	mzmax	rtmin	rtmax	area	height	width	isotope_group_id	isotope_base_peak	isotope_group_size	isotope_charge
int	int	MzType	RtType	MzType	MzType	RtType	RtType	float	float	RtType	int	int	int	int
3495	2379	599.389374	1.01 m	599.388489	599.390259	0.97 m	1.08 m	3.64e+06	1.41e+06	0.11 m	-	-	-	-
3494	2371	597.875854	0.78 m	597.875366	597.877563	0.74 m	0.84 m	3.32e+05	1.28e+05	0.09 m	414	3492	2	2
3492	2370	597.374146	0.78 m	597.373413	597.375305	0.73 m	0.83 m	5.35e+05	1.74e+05	0.09 m	414	3492	2	2
3493	2370	597.374146	0.54 m	597.373596	597.376770	0.53 m	0.61 m	2.21e+05	7.46e+04	0.07 m	-	-	-	-
3491	2368	596.317200	0.50 m	596.312622	596.318970	0.47 m	0.54 m	4.72e+05	2.11e+05	0.07 m	40	3489	3	1
3490	2367	595.315369	0.48 m	595.314880	595.316895	0.47 m	0.56 m	2.46e+06	1.19e+06	0.10 m	40	3489	3	1
3489	2365	594.312195	0.48 m	594.310303	594.312561	0.47 m	0.56 m	8.60e+06	4.07e+06	0.10 m	40	3489	3	1
3486	2351	591.876099	0.72 m	591.875061	591.877075	0.67 m	0.78 m	3.99e+05	1.83e+05	0.10 m	331	3484	4	2
3487	2351	591.875244	1.15 m	591.873657	591.876282	1.12 m	1.24 m	2.62e+05	6.80e+04	0.12 m	331	3484	4	2
3488	2351	591.875092	0.96 m	591.873291	591.877380	0.91 m	0.98 m	1.56e+05	6.00e+04	0.07 m	331	3484	4	2

Note, only grouped peaks obtain additional identifiers, hence columns isotope_group_id and isotope_base_peak contain None values if peaks were not grouped.

Comparing feature_finder_metabo and ADAP¶

We can use the emzed Table method join to compare the feature detection results. Ideally, both methods will result in the same mz and rt values. We can test this using Table.join allowing only small mz and rt tolerances:

[18]:

adap = ta_caulo_final
ff = t_caulo
mztol = (0.001, 0.0)
rttol = (2.0, 0.0)  # in the range of about one spectrum distance
comp_adap_ff = adap.join(
    ff,
    adap.mz.approx_equal(ff.mz, *mztol)
    & adap.rt.approx_equal(ff.rt, *rttol),
)
print("number of common peaks:", len(comp_adap_ff))

shutdown send alive token

number of common peaks: 736

Note, the length of table comp_adap_ff does not necessarily correspond to the number of common peaks since in principle, a peak in one table can match several peaks in the other table and vice versa, leading to multiple entries for each peak. Here, a more general way to determine the number of common peaks:

[19]:

ids = comp_adap_ff.id.to_list()
no_peaks = len(set(ids))
print("number of common peaks:", no_peaks)

number of common peaks: 736

Ideally, both feature detection approaches result identical m/z and rt values for the same peaks. We can add a column to show the mz differences:

[20]:

t = comp_adap_ff
t.add_or_replace_column(
    "mz_delta", t.mz - t.mz__0, emzed.MzType, format_="%.1e"
)
t.add_or_replace_column(
    "mz_delta",
    t.apply(abs, t.mz_delta),
    emzed.MzType,
    format_="%.1e",
)
t = t.sort_by("mz_delta", ascending=False)
t[:10]

[20]:

