METHOD FOR EVALUATING MASS SPECTROMETRY DATA AND MASS SPECTROMETRIC METHOD AS WELL AS A MALDI TOF MASS SPECTROMETER

Description

[0001] The invention relates to a method for evaluating mass spectrometry data for analyzing peptides from biological samples, in particular MALDI-TOF mass spectrometry according to claims 1 and 8. Furthermore, the invention relates to a mass spectrometric method for analyzing peptides from biological samples, in particular using a MALDI-TOF mass spectrometer according to claim 12. Finally, the invention is directed to a mass spectrometric method for analyzing peptides from biological samples according to claims 13 and 17, as well as a MALDI-TOF mass spectrometer according to claim 19.

[0002] In matrix-assisted laser desorption / ionization time-of-flight mass spectrometry (MALDI-TOF-MS), a biological tissue sample is coated with a matrix solution after suitable sample preparation and then bombarded with a laser in a vacuum. This dissolves biological macromolecules from the tissue and ionizes them, typically with a singly positive charge. The ions are then accelerated in an electric field and registered by a detector. The m / z value, i.e. the relationship between the mass and charge of the molecule, can be determined from the time of flight. The measured mass spectrum represents the relative number of registered ions (spectral intensity) as a function of their m / z values. Assuming singly positive ionization, the m / z value is equivalent to the mass m of the ionized molecule. For simplicity, the mass m of the ionized molecule is referred to below as the m / z value.

[0003] The m / z value or molecular mass is given in Daltons (Da) as a multiple of the atomic mass unit (1 Da = 1 amu, atomic mass unit).

[0004] The mass of a molecule in Da approximately corresponds to the total number of protons and neutrons that make up the molecule's atomic nuclei. The difference between this integer nominal mass and the actual mass is called the mass defect. The mass defect of a molecule is the sum of the mass defects of the individual atoms, which in turn vary for each chemical element or isotope.

[0005] The term "mass defect" is not used consistently in the literature. A first meaning of the term refers to the difference in masses in the SI unit kg. A second meaning also refers to the difference in masses, but is based on the atomic mass unit u, which is defined with reference to the carbon isotope 12C. Due to this definition, the mass defect of the carbon isotope 12C is zero. For this second meaning, the term "mass excess" is also used instead of the term "mass defect" to emphasize the difference from the first meaning mentioned above. In biology and chemistry, however, especially in connection with mass spectrometric methods, the term "mass defect" is still used, i.e., in the sense of the second meaning, as is the case here.

[0006] The acquisition of MALDI-TOF mass spectrometry data from biological tissue sections yields a wealth of information about the proteomic structure of the tissue samples. At the same time, the measurement is subject to a number of potential interferences that can lead to distortions and falsifications of the obtained information. Due to the high complexity of the data, an objective assessment of its quality and accuracy is often impossible. Currently, there are no widely accepted and easily applicable metrics that allow a statement about the data quality of a measurement or whether two measurements provide data of comparable quality.

[0007] US 2016 / 0003842 A1 discloses a mass spectrometric method for the analysis of peptides. The goal is the detection of so-called glycopeptides. Figures 2a and 2b of US 2016 / 0003842 A1 show mass defects plotted against nominal masses (m / z). A distinction is made between regions containing peptides and regions enriched with glycopeptides.

[0008] A common distortion of measured data is a systematic error in the measured masses, which exceeds the metrological tolerance limit even under carefully controlled experimental conditions. Conventional methods for correcting such mass distortions are often either too inaccurate or too time-consuming.

[0009] The study "Shifting Unoccupied Spectral Space in Mass Spectrum of Peptide Fragment Ions" (Bekim Bajrami et al., J Am Soc Mass Spectrom 2009, 20, 2124-2134) deals with theoretical calculations and laboratory experiments to investigate the effect of different esterifications on the average mass defect distribution of peptide fragment ions and on the specificity of two positive phosphoryl ions.

[0010] The paper "Improving Mass Defect Filters for Human Proteins" (Melinda L. Toumi and Heather Desaire, Journal of Proteome Research 2010, 9, 5492-5495) presents calculations of mass defect values ​​for two sets of human tryptic peptides and the comparison of the data with theoretical peptide mass defect calculations.

[0011] The publication "Improved Mass Defect Model for Theoretical Tryptic Peptides" (Indranil Mitra et al., Anal. Chem. 2012, 84, 3026-3032) proposes a specific model for peptide mass defects and peak widths that can improve peptide identification by filtering out chemical noise in mass spectral data.

[0012] Patent publication US 2012 / 0232805 A1 discloses a computerized method and apparatus for enumerating one or more amino acid compositions for all peptides having a length less than or equal to the maximum length and a mass less than or equal to the mass limit, comprising one or more processors, a data store communicatively coupled to the one or more processors, and a user interface communicatively coupled to the one or more processors.

[0013] The publication "Mass defect filter technique and its applications to drug metabolite identification by high-resolution mass spectrometry" (Haiying Zhang et al., J. Mass. Spectrom. 2009, 44, 999-1016) provides an overview of concepts of mass defect filtering algorithms, mass defect patterns of different classes of drug metabolites, and the design of filter templates for detecting these metabolites.

[0014] The object of the present invention is to provide a method for quality control of mass spectrometry data (in the analysis of peptides from biological samples) or a mass spectrometric method with a corresponding control or a mass spectrometric method with a signal correction.