id	parent_id	mz	rt	mzmin	mzmax	rtmin	rtmax	area	height	width	isotope_group_id	isotope_base_peak	isotope_group_size	isotope_charge	id__0	feature_id__0	feature_size__0	mz__0	mzmin__0	mzmax__0	rt__0	rtmin__0	rtmax__0	intensity__0	quality__0	fwhm__0	z__0	source__0	mz_delta
int	int	MzType	RtType	MzType	MzType	RtType	RtType	float	float	RtType	int	int	int	int	int	int	int	MzType	MzType	MzType	RtType	RtType	RtType	float	float	RtType	int	str	MzType
2718	618	245.605164	1.51 m	245.603516	245.610641	1.49 m	1.56 m	1.67e+05	6.15e+04	0.07 m	-	-	-	-	575	341	2	245.606116	245.602371	245.610641	1.53 m	1.49 m	1.58 m	2.02e+05	1.67e-04	0.06 m	2	AA_sample_caulobacter.mzML	9.5e-04
3187	1661	447.270050	0.53 m	447.269714	447.270508	0.52 m	0.59 m	9.05e+05	3.96e+05	0.07 m	-	-	-	-	581	346	1	447.269209	447.259308	447.272034	0.53 m	0.47 m	0.66 m	7.04e+05	1.64e-04	0.03 m	0	AA_sample_caulobacter.mzML	8.4e-04
3338	2033	520.332031	0.74 m	520.330017	520.333252	0.70 m	0.79 m	6.78e+05	2.83e+05	0.09 m	-	-	-	-	660	412	1	520.332712	520.330017	520.344360	0.74 m	0.71 m	0.82 m	5.35e+05	1.25e-04	0.03 m	0	AA_sample_caulobacter.mzML	6.8e-04
3422	2236	564.327545	0.52 m	564.321594	564.328308	0.50 m	0.57 m	1.33e+06	5.86e+05	0.07 m	-	-	-	-	390	195	2	564.328171	564.321594	564.341370	0.52 m	0.50 m	0.64 m	1.17e+06	3.53e-04	0.05 m	1	AA_sample_caulobacter.mzML	6.3e-04
2675	534	230.838531	1.53 m	230.835220	230.840469	1.46 m	1.60 m	3.98e+06	9.45e+05	0.14 m	132	2675	2	3	244	110	1	230.837950	230.835220	230.840469	1.53 m	1.45 m	1.59 m	3.26e+06	7.61e-04	0.08 m	0	AA_sample_caulobacter.mzML	5.8e-04
3491	2368	596.317200	0.50 m	596.312622	596.318970	0.47 m	0.54 m	4.72e+05	2.11e+05	0.07 m	40	3489	3	1	113	45	3	596.317778	596.312622	596.320618	0.49 m	0.47 m	0.56 m	4.03e+05	2.15e-03	0.05 m	1	AA_sample_caulobacter.mzML	5.8e-04
3262	1853	484.314926	1.57 m	484.312286	484.315338	1.54 m	1.64 m	4.49e+05	1.68e+05	0.10 m	123	3257	6	3	332	155	3	484.314352	484.312286	484.315338	1.58 m	1.50 m	1.63 m	4.71e+05	5.00e-04	0.06 m	3	AA_sample_caulobacter.mzML	5.7e-04
3375	2097	533.306763	0.52 m	533.304932	533.307312	0.50 m	0.59 m	1.41e+06	6.08e+05	0.09 m	181	3375	2	1	395	199	2	533.306200	533.296936	533.307312	0.52 m	0.46 m	0.61 m	1.19e+06	3.38e-04	0.04 m	1	AA_sample_caulobacter.mzML	5.6e-04
3293	1933	499.283325	1.55 m	499.281006	499.284668	1.48 m	1.58 m	6.57e+05	1.74e+05	0.10 m	23	3286	3	1	41	15	3	499.282783	499.279083	499.284668	1.54 m	1.47 m	1.58 m	6.34e+05	5.55e-03	0.06 m	1	AA_sample_caulobacter.mzML	5.4e-04
3434	2261	570.339050	0.80 m	570.331909	570.341797	0.75 m	0.84 m	3.59e+05	1.34e+05	0.09 m	291	3431	3	3	535	308	2	570.338545	570.331909	570.341797	0.80 m	0.75 m	0.86 m	2.92e+05	1.96e-04	0.05 m	3	AA_sample_caulobacter.mzML	5.1e-04

We observe rather small m/z variances at the 4th digit which are below mass accuracy of the instrument. Exceptions seem to be due to bad quality i.e. peak id==2722 turned out to be a satellite peak which has not been removed. Summarized, peaks detected with both methods give the same results for common detected peaks.

Next we evaluate peaks exclusively extracted with only one out of the two detectors: the joined table provides peaks ids of those peaks detected with both tools. With the method left_join we can also find peaks detected with only one out of the two tools, since rows of the joined table contain only None values.

[21]:

# mzmine2
comp_adap_ff = adap.left_join(
    ff,
    adap.mz.approx_equal(ff.mz, *mztol)
    & adap.rt.approx_equal(ff.rt, *rttol),
)
adap_only = comp_adap_ff.filter(comp_adap_ff.id__0.is_none())
# open ms
comp_ff_adap = ff.left_join(
    adap,
    ff.mz.approx_equal(adap.mz, *mztol)
    & ff.rt.approx_equal(adap.rt, *rttol),
)
ff_only = comp_ff_adap.filter(comp_ff_adap.id__0.is_none())
print(f"only by ADAP: {len(adap_only)}")
print(f"only by run_feature_finder_metabo: {len(ff_only)}")

only by ADAP: 379
only by run_feature_finder_metabo: 308

Note, here we do not build a set from id values to count the number of peaks, since by definition there are exists no similar peak in the joint table.

Figure 5: Top 10 peaks exclusively detected with ff_metabo (A) and ADAP (B).

When evaluating exclusive peaks, both approaches miss significant peaks (Figure 5). For given parameters, ff_metabo detects about 4 % more peaks than ADAP but it ff_metabo misses more high quality peaks.

SUMMARY: Both approaches give similar results with a slightly better performance of the ADAP. However, some further parameter optimization might close the gap between the two methods. ADAP has a clear advantage in case of Orbitrap instruments data since it can directly remove shoulder peaks. On the other hand ff_metabo is about 100x faster than ADAP. This is a clear advantage when analyzing huge data sets.

Peak deconvolution (adduct grouping)¶

[22]:

emzed.adducts.all[:10]

[22]:

[23]:

emzed.adducts.all[-10:]

[23]:

[24]:

adds = emzed.adducts.positive_single_charged
adds

[24]:

[25]:

annotate = emzed.annotate.annotate_adducts
help(annotate)  # more explanat

Help on function annotate_adducts in module emzed.annotate.annotate_adducts:

annotate_adducts(peaks, adducts, mz_tol, rt_tol, explained_abundance=0.2)
    attempts to group peaks as adducts.

    required column names are `mz` and `rt` only. all other columns will be ignored.

[26]:

%%capture
ids = [20, 21, 22, 24, 25, 32, 36, 38, 39]
adducts = adds.filter(adds.id.is_in(ids))
t_annotate = annotate(adap, adducts, 0.002, 1.5, 0.3)

[27]:

t_annotate.sort_by("adduct_cluster_id", ascending=False)[:10]

[27]:

[28]:

print("before:", len(adap), "\nafter :", len(t_annotate))

before: 1115
after : 311

[29]:

from collections import Counter

id2counts = Counter(t_annotate.id)
# peaks with ids present more than ones show ambiguous grouping
ids = [key for key, counts in id2counts.items() if counts > 1]
# next we can filter our table for selected peaks
t = t_annotate.filter(t_annotate.id.is_in(ids))
# we now obtain the associated adduct_cluster_ids
ac_is = t.adduct_cluster_id.to_list()
# now we can access the subtable of interest:
t_cases = t_annotate.filter(t_annotate.adduct_cluster_id.is_in(ac_is))
t_cases[:10]

[29]:

[30]:

rt_tolerance = (False, 1.0)
adducts = (
    ("[M+Na-H]", 21.9825),
    ("[M+K-H]", 37.9559),
    ("[M+Mg-2H]", 21.9694),
    ("[M+NH3]", 17.0265),
    ("[M+H3PO4]", 97.9769),
    ("[M+H2SO4]", 97.9674),
    ("[M+H2CO3]", 62.0004),
    ("[(Deuterium)]glycerol", 5.0),
)
mz_tolerance = (0.001, 0)
max_adduct_height = 1.0

[31]:

adds = emzed.adducts.positive_single_charged
ids = [21, 22, 24, 25, 32]
adducts = adds.filter(adds.id.is_in(ids))
a = adducts

# we calculate the mass shift relative to the M+H ion
expr = (
    a.apply(emzed.mass.of, a.adduct_add)
    - a.apply(emzed.mass.of, a.adduct_sub)
    - emzed.mass.H1
)
a.add_column("mass_shift", expr, float)
# We build the (name, mass_shift) tuple with mass shift relative to M+H
adducts = list(zip(a.adduct_name, a.mass_shift))
adducts

[31]:

[('M+NH4', 17.0265490957),
 ('M+Na', 21.981944248999998),
 ('M+H-H2O', -18.0105650638),
 ('M+K', 37.955881648100004),
 ('M+2K-H', 75.91176329620001)]

[32]:

annotated = ta_caulo_annotated.filter(
    ta_caulo_annotated.adduct_annotation.is_not_none()
)
annotated[:10]

[32]:

Pubchem database matching¶

The ``emzed.db`` module supports database matching using a local subset of the pubchem data base, which comprises all compounds annotated in KEGG, BIOCYC, and HMDB. When using the local database for the first time, you have to run the command:

[33]:

emzed.db.update_pubchem()

update pubchem: 00:00:00.0 [ ]

</pre>

update pubchem: 00:00:00.0 [ ]

end{sphinxVerbatim}

update pubchem: 00:00:00.0 [ ]

update pubchem: 00:00:00.1 [| ]

</pre>

update pubchem: 00:00:00.1 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.1 [| ]

update pubchem: 00:00:00.1 [|| ]

</pre>

update pubchem: 00:00:00.1 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.1 [|| ]

update pubchem: 00:00:00.2 [||| ]

</pre>

update pubchem: 00:00:00.2 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.2 [||| ]

update pubchem: 00:00:00.3 [|||| ]

</pre>

update pubchem: 00:00:00.3 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.3 [|||| ]

update pubchem: 00:00:00.4 [||||| ]

</pre>

update pubchem: 00:00:00.4 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.4 [||||| ]

update pubchem: 00:00:00.4 [|||||| ]

</pre>

update pubchem: 00:00:00.4 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.4 [|||||| ]

update pubchem: 00:00:00.5 [||||||| ]

</pre>

update pubchem: 00:00:00.5 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.5 [||||||| ]

update pubchem: 00:00:00.6 [|||||||| ]

</pre>

update pubchem: 00:00:00.6 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.6 [|||||||| ]

update pubchem: 00:00:00.6 [||||||||| ]