[0015] To achieve this objective, the method has the features of claim 1. The basis is therefore a method for evaluating data from a mass spectrometry for the analysis of peptides from biological samples, in particular a MALDI-TOF mass spectrometry, comprising the following steps: a) Provision of expected mass defects; b) Determination of measured mass defects, namely the mass defects resulting from the mass spectrometry data; c) Comparison of the measured mass defects with the expected mass defects.

[0016] Depending on the extent of the deviation of the measured mass defects from the expected mass defects, the data or an underlying measurement can be evaluated as erroneous or acceptable. A suitable signal source, such as a display on a screen, can also output a signal corresponding to the evaluation of the data.

[0017] If the data or measurement is deemed acceptable, further processing will be carried out and / or further measurements will be performed. Data deemed to be faulty may, for example, be discarded for further processing and / or the device used to perform the mass spectrometry will be reviewed.

[0018] The method is based, among other things, on a computer-aided visualization of the mass defects of the peaks detected in a spectrum. This exploits the fact that many of the peaks originate from peptides whose mass defects follow a characteristic pattern. By comparing the measured mass defects with a theoretically expected pattern of mass defects, conclusions about the quality of the measured data can be drawn both computationally and visually.

[0019] The mass defects are, of course, not measured directly in the mass spectrometric method, but rather calculated from the measured masses (which are determined in a TOF mass spectrometer from measured times of flight). For simplicity, however, the term "measured mass defect" is used to distinguish it from the "expected mass defect." The latter is calculated based on the specific properties of the peptides.

[0020] In a further development of the invention, it is provided that the expected mass defects are calculated from m N r p , where m N is the nominal mass of a peptide and r p is preferably between 10 -3< and 10 -4<, in particular about 4.95 × 10 -4<.

[0021] For the calculation of the expected mass defects, for a given (integer) nominal mass m N is the difference to the mass resulting from the product with 1+rpm taken into account. r The preferred factor for p is 4.95×10 -4< . Other values ​​are also possible.

[0022] According to a further idea of ​​the invention, it is provided that the mass defect leads to a measured mass m is calculated from m − floor m 1 + r p + 0,5 , where the function floor(x) for any x > 0 the integer part of x designated.

[0023] The measured mass defect is therefore the difference between m and the nominal mass m N, whose corresponding expected peptide mass m p = (1+ r p ) m N closest to m lies.

[0024] In a further development of the invention it is provided that the discrepancy δ P between the measured and the expected mass defect is calculated directly from the measured mass m as δ P m = φ m 1 + r P + 0,5 − 0,5 where the function ϕ(x) = x-floor(x) is the fractional part of x for any x > 0.

[0025] According to a further idea of ​​the invention, it is provided that in order to compare the measured mass defects with the expected mass defects over subintervals of a mass axis, the median of the measured mass defects is formed and compared with the expected mass defect.

[0026] In a further development of the invention, the measured values ​​(of the determined masses) are corrected upon further use of the data, namely depending on the deviation of the measured mass defects from the expected mass defects. In the simplest case, the determined masses are corrected by the differences in the mass defects.

[0027] According to a further aspect of the invention, the measured mass defects are calculated for local maxima of the spectral intensities. Each local maximum is considered a peak of a specific measured mass.

[0028] A further method according to the invention has the features of claim 8. The basis is therefore a method for evaluating data from a mass spectrometry for the analysis of peptides from biological samples, in particular a MALDI-TOF mass spectrometry, preferably according to one of the previously discussed methods, with the following steps: a) Provision of expected mass defects; b) Determination of measured mass defects, namely the mass defects resulting from the mass spectrometry data; c) Determination of the discrepancies between the measured mass defects and the expected mass defects. d) Determination of the scatter of the discrepancies around their mean value; e) Comparison of the scatter with a defined permissible scatter.

[0029] Depending on the deviation of the variance from the defined permissible variance, the data or an underlying measurement can be evaluated as faulty or acceptable. A suitable signal source, such as a display on a screen, can also output a signal corresponding to the evaluation of the data.

[0030] If the data or measurement is deemed acceptable, further processing will be carried out and / or further measurements will be performed. Data deemed to be faulty may, for example, be discarded for further processing and / or the device used to perform the mass spectrometry will be reviewed.

[0031] In a further development of the invention, it is provided that, in order to determine a scatter of the mass defect discrepancies, interquartile ranges of the determined discrepancies are determined over subintervals of the mass axis, and that if a limit value of the scatter is exceeded, in particular the data are rejected as faulty.

[0032] According to a further idea of ​​the invention, a corridor for permissible scattering is formed by the limit values d P 1,2 m = ± μ v m where v m = σ P 2 m + Δ m m 2 12 , and σ P m N = σ 0 + s P m N , mit σ 0 ≈ 0,02 und s P ≈ 2,0 × 10 − 5 , and Δ m ( m) the width of m / z bins at the mass position m and m / z bins represent the intervals resulting from discretization of the mass axis, and µ > 0 indicates a scaling factor, preferably µ = 2, and in particular the data are rejected as erroneous if the scatter of the mass defect discrepancies lies outside the thus specified corridor.

[0033] In a further development of the invention, it is provided that the mass range with a detectable peptide signal is determined as the totality of all subintervals of the mass axis for which the quotient of the actual scattering and the maximum permissible scattering d P ( m ) a specified threshold t does not exceed, preferably t = 1.2. The lower and upper limits of this mass range indicate the extent of the peptide signal range.