</pre>

update pubchem: 00:00:00.6 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.6 [||||||||| ]

update pubchem: 00:00:00.7 [|||||||||| ]

</pre>

update pubchem: 00:00:00.7 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.7 [|||||||||| ]

update pubchem: 00:00:00.8 [||||||||||| ]

</pre>

update pubchem: 00:00:00.8 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.8 [||||||||||| ]

update pubchem: 00:00:00.9 [|||||||||||| ]

</pre>

update pubchem: 00:00:00.9 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.9 [|||||||||||| ]

update pubchem: 00:00:00.9 [||||||||||||| ]

</pre>

update pubchem: 00:00:00.9 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:00.9 [||||||||||||| ]

update pubchem: 00:00:01.0 [|||||||||||||| ]

</pre>

update pubchem: 00:00:01.0 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:01.0 [|||||||||||||| ]

update pubchem: 00:00:01.1 [|||||||||||||||]

</pre>

update pubchem: 00:00:01.1 [|||||||||||||||]

end{sphinxVerbatim}

update pubchem: 00:00:01.1 [|||||||||||||||]

update pubchem: 00:00:01.2 [|||||||||||||| ]

</pre>

update pubchem: 00:00:01.2 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:01.2 [|||||||||||||| ]

update pubchem: 00:00:01.2 [||||||||||||| ]

</pre>

update pubchem: 00:00:01.2 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:01.2 [||||||||||||| ]

update pubchem: 00:00:01.5 [|||||||||||| ]

</pre>

update pubchem: 00:00:01.5 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:01.5 [|||||||||||| ]

update pubchem: 00:00:01.6 [||||||||||| ]

</pre>

update pubchem: 00:00:01.6 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:01.6 [||||||||||| ]

update pubchem: 00:00:01.8 [|||||||||| ]

</pre>

update pubchem: 00:00:01.8 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:01.8 [|||||||||| ]

update pubchem: 00:00:02.0 [||||||||| ]

</pre>

update pubchem: 00:00:02.0 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:02.0 [||||||||| ]

update pubchem: 00:00:02.1 [|||||||| ]

</pre>

update pubchem: 00:00:02.1 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:02.1 [|||||||| ]

update pubchem: 00:00:02.4 [||||||| ]

</pre>

update pubchem: 00:00:02.4 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:02.4 [||||||| ]

update pubchem: 00:00:02.7 [|||||| ]

</pre>

update pubchem: 00:00:02.7 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:02.7 [|||||| ]

update pubchem: 00:00:02.8 [||||| ]

</pre>

update pubchem: 00:00:02.8 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:02.8 [||||| ]

update pubchem: 00:00:02.9 [|||| ]

</pre>

update pubchem: 00:00:02.9 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:02.9 [|||| ]

update pubchem: 00:00:03.1 [||| ]

</pre>

update pubchem: 00:00:03.1 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:03.1 [||| ]

update pubchem: 00:00:03.5 [|| ]

</pre>

update pubchem: 00:00:03.5 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:03.5 [|| ]

update pubchem: 00:00:03.6 [| ]

</pre>

update pubchem: 00:00:03.6 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:03.6 [| ]

update pubchem: 00:00:03.7 [ ]

</pre>

update pubchem: 00:00:03.7 [ ]

end{sphinxVerbatim}

update pubchem: 00:00:03.7 [ ]

update pubchem: 00:00:03.8 [| ]

</pre>

update pubchem: 00:00:03.8 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:03.8 [| ]

update pubchem: 00:00:03.9 [|| ]

</pre>

update pubchem: 00:00:03.9 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:03.9 [|| ]

update pubchem: 00:00:04.2 [||| ]

</pre>

update pubchem: 00:00:04.2 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:04.2 [||| ]

update pubchem: 00:00:04.4 [|||| ]

</pre>

update pubchem: 00:00:04.4 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:04.4 [|||| ]

update pubchem: 00:00:04.5 [||||| ]

</pre>

update pubchem: 00:00:04.5 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:04.5 [||||| ]

update pubchem: 00:00:05.0 [|||||| ]

</pre>

update pubchem: 00:00:05.0 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:05.0 [|||||| ]

update pubchem: 00:00:05.2 [||||||| ]

</pre>

update pubchem: 00:00:05.2 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:05.2 [||||||| ]

update pubchem: 00:00:05.2 [|||||||| ]

</pre>

update pubchem: 00:00:05.2 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:05.2 [|||||||| ]

update pubchem: 00:00:05.4 [||||||||| ]

</pre>

update pubchem: 00:00:05.4 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:05.4 [||||||||| ]

update pubchem: 00:00:05.8 [|||||||||| ]

</pre>

update pubchem: 00:00:05.8 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:05.8 [|||||||||| ]

update pubchem: 00:00:05.9 [||||||||||| ]

</pre>

update pubchem: 00:00:05.9 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:05.9 [||||||||||| ]

update pubchem: 00:00:06.0 [|||||||||||| ]