[0034] Depending on whether a tolerance value for the upper limit or the extent of the peptide signal range is exceeded, or whether a tolerance value for the lower limit of the peptide signal range is exceeded, the data or an underlying measurement can be evaluated as faulty or acceptable. A suitable signal source can also output a signal corresponding to the evaluation of the data, such as a display on a screen.

[0035] If the data or measurement is deemed acceptable, further processing will be carried out and / or further measurements will be performed. Data deemed to be faulty may, for example, be discarded for further processing and / or the device used to perform the mass spectrometry will be reviewed.

[0036] The mass spectrometric method according to the invention has the features of claim 12. For the analysis of peptides from biological samples, in particular using a MALDI-TOF mass spectrometer, the following steps are provided: (a) performing one or more mass spectrometric analyses on the biological sample and providing data resulting from the mass spectrometric analyses; (b) performing one of the data analysis procedures as previously described.

[0037] A further solution to the problem mentioned above comprises the measures of claim 13. Accordingly, a method is provided with the following steps: a) measuring m / z values ​​for different peptides by means of a measuring apparatus, in particular using a MALDI-TOF mass spectrometer, b) assigning the measured m / z values ​​to corresponding m / z bins of a 2D histogram, c) plotting spectral intensities of the m / z bins in the 2D histogram, wherein the m / z bins are plotted on an abscissa axis and the discrepancy between measured and expected mass defect is plotted on an ordinate axis, d) wherein the diagram area spanned by the two axes is divided into a plurality, preferably 20 to 50, of rectangles, e) wherein the measured values ​​are interpolated up to an m / z resolution corresponding to the selected subdivision of the ordinate axis, and f) wherein for each rectangle those intensity values ​​of the interpolated spectrum are summed up whose mass defect discrepancy in the respective Subintervals of the axes fall,g) where different intensity values ​​are marked differently and the same intensity values ​​are marked the same.

[0038] Preferably, the invention can provide that for each sub-interval of the abscissa axis, in particular the horizontal mass axis, the corresponding intensity values ​​of the ordinate axis are statistically evaluated in order to determine cluster points and / or scatter values.

[0039] A further embodiment may provide that circular statistics are used to describe the distribution of the summed intensity values ​​in the vertical direction, in particular a first circular moment Z be used as (complex-valued) statistics.

[0040] In particular, it is conceivable that in order to determine a mass shift profile (vector of all circular moments Zfor all subintervals of the abscissa axis, especially the horizontal mass axis), the steps for forming the histogram and calculating the circular moments are combined and according to the formula Z k = 1 ∫ I k S ˜ t dt ∫ I k S ˜ t e iωt dt , mit ω = 2 π 1 + r P as Fourier integrals of the continuously interpolated spectrum S over the subintervals I k the abscissa axis.

[0041] A further solution to the problem mentioned above comprises the measures of claim 17. Accordingly, a mass spectrometric method for handling a single spectrum is provided, in which for a mass shift normalization to an ensemble for each spectrum a) a mass shift profile is determined, b) a common, average reference profile is formed from all individual mass shift profiles by element-wise calculation of the arithmetic mean, c) each spectrum is modified so that the mass shift profile of the modified spectrum corresponds to the reference profile,

[0042] In addition, it can be provided that for the normalization of an individual spectrum to the reference profile, relative shift values ​​are determined for each of the individual subintervals for which the mass shift profiles are calculated and are interpolated over the entire mass axis, whereby the measured values ​​of the spectrum are corrected by these interpolated shift values.

[0043] A MALDI-TOF mass spectrometer for solving the aforementioned problem is described in claim 19. Accordingly, this mass spectrometer has a control unit for carrying out the inventive method according to at least one of claims 1 to 18.

[0044] Further features of the invention will become apparent from the description and the claims. Advantageous embodiments of the invention are explained in more detail below with reference to the drawings. They show: Figure 1 a section of a mean spectrum with a clearly visible, wavy baseline with a characteristic wavelength of slightly more than 1 Da; Figure 2 a conventional mass defect diagram of a mean spectrum with a recognizable peptide band; Figure 3 a peptide mass defect diagram (PMD) of the mean spectrum to Figure 2, namely with representation of a mass shift versus mass (m / z), with horizontal reference line for an expected mass defect and reference corridor (dashed lines), mass defect discrepancy and its scatter (solid lines) as well as expected scatter (dash-dot-dash lines) and upper limit of the mass range with recognizable peptide signal (dotted line); figure a PMD of a mean spectrum with a strongly reduced signal-to-noise ratio, recognizable by the shortened peptide signal range, which only extends up to approximately 1300 Da; Figure 5 a PMD of a spectrum with a significant mass shift of approximately 0.15 to 0.35 Da; Figure 6 a PMD with a clearly recognizable line structure due to an equidistantly sampled mass axis; Figure 7 : a PMD of a mean spectrum with a very coarsely sampled mass axis; due to the coarse sampling, the structure of the peptide band is barely recognizable; Figure 8a PMD of a mean spectrum with incorrectly performed calibration, recognizable by the break in the peptide band from approx. m / z=1700; Figure 9 a peptide mass defect histogram (PMH) of a mean spectrum with mass shift comparable to that in Figure 5 shown; Figure 10 Mean spectra before (top) and after (bottom) mass shift normalization.

[0045] The molecules detected in a MALDI measurement primarily include metabolites and peptides. Metabolites are metabolic products and can have different chemical forms, such as lipids, carbohydrates, or degradation products of substances absorbed from food or the environment. Their masses are typically less than 1000 Da. Peptides, in contrast, are chains of amino acids with masses of up to 5000 Da and more.

[0046] All 23 amino acids found in proteins – and thus all peptides – consist of the five chemical elements carbon, hydrogen, oxygen, nitrogen, and sulfur. The relative ratio of these elements to each other is approximately the same for all peptides, regardless of their total mass, so that the mass defect of a peptide is essentially determined by its nominal mass. This results in a nearly linear

[0047] Relationship between the mass m of a peptide and its nominal mass m N: m ≈ m P m N = 1 + r P m N , mit r p ≈ 4,95 × 10 − 4 .

[0048] The scatter of the true masses around the theoretical mean m P is relatively small, its standard deviation can be σ P m N = σ 0 + s P m N , mit σ 0 ≈ 0,02 , s P ≈ 2,0 × 10 − 5 be estimated.

[0049] Due to the large number of different proteins and resulting peptides in biological tissue cells, a typical MALDI spectrum shows signal intensities at practically all m P (m N ) for a wide range of nominal masses m N. If a sum or average spectrum is formed from several spectra obtained during the measurement of a tissue sample, see Figure 1 , a characteristic, uniform wavy line can be seen as the baseline 20, which is particularly prominent above approximately 1000 Da and has a wavelength of approximately 1+ r P Da has.

[0050] Determining the mass of a molecule is subject to error, which essentially stems from two causes: First, the time of flight of a molecule can only be measured with a certain degree of accuracy and at discrete intervals, resulting in a discretization of the mass axis, i.e., a division into consecutive intervals (m / z bins). Typically, the width of the m / z bins is not constant but increases with higher masses.

[0051] Second, the flight time of the molecule depends not only on its mass but also on its initial state within the ion cloud at the beginning of the acceleration. This initial state, especially the speed and direction of motion of the molecule, is largely unknown and leads to a significant measurement error, which is usually corrected by calibration after the measurement.

[0052] Common calibration methods include external calibration and statistical peptide calibration. In external calibration, several drops of a solution containing defined components are placed next to the tissue sample prior to measurement. The measured spectra are compared with the expected masses of the known components after measurement, and a calibration curve is determined for the m / z axis of a spectrum. In peptide calibration, the relationship between the true mass of a peptide and its nominal mass described above is exploited to shift the peak positions presumably belonging to a peptide to the theoretically expected m / z values. See Wool A, Smilansky Z: Precalibration of matrix-assisted laser desorption / ionization-time of flight spectra for peptide mass fingerprinting. Proteomics 2002, 2, 1365-1373.

[0053] Both calibration methods cannot fully correct the errors in the m / z values. Furthermore, external calibration requires manual interaction, while peptide calibration is computationally and time-consuming.

[0054] Since this method compensates for mass errors globally for all spectra of a measurement, different errors in the spectra of a data set cannot be corrected this way. As an alternative, internal calibration is used, in which the calibration solution is distributed over the tissue sample to be measured, thus allowing individual correction for each spectrum of the measurement.

[0055] For practical reasons, a calibration solution can only contain a small number of known substances. This limits the number of sampling points from which the calibration curve is determined, and thus the accuracy of the calibration. Furthermore, this form of calibration requires manual user interaction.

[0056] In contrast, methods of statistical peptide calibration (see Wool A, Smilansky Z: Precalibration of matrix-assisted laser desorption / ionization-time of flight spectra for peptide mass fingerprinting. Proteomics 2002, 2, 1365-1373; Wolski WE, Lalowski M, Jungblut P, ​​and Reinert K. Calibration of mass spectrometric peptide mass fingerprint data without specific external or internal calibrants. BMC bioinformatics, 6(1):203, 2005) are fully automated and do not require a calibration solution. With these methods, the correction is achieved by comparing the masses measured in the tissue with a theoretical peptide mass model (see above) and a peptide database. These methods require prior peak picking, i.e., the identification of relevant peaks in a spectrum. They are very time-consuming and can lead to erroneous results due to incorrect assignment between peak and peptide database.

[0057] To visualize the mass defects observed in a spectrum, the m / z values ​​of the peaks found in a spectrum are plotted in a diagram whose horizontal axis corresponds to the mass m (or the m / z value), while on the vertical axis the fractional part m -floor( m ) is removed, see Figure 2Such diagrams are used to visualize different compositions of complex molecular mixtures. One variant is the so-called Kendrick mass defect diagram, which is used to characterize chemical compounds of a specific group; see Wikipedia: Kendrick mass. https: / / en.wikipedia.org / wiki / Kendrick_mass. In the context of peptide studies, mass defect diagrams can be used to differentiate between peptides and so-called glycopeptides; see US 2016 / 0003842 A1 and Froehlich J et al.: A Classifier Based on Accurate Mass Measurements to Aid Large Scale, Unbiased Glycoproteomics. Mol. Cell. Proteomics 2013, 12, 1017-1025.