</pre>

update pubchem: 00:00:06.0 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.0 [|||||||||||| ]

update pubchem: 00:00:06.1 [||||||||||||| ]

</pre>

update pubchem: 00:00:06.1 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.1 [||||||||||||| ]

update pubchem: 00:00:06.1 [|||||||||||||| ]

</pre>

update pubchem: 00:00:06.1 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.1 [|||||||||||||| ]

update pubchem: 00:00:06.2 [|||||||||||||||]

</pre>

update pubchem: 00:00:06.2 [|||||||||||||||]

end{sphinxVerbatim}

update pubchem: 00:00:06.2 [|||||||||||||||]

update pubchem: 00:00:06.3 [|||||||||||||| ]

</pre>

update pubchem: 00:00:06.3 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.3 [|||||||||||||| ]

update pubchem: 00:00:06.3 [||||||||||||| ]

</pre>

update pubchem: 00:00:06.3 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.3 [||||||||||||| ]

update pubchem: 00:00:06.4 [|||||||||||| ]

</pre>

update pubchem: 00:00:06.4 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.4 [|||||||||||| ]

update pubchem: 00:00:06.5 [||||||||||| ]

</pre>

update pubchem: 00:00:06.5 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.5 [||||||||||| ]

update pubchem: 00:00:06.6 [|||||||||| ]

</pre>

update pubchem: 00:00:06.6 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.6 [|||||||||| ]

update pubchem: 00:00:06.6 [||||||||| ]

</pre>

update pubchem: 00:00:06.6 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.6 [||||||||| ]

update pubchem: 00:00:06.7 [|||||||| ]

</pre>

update pubchem: 00:00:06.7 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.7 [|||||||| ]

update pubchem: 00:00:06.8 [||||||| ]

</pre>

update pubchem: 00:00:06.8 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.8 [||||||| ]

update pubchem: 00:00:06.8 [|||||| ]

</pre>

update pubchem: 00:00:06.8 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.8 [|||||| ]

update pubchem: 00:00:06.9 [||||| ]

</pre>

update pubchem: 00:00:06.9 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:06.9 [||||| ]

update pubchem: 00:00:07.0 [|||| ]

</pre>

update pubchem: 00:00:07.0 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.0 [|||| ]

update pubchem: 00:00:07.1 [||| ]

</pre>

update pubchem: 00:00:07.1 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.1 [||| ]

update pubchem: 00:00:07.1 [|| ]

</pre>

update pubchem: 00:00:07.1 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.1 [|| ]

update pubchem: 00:00:07.2 [| ]

</pre>

update pubchem: 00:00:07.2 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.2 [| ]

update pubchem: 00:00:07.3 [ ]

</pre>

update pubchem: 00:00:07.3 [ ]

end{sphinxVerbatim}

update pubchem: 00:00:07.3 [ ]

update pubchem: 00:00:07.4 [| ]

</pre>

update pubchem: 00:00:07.4 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.4 [| ]

update pubchem: 00:00:07.4 [|| ]

</pre>

update pubchem: 00:00:07.4 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.4 [|| ]

update pubchem: 00:00:07.5 [||| ]

</pre>

update pubchem: 00:00:07.5 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.5 [||| ]

update pubchem: 00:00:07.6 [|||| ]

</pre>

update pubchem: 00:00:07.6 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.6 [|||| ]

update pubchem: 00:00:07.7 [||||| ]

</pre>

update pubchem: 00:00:07.7 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.7 [||||| ]

update pubchem: 00:00:07.7 [|||||| ]

</pre>

update pubchem: 00:00:07.7 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.7 [|||||| ]

update pubchem: 00:00:07.8 [||||||| ]

</pre>

update pubchem: 00:00:07.8 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.8 [||||||| ]

update pubchem: 00:00:07.9 [|||||||| ]

</pre>

update pubchem: 00:00:07.9 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.9 [|||||||| ]

update pubchem: 00:00:07.9 [||||||||| ]

</pre>

update pubchem: 00:00:07.9 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:07.9 [||||||||| ]

update pubchem: 00:00:08.0 [|||||||||| ]

</pre>

update pubchem: 00:00:08.0 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.0 [|||||||||| ]

update pubchem: 00:00:08.1 [||||||||||| ]

</pre>

update pubchem: 00:00:08.1 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.1 [||||||||||| ]

update pubchem: 00:00:08.2 [|||||||||||| ]

</pre>

update pubchem: 00:00:08.2 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.2 [|||||||||||| ]

update pubchem: 00:00:08.2 [||||||||||||| ]

</pre>

update pubchem: 00:00:08.2 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.2 [||||||||||||| ]

update pubchem: 00:00:08.3 [|||||||||||||| ]

</pre>

update pubchem: 00:00:08.3 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.3 [|||||||||||||| ]

update pubchem: 00:00:08.4 [|||||||||||||||]

</pre>

update pubchem: 00:00:08.4 [|||||||||||||||]

end{sphinxVerbatim}

update pubchem: 00:00:08.4 [|||||||||||||||]

update pubchem: 00:00:08.5 [|||||||||||||| ]

</pre>

update pubchem: 00:00:08.5 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.5 [|||||||||||||| ]

update pubchem: 00:00:08.5 [||||||||||||| ]

</pre>

update pubchem: 00:00:08.5 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.5 [||||||||||||| ]

update pubchem: 00:00:08.6 [|||||||||||| ]

</pre>

update pubchem: 00:00:08.6 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.6 [|||||||||||| ]

update pubchem: 00:00:08.7 [||||||||||| ]

</pre>

update pubchem: 00:00:08.7 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.7 [||||||||||| ]

update pubchem: 00:00:08.7 [|||||||||| ]

</pre>

update pubchem: 00:00:08.7 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.7 [|||||||||| ]

update pubchem: 00:00:08.8 [||||||||| ]

</pre>

update pubchem: 00:00:08.8 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.8 [||||||||| ]

update pubchem: 00:00:08.9 [|||||||| ]

</pre>

update pubchem: 00:00:08.9 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:08.9 [|||||||| ]

update pubchem: 00:00:09.0 [||||||| ]

</pre>

update pubchem: 00:00:09.0 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.0 [||||||| ]

update pubchem: 00:00:09.0 [|||||| ]

</pre>

update pubchem: 00:00:09.0 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.0 [|||||| ]

update pubchem: 00:00:09.1 [||||| ]

</pre>

update pubchem: 00:00:09.1 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.1 [||||| ]

update pubchem: 00:00:09.2 [|||| ]

</pre>

update pubchem: 00:00:09.2 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.2 [|||| ]

update pubchem: 00:00:09.2 [||| ]

</pre>

update pubchem: 00:00:09.2 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.2 [||| ]

update pubchem: 00:00:09.3 [|| ]

</pre>

update pubchem: 00:00:09.3 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.3 [|| ]

update pubchem: 00:00:09.4 [| ]

</pre>

update pubchem: 00:00:09.4 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.4 [| ]

update pubchem: 00:00:09.5 [ ]

</pre>

update pubchem: 00:00:09.5 [ ]

end{sphinxVerbatim}

update pubchem: 00:00:09.5 [ ]

update pubchem: 00:00:09.5 [| ]

</pre>

update pubchem: 00:00:09.5 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.5 [| ]

update pubchem: 00:00:09.6 [|| ]

</pre>

update pubchem: 00:00:09.6 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.6 [|| ]

update pubchem: 00:00:09.7 [||| ]

</pre>

update pubchem: 00:00:09.7 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.7 [||| ]

update pubchem: 00:00:09.7 [|||| ]

</pre>

update pubchem: 00:00:09.7 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.7 [|||| ]

update pubchem: 00:00:09.8 [||||| ]

</pre>

update pubchem: 00:00:09.8 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.8 [||||| ]

update pubchem: 00:00:09.9 [|||||| ]

</pre>

update pubchem: 00:00:09.9 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:09.9 [|||||| ]

update pubchem: 00:00:10.0 [||||||| ]

</pre>

update pubchem: 00:00:10.0 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.0 [||||||| ]

update pubchem: 00:00:10.0 [|||||||| ]

</pre>

update pubchem: 00:00:10.0 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.0 [|||||||| ]

update pubchem: 00:00:10.1 [||||||||| ]

</pre>

update pubchem: 00:00:10.1 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.1 [||||||||| ]

update pubchem: 00:00:10.2 [|||||||||| ]

</pre>

update pubchem: 00:00:10.2 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.2 [|||||||||| ]

update pubchem: 00:00:10.2 [||||||||||| ]

</pre>

update pubchem: 00:00:10.2 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.2 [||||||||||| ]

update pubchem: 00:00:10.3 [|||||||||||| ]

</pre>

update pubchem: 00:00:10.3 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.3 [|||||||||||| ]

update pubchem: 00:00:10.4 [||||||||||||| ]

</pre>

update pubchem: 00:00:10.4 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.4 [||||||||||||| ]

update pubchem: 00:00:10.5 [|||||||||||||| ]

</pre>

update pubchem: 00:00:10.5 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.5 [|||||||||||||| ]

update pubchem: 00:00:10.5 [|||||||||||||||]

</pre>

update pubchem: 00:00:10.5 [|||||||||||||||]

end{sphinxVerbatim}

update pubchem: 00:00:10.5 [|||||||||||||||]

update pubchem: 00:00:10.6 [|||||||||||||| ]

</pre>

update pubchem: 00:00:10.6 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.6 [|||||||||||||| ]

update pubchem: 00:00:10.7 [||||||||||||| ]

</pre>

update pubchem: 00:00:10.7 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.7 [||||||||||||| ]

update pubchem: 00:00:10.8 [|||||||||||| ]