[0058] In addition, peptides can be specifically chemically modified so that they have a mass defect that deviates significantly from the peptide mass model (also known as the averagine model). This deviation can be used to distinguish them from unmodified peptides. See Chen X, Savickas P, Vestal M. Methods and systems for mass defect filtering of mass spectrometry data. US Patent 7,634,364, filed 2006-06-23, granted 2009-12-15; Yao X, Diego P, Ramos AA, Shi Y. Averagine-scaling analysis and fragment ion mass defect labeling in peptide mass spectrometry. Anal. Chem. 2008 Oct 1;80(19):7383-91. doi: 10.1021 / ac801096e; Sleno L. The use of mass defect in modern mass spectrometry. J. Mass. Spectrom. 2012, 47: 226-236. doi:10.1002 / jms.2953

[0059] In this method of mass defect filtering, the mass defect determined for a spectral peak is used to further chemically characterize the corresponding molecule. A sufficiently high accuracy in mass determination is therefore required for this method.

[0060] For graphical illustration of mass defect filtering, a representation deviating from the usual mass scale is occasionally used, in which the deviation from the nearest mass corresponding to the averagine model is plotted in the vertical direction instead of the mass defect, see Yao X, Diego P, Ramos AA, Shi Y. Averagine-scaling analysis and fragment ion mass defect labeling in peptide mass spectrometry, Anal. Chem. 2008 Oct 1; 80(19):7383-91. doi: 10.1021 / ac801096e.

[0061] In their respective contexts, these diagrams serve only to illustrate the method using exemplary, synthetically calculated peptide masses. An application of this representation to actually measured data is unknown.

[0062] A diagram can be created from a spectrum averaged over several spectra from a MALDI measurement, which is referred to as a peptide mass defect diagram (PMD). This is done by determining a list of all local maxima and their respective m / z values, as well as the deviation from the nearest mass corresponding to the theoretical peptide mass model for each m / z value. Assuming that the measured signals are attributable to peptides, a peptide mass defect diagram is then calculated for each m / z value. m the nominal mass m N, for which the deviation between m and the mass expected according to the theoretical peptide mass model (Averagine model, see above) m P ( mN ) is minimized in absolute value (see below). The minimum deviation δ P ( m ), which can take values ​​from -0.5 to 0.5, is called the peptide model distance. The peptide model distance corresponds to the discrepancy between the measured and expected mass defect described above.

[0063] The positions of all local maxima are now plotted in a diagram whose horizontal axis corresponds to the mass or m / z value, and whose vertical axis represents the deviation from the peptide mass model determined above.

[0064] Compared to the known mass defect diagram, the PMD is generated by a transformation that maps the positions of the theoretically expected peptide mass defects to the reference line 22, which is a horizontal zero line. Furthermore, the PMD differs from known representations in that no prior specific signal analysis is performed, in particular no identification of significant peptide peaks (peak picking). Rather, the PMD essentially reflects statistical properties of the spectral background signal (cf. Fig. 1), which is merely assumed to originate in large proportions from peptide molecules. The vertical axis extends from -0.5 to 0.5 and represents the signed discrepancy between the mass defect expected for a peptide and the actually measured mass defect. In addition to the reference line 22, a reference corridor with lines 23, 24 can be drawn, which represents the expected scatter of the peptide mass defects around their mean value, taking into account the discretization of the mass axis given for a measurement.

[0065] The following quality characteristics of a spectrum can be easily read from a PMD: 1. Mass range with peptide signals: A clearly visible band ("peptide band 21") near reference line 22 indicates the presence of peptide signals in the respective mass range. Where the band structure disappears into an unstructured point cloud (typically visible at the upper end of the mass axis, to the right of an upper limit 30), the peptide signal disappears into the noise ( Figures 3 , 4 ). 2. Mass shift: If peptide band 21 deviates noticeably from reference line 22, this indicates a discrepancy between the true and measured molecular masses. The size of the discrepancy corresponds to the vertical offset between reference line 22 and center line 25 of peptide band 21. The reference corridor indicates the range in which a shift can be explained by the respective discretization of the mass axis ( Figures 3 , 5). 3. Equidistant or too coarsely sampled mass axis: A MALDI-TOF mass spectrometer typically does not discretize the mass axis equidistantly, but rather with increasing bin widths for higher masses. During post-processing of spectral data, the data are often resampled to an equidistant mass axis with lower resolution, resulting in a loss of accuracy. An equidistant mass axis is clearly recognizable in the PMD by a linear structure of the plotted points ( Figure 6 ). If the resolution of the data after resampling is too low, the structure of peptide band 21 ( Figures 6 , 7 ). 4. Incorrect calibration of the mass axis: Errors can occur during calibration of the mass axis, leading to a discontinuous distortion of the mass axis. Such distortions appear in the PMD as a break or offset in peptide band 21 ( Figure 8 ).

[0066] Compared to visualization using conventional mass defect diagrams, the above-mentioned quality characteristics of a spectrum are much more clearly recognizable in a PMD; in particular, even smaller mass shifts or those limited to partial sections of the mass axis can be more easily detected as deviations from the horizontal reference line 22.

[0067] Instead of a mean spectrum, a PMD can also be created for a single spectrum or for the maximum spectrum over several individual spectra (so-called skyline spectrum), but this representation is less meaningful.

[0068] In addition to pure visualization, the information presented in a PMD can also be quantitatively evaluated as follows (see also mathematical formulation below): 1. Determination of the discrepancy between the measured and expected peptide mass defect as a function of mass. For this purpose, the median of the mass defects is calculated over subintervals of the mass axis and compared with the expected value. 2. Determination of the dispersion of the mass defects around their mean. For this purpose, the interquartile range of the mass defects is determined over subintervals and converted into a specified multiple of the standard deviation of an assumed normal distribution. 3. Determination of the mass range with a detectable peptide signal. For this purpose, the dispersion of the mass defects determined from the data (area between lines 26, 27) is compared with the width of the reference corridor (area between lines 23, 24), and the range in which the deviation remains within a selected tolerance is determined.

[0069] This quantitative information can be displayed in the PMD, numerically displayed, or further processed to evaluate the quality of the measurement.

[0070] With S = s j m j j = 1 … n , mit n ∈ ℕ , 0 < m 1 < ⋯ < m n is a (single, average or skyline) spectrum consisting of the n Intensities s 1 ... sn to the m / z values m 1 ... mn . With floor x für x > 0 the integer part of a positive number x designated, with φ x = x − floor x für x > 0 the decimal part of a positive number x designated.

[0071] The PMD of the local maxima of S consists of the graphical representation of the points m i , δ P m i : s i > max 0 < j − i ≤ u s j where u ∈ ℕ denotes the radius of the local environment over which the local maxima are formed, and the function δ P m = φ m 1 + r P + 0,5 − 0,5 describes the signed discrepancy between the expected and the actually measured mass defect for a peptide. The above representation of the discrepancy δ P ( m ) is as follows: The theoretically expected mass defect of a peptide with nominal mass m N is m P − m N = 1 + r P m N − m N = r P m N .

[0072] To an actually measured mass m of a peptide, its nominal mass is defined as the integer mass m N is assumed for which the absolute difference m − 1 + r P m N This leads to m N = floor m 1 + r P + 0,5 .

[0073] The discrepancy δ P ( m ) results from the difference between measured and expected mass defect to δ P m = m − m N − m P − m N 1 + r P = m − 1 + r P m N 1 + r P = m 1 + r P − floor m 1 + r P + 0,5 = φ m 1 + r P + 0,5 − 0,5

[0074] The weighting of the difference of the mass defects with 1 / (1+ r P ) is used to normalize δ P ( m ) to the value range [-0.5 ... 0.5].

[0075] The reference line 22 of the theoretically expected mean mass defects of peptides is defined by the zero line δ P = 0. To determine the reference corridor (lines 23, 24), the expected variance v ( m ) of the positions of the local maxima as a function of the mass, which is determined by the sum of the variance of the true peptide masses σ 2< P and the variance resulting from the discretization of the mass axis can be estimated, v m = σ P 2 m + Δ m m 2 12 .

[0076] Where Δ m ( m ) the width of the m / z bins at the mass position m. The reference corridor is defined by boundary lines 23, 24 and 25 respectively. d P 1,2 m = ± μ v m formed, where the scaling factor µ > 0 indicates the width of the corridor as a multiple of a standard deviation (typically µ = 2).

[0077] To the spectrum S be a partitioning Ithe mass axis into pairwise disjoint intervals I k given: I = I k k = 1 , … , K mit K ∈ ℕ , U k I k = m 1 m n .

[0078] To a PMD in which the points m i , δ P m i i ∈ L mit L = i ∈ 1 … n : s i > max 0 < j − i ≤ u s j are shown, is used to determine the mass discrepancy E ( m ) for the subintervals I k the discrepancy E k = median δ P m i : i ∈ L ∩ I k formed. The E k are represented as points above the respective centers of the corresponding subintervals I k shown, with appropriate interpolation (e.g. linear) taking place between them.

[0079] The scatter e ( m ) of the mass defects is formed analogously from the interquartile ranges (IQR), e k = μ IQR δ P m i : i ∈ L ∩ I k 2 2 erf − 1 0,5 , where the scaling factor µ > 0 in turn indicates the width of the corridor as a multiple of a standard deviation, typically µ = 2, and erf denotes the Gaussian error function.

[0080] The mass range with detectable peptide signal is determined as that part of the mass axis for which the ratio between observed (lines 26, 27) and expected scatter (lines 28, 29) is below a specified tolerance threshold t remains: M P = m ∈ m 1 m n : e m μ v m ≤ t .

[0081] A typical tolerance value is t = 1,2. The positions of the outer edges of M P can be drawn as vertical lines in the PMD.

[0082] The previously described representation of a spectrum in a PMD can, in principle, be applied to both average and individual spectra. However, it requires the identification of local maxima in the spectrum in question and thus a sufficiently high signal-to-noise ratio, which is typically not present in individual spectra.

[0083] This disadvantage can be overcome by displaying the spectra in a peptide mass defect histogram (PMH). For this purpose, all spectral intensities for all m / z bins of a spectrum are plotted in a 2D histogram, where the horizontal axis corresponds to the mass axis and the vertical axis represents the peptide model distance to the respective mass (see below). Both axes are evenly divided into preselected numbers of subintervals (typically 20–50, but this can be different for each axes), resulting in a partitioning of the diagram area into rectangular tiles.

[0084] The spectrum under investigation is then interpolated to an m / z resolution corresponding to the selected subdivision of the mass defect axis. For each tile, all intensity values ​​of the interpolated spectrum whose masses and mass defects fall within the respective subintervals of the horizontal and vertical axes are then summed.

[0085] For graphical representation, all tiles can be visualized with a suitably selected grayscale or color scale according to the summed intensities. As in the PMD, the reference line 31 and the reference corridor 32, 33 are also drawn ( Fig. 9 ). The PMH can be interpreted in the same way as the PMD.

[0086] Analogous to the quantitative analysis of a PMD, characteristic quantities of a spectrum can also be calculated from a PMH—and thus also for individual spectra. For this purpose, an evaluation of the vertically arranged, summed intensity values ​​is performed for each subinterval of the horizontal mass axis to determine cluster points and scatter values.

[0087] It should be noted that the upper and lower boundary lines of a PMH, i.e., the points corresponding to the extreme distance values ​​+0.5 and -0.5, can be considered as identified with each other. Therefore, circular statistics are suitable for describing the distribution of the summed intensity values ​​in the vertical direction. In particular, the first circular moment Z be used as (complex-valued) statistics (mathematical formulation below). The circular moments Zfor all subintervals of the mass axis together are referred to here as the mass shift profile of the spectrum under consideration. The complex argument of Z corresponds (up to a factor of 2 π / (1+r P ) ) of the discrepancy between measured and expected masses. The amount of Z provides a reciprocal measure for the scatter of the measured peptide model distances: The value Z =0 corresponds to a maximum dispersion of all measurements over the interval [-0.5 .. 0.5], while in the extreme case a minimum dispersion, where all distance values ​​are identical, Z takes a value with magnitude 1.

[0088] For the actual calculation of the mass shift profile ZThe two steps of generating the 2D histogram and calculating the circular moments can be combined and expressed as Fourier integrals of the spectrum over the subintervals of the mass axis (see below). Numerically, these integrals can be approximated using suitable integration formulas (e.g., the trapezoidal formula or Simpson's rule). This also eliminates the need for finer discretization and interpolation of the spectrum, allowing calculations to be performed directly using the discrete spectral intensities at the original resolution.

[0089] The mass shift profile provides an estimate of the measurement errors of the measured masses in a spectrum compared to the true masses. In practice, it is often desirable to correct these shifts and thus achieve a higher accuracy of the measured masses in a spectrum.

[0090] On the other hand, the mass shift profile is obtained by comparing the measured data with the relatively simple, linear Averagine model (see above). Therefore, the estimation of mass errors using the mass shift profile cannot be more accurate than the accuracy of the model itself, which is not sufficiently high for many applications, at least in the lower mass range up to approximately 1000 Da.

[0091] Correcting the measured masses by the estimated measurement errors could therefore lead to parts of the measurement becoming less accurate.

[0092] For many applications, however, absolute mass accuracy is not crucial, but rather the best possible comparability between individual spectra from one and the same or multiple measurements. In these cases, the absolute measurement error of the measured masses of a spectrum is less relevant than the differences in measurement errors within an ensemble of spectra.

[0093] The mass shift normalization procedure consists of first determining the respective mass shift profile for each spectrum in an ensemble (see above), then creating a common, average reference profile from all individual mass shift profiles, and finally modifying each spectrum so that the mass shift profile of the modified spectrum corresponds to the reference profile. This adjustment reduces the relative deviation between the signal peaks belonging to one and the same peptide in the individual spectra and increases the comparability of the spectra ( Fig. 10 , below).

[0094] The reference profile is determined by calculating the arithmetic mean for each element (see below). To normalize a single spectrum to the reference profile, relative shift values ​​are determined for each subinterval for which the mass shift profiles were calculated and interpolated across the entire mass axis. The measured mass values ​​of the spectrum are then corrected by these interpolated shift values.

[0095] By applying these shifts, each individual spectrum receives its own mass axis. For a joint evaluation of an ensemble of spectra, it is usually desirable for all spectra to be defined along a common mass axis. This can be achieved by forming a common mass axis (e.g., by averaging across all individual mass axes or by selecting an arbitrary mass axis as the reference mass axis) and then interpolating each normalized spectrum to the common mass axis. Peptide mass defect histogram:

[0096] With S = s j m j j = 1 … n , mit n ∈ ℕ , 0 < m 1 < ⋯ < m n

[0097] As above, a (single, average or skyline) spectrum is defined, consisting of the n Intensities s 1 ... sn to the m / z values m 1 ... mn .

[0098] For a spectrum S, as above, a partitioning Ithe mass axis and a further partitioning J of the interval [-0.5 ... 0.5], I = I k k = 1 , … , K mit K ∈ ℕ , ∪ k I k = m 1 m n , J = J l l = 1 , … , L mit L ∈ ℕ , ∪ l J l = − 1 2 , 1 2 .

[0099] For the partitions I and J, Γ k , l = m ∈ I k : δ P m ∈ J l a finer partitioning of the mass axis, in which the subintervals Γ k,l are assigned to the individual tiles of the 2D histogram. Furthermore, an interpolant of the spectrum S is defined by a continuous function S ( m ) given, S ˜ : m 1 m n → ℝ , mit S ˜ m j = s j , j = 1 … n .