</pre>

update pubchem: 00:00:10.8 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.8 [|||||||||||| ]

update pubchem: 00:00:10.8 [||||||||||| ]

</pre>

update pubchem: 00:00:10.8 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.8 [||||||||||| ]

update pubchem: 00:00:10.9 [|||||||||| ]

</pre>

update pubchem: 00:00:10.9 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:10.9 [|||||||||| ]

update pubchem: 00:00:11.0 [||||||||| ]

</pre>

update pubchem: 00:00:11.0 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.0 [||||||||| ]

update pubchem: 00:00:11.0 [|||||||| ]

</pre>

update pubchem: 00:00:11.0 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.0 [|||||||| ]

update pubchem: 00:00:11.1 [||||||| ]

</pre>

update pubchem: 00:00:11.1 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.1 [||||||| ]

update pubchem: 00:00:11.2 [|||||| ]

</pre>

update pubchem: 00:00:11.2 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.2 [|||||| ]

update pubchem: 00:00:11.3 [||||| ]

</pre>

update pubchem: 00:00:11.3 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.3 [||||| ]

update pubchem: 00:00:11.3 [|||| ]

</pre>

update pubchem: 00:00:11.3 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.3 [|||| ]

update pubchem: 00:00:11.4 [||| ]

</pre>

update pubchem: 00:00:11.4 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.4 [||| ]

update pubchem: 00:00:11.5 [|| ]

</pre>

update pubchem: 00:00:11.5 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.5 [|| ]

update pubchem: 00:00:11.5 [| ]

</pre>

update pubchem: 00:00:11.5 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.5 [| ]

update pubchem: 00:00:11.6 [ ]

</pre>

update pubchem: 00:00:11.6 [ ]

end{sphinxVerbatim}

update pubchem: 00:00:11.6 [ ]

update pubchem: 00:00:11.7 [| ]

</pre>

update pubchem: 00:00:11.7 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.7 [| ]

update pubchem: 00:00:11.8 [|| ]

</pre>

update pubchem: 00:00:11.8 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.8 [|| ]

update pubchem: 00:00:11.8 [||| ]

</pre>

update pubchem: 00:00:11.8 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.8 [||| ]

update pubchem: 00:00:11.9 [|||| ]

</pre>

update pubchem: 00:00:11.9 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:11.9 [|||| ]

update pubchem: 00:00:12.0 [||||| ]

</pre>

update pubchem: 00:00:12.0 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.0 [||||| ]

update pubchem: 00:00:12.0 [|||||| ]

</pre>

update pubchem: 00:00:12.0 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.0 [|||||| ]

update pubchem: 00:00:12.1 [||||||| ]

</pre>

update pubchem: 00:00:12.1 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.1 [||||||| ]

update pubchem: 00:00:12.2 [|||||||| ]

</pre>

update pubchem: 00:00:12.2 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.2 [|||||||| ]

update pubchem: 00:00:12.3 [||||||||| ]

</pre>

update pubchem: 00:00:12.3 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.3 [||||||||| ]

update pubchem: 00:00:12.3 [|||||||||| ]

</pre>

update pubchem: 00:00:12.3 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.3 [|||||||||| ]

update pubchem: 00:00:12.4 [||||||||||| ]

</pre>

update pubchem: 00:00:12.4 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.4 [||||||||||| ]

update pubchem: 00:00:12.5 [|||||||||||| ]

</pre>

update pubchem: 00:00:12.5 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.5 [|||||||||||| ]

update pubchem: 00:00:12.5 [||||||||||||| ]

</pre>

update pubchem: 00:00:12.5 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.5 [||||||||||||| ]

update pubchem: 00:00:12.6 [|||||||||||||| ]

</pre>

update pubchem: 00:00:12.6 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.6 [|||||||||||||| ]

update pubchem: 00:00:12.7 [|||||||||||||||]

</pre>

update pubchem: 00:00:12.7 [|||||||||||||||]

end{sphinxVerbatim}

update pubchem: 00:00:12.7 [|||||||||||||||]

update pubchem: 00:00:12.8 [|||||||||||||| ]

</pre>

update pubchem: 00:00:12.8 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.8 [|||||||||||||| ]

update pubchem: 00:00:12.8 [||||||||||||| ]

</pre>

update pubchem: 00:00:12.8 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.8 [||||||||||||| ]

update pubchem: 00:00:12.9 [|||||||||||| ]

</pre>

update pubchem: 00:00:12.9 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:12.9 [|||||||||||| ]

update pubchem: 00:00:13.0 [||||||||||| ]

</pre>

update pubchem: 00:00:13.0 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.0 [||||||||||| ]

update pubchem: 00:00:13.0 [|||||||||| ]

</pre>

update pubchem: 00:00:13.0 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.0 [|||||||||| ]

update pubchem: 00:00:13.1 [||||||||| ]

</pre>

update pubchem: 00:00:13.1 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.1 [||||||||| ]

update pubchem: 00:00:13.2 [|||||||| ]