[0100] The Matrix H ( S ) = ( hk,l ) be through h k , l = 1 + r P ∫ I k S ˜ t dt ∫ Γ k , l S ˜ t dt für k = 1 … K , l = 1 … L For the numerical calculation of the hk,l can e.g. S as a linear interpolator of S The integrals can then be evaluated exactly. To form the PMH, the matrix H ( S ) as a grayscale or false color image. Mass shift profile:

[0101] With the notations and definitions of the previous section, the first circular moment of the columns of H given by the complex-valued quantities Z k = ∑ l = 1 L h k , l e 2 πi δ ¯ l , where δ 1 the mean values ​​of the intervals J l In the limit of an infinitesimally fine partitioning J (ie L → ∞) the Z k through Z k = 1 ∫ I k S ˜ t dt ∫ I k S ˜ t e iωt dt , mit ω = 2 π 1 + r P For the concrete numerical calculation of these integrals, a suitable integration formula (e.g. trapezoidal formula or Simpson's rule) and as support points of S just the discrete measuring points ( mj ) j= 1... n be elected. Because S ( mj ) = sj There is no need to change the spectrum S to interpolate explicitly. Mass shift normalization:

[0102] An ensemble of N spectra S i< ( i = 1 ... N) is given, which have a common mass axis ( mj ) j =1... n Furthermore, a partitioning I the mass axis in K Subintervals as given above. The partitioning for the individual spectra S i< calculated mass shift profiles are Z i< designated: Z i = Z k i k = 1 … K für i = 1 … N

[0103] About the mass shift profiles Z i< the reference profile Z calculated by element-wise arithmetic averaging: Z ¯ = Z ¯ k k = 1 … K mit Z ¯ k = 1 N ∑ i = 1 N Z k i .

[0104] Be now S = ( sj ,mj ) i =1... n any spectrum defined over the same mass axis with mass shift profile Z = ( Z k ) k=1 ...K . For each subinterval of the partitioning I a relative shift is now Δ k = 1 2 π arg Z k − arg Z ¯ k certainly, where arg( z ) ∈ (- π,π] denotes the complex argument function. The individual shifts Δ k are assigned to the midpoints of the subintervals I k and interpolated over the entire mass axis (typically using linear interpolation). This creates a displacement vector Δ ∗ = Δ j ∗ j = 1 … n The normalized spectrum S * is calculated by applying the shift values ​​to the m / z values ​​of the spectrum S receive, S ∗ = s j m j ∗ j = 1 … n , mit m j ∗ − m j + Δ j ∗ .

Claims

1. Control unit-implemented method to evaluate data of a mass spectrometry, particularly a MALDI-TOF mass spectrometry, for the analysis of peptides from biological samples, comprising the following steps: a) providing expected mass defects; b) determining measured mass defects, i.e., the mass defects resulting from the data of the mass spectrometry; c) comparing the measured mass defects with the expected mass defects, wherein a peptide mass defect diagram (PMD) is created from an average spectrum formed over several spectra of a MALDI measurement, in that a list of all local maxima and their respective m / z values as well as, for each m / z value, the deviation from the respective nearest mass corresponding to the theoretical peptide mass model is determined, wherein, assuming that the measured signals can be traced back to peptides, for every m / z value m the nominal mass mN is determined, for which the modulus of the deviation between m and the mass expected according to the theoretical peptide mass model mP(mN) is minimized, wherein the minimum deviation δP(m), which is called peptide model distance, can assume values from -0.5 to 0.5, and further wherein the positions of all local maxima are entered into a diagram whose horizontal axis corresponds to the mass or the m / z value, and on whose vertical axis the deviation from the peptide mass model determined above is plotted.

2. Method according to Claim 1, characterized in that the expected mass defects are calculated from m N r p , where mN designates the nominal mass of a peptide and rp amounts preferably to between 10-3 and 10-4, in particular, about 4.95 × 10-4.

3. Method according to Claim 1 or 2, characterized in that the mass defect for a measured mass m is calculated from m − floor m 1 + r p + 0.5 , where the function floor(x) for an arbitrary x > 0 designates the integer part of x.

4. Method according to Claim 1 or any one of the further Claims, characterized in that the discrepancy δP between the measured and the expected mass defect is calculated directly from the measured mass m as δ p m = φ m 1 + r p + 0.5 − 0.5 where the function ϕ(x) = x-floor(x) designates the decimal places of x for any x > 0.

5. Method according to Claim 1 or any one of the further Claims, characterized in that the median of the measured mass defects is formed and compared with the expected mass defect in order to compare the measured mass defects with the expected mass defects over partial intervals of a mass axis.

6. Method according to Claim 1 or any one of the further Claims, characterized in that the measured values are corrected when the data are to be used further, i.e., depending on the deviation of the measured mass defects from the expected mass defects.

7. Method according to Claim 1 or any one of the further Claims, characterized in that the measured mass defects are calculated for local maxima of the spectral intensities.

8. Control unit-implemented method to evaluate data of a mass spectrometry, particularly a MALDI-TOF mass spectrometry, for the analysis of peptides from biological samples, preferably according to any one of the preceding Claims, comprising the following steps: a) providing expected mass defects; b) determining measured mass defects, i.e., the mass defects resulting from the data of the mass spectrometry; c) determining the discrepancies between the measured mass defects and the expected mass defects; d) determining the variance of the discrepancies about their mean; e) comparing the variance with a defined acceptable variance, wherein a peptide mass defect diagram (PMD) is created from an average spectrum formed over several spectra of a MALDI measurement, in that a list of all local maxima and their respective m / z values as well as, for each m / z value, the deviation