</pre>

update pubchem: 00:00:13.2 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.2 [|||||||| ]

update pubchem: 00:00:13.3 [||||||| ]

</pre>

update pubchem: 00:00:13.3 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.3 [||||||| ]

update pubchem: 00:00:13.3 [|||||| ]

</pre>

update pubchem: 00:00:13.3 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.3 [|||||| ]

update pubchem: 00:00:13.4 [||||| ]

</pre>

update pubchem: 00:00:13.4 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.4 [||||| ]

update pubchem: 00:00:13.5 [|||| ]

</pre>

update pubchem: 00:00:13.5 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.5 [|||| ]

update pubchem: 00:00:13.6 [||| ]

</pre>

update pubchem: 00:00:13.6 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.6 [||| ]

update pubchem: 00:00:13.6 [|| ]

</pre>

update pubchem: 00:00:13.6 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.6 [|| ]

update pubchem: 00:00:13.7 [| ]

</pre>

update pubchem: 00:00:13.7 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.7 [| ]

update pubchem: 00:00:13.8 [ ]

</pre>

update pubchem: 00:00:13.8 [ ]

end{sphinxVerbatim}

update pubchem: 00:00:13.8 [ ]

update pubchem: 00:00:13.8 [| ]

</pre>

update pubchem: 00:00:13.8 [| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.8 [| ]

update pubchem: 00:00:13.9 [|| ]

</pre>

update pubchem: 00:00:13.9 [|| ]

end{sphinxVerbatim}

update pubchem: 00:00:13.9 [|| ]

update pubchem: 00:00:14.0 [||| ]

</pre>

update pubchem: 00:00:14.0 [||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.0 [||| ]

update pubchem: 00:00:14.1 [|||| ]

</pre>

update pubchem: 00:00:14.1 [|||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.1 [|||| ]

update pubchem: 00:00:14.1 [||||| ]

</pre>

update pubchem: 00:00:14.1 [||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.1 [||||| ]

update pubchem: 00:00:14.2 [|||||| ]

</pre>

update pubchem: 00:00:14.2 [|||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.2 [|||||| ]

update pubchem: 00:00:14.3 [||||||| ]

</pre>

update pubchem: 00:00:14.3 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.3 [||||||| ]

update pubchem: 00:00:14.3 [|||||||| ]

</pre>

update pubchem: 00:00:14.3 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.3 [|||||||| ]

update pubchem: 00:00:14.4 [||||||||| ]

</pre>

update pubchem: 00:00:14.4 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.4 [||||||||| ]

update pubchem: 00:00:14.5 [|||||||||| ]

</pre>

update pubchem: 00:00:14.5 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.5 [|||||||||| ]

update pubchem: 00:00:14.6 [||||||||||| ]

</pre>

update pubchem: 00:00:14.6 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.6 [||||||||||| ]

update pubchem: 00:00:14.6 [|||||||||||| ]

</pre>

update pubchem: 00:00:14.6 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.6 [|||||||||||| ]

update pubchem: 00:00:14.7 [||||||||||||| ]

</pre>

update pubchem: 00:00:14.7 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.7 [||||||||||||| ]

update pubchem: 00:00:14.8 [|||||||||||||| ]

</pre>

update pubchem: 00:00:14.8 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.8 [|||||||||||||| ]

update pubchem: 00:00:14.8 [|||||||||||||||]

</pre>

update pubchem: 00:00:14.8 [|||||||||||||||]

end{sphinxVerbatim}

update pubchem: 00:00:14.8 [|||||||||||||||]

update pubchem: 00:00:14.9 [|||||||||||||| ]

</pre>

update pubchem: 00:00:14.9 [|||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:14.9 [|||||||||||||| ]

update pubchem: 00:00:15.0 [||||||||||||| ]

</pre>

update pubchem: 00:00:15.0 [||||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.0 [||||||||||||| ]

update pubchem: 00:00:15.1 [|||||||||||| ]

</pre>

update pubchem: 00:00:15.1 [|||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.1 [|||||||||||| ]

update pubchem: 00:00:15.1 [||||||||||| ]

</pre>

update pubchem: 00:00:15.1 [||||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.1 [||||||||||| ]

update pubchem: 00:00:15.2 [|||||||||| ]

</pre>

update pubchem: 00:00:15.2 [|||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.2 [|||||||||| ]

update pubchem: 00:00:15.3 [||||||||| ]

</pre>

update pubchem: 00:00:15.3 [||||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.3 [||||||||| ]

update pubchem: 00:00:15.3 [|||||||| ]

</pre>

update pubchem: 00:00:15.3 [|||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.3 [|||||||| ]

update pubchem: 00:00:15.4 [||||||| ]

</pre>

update pubchem: 00:00:15.4 [||||||| ]

end{sphinxVerbatim}

update pubchem: 00:00:15.4 [||||||| ]

update pubchem: 00:00:15.5 [|||||| ]

</pre>

update pubchem: 00:00:15.5 [|||||| ]

end{sphinxVerbatim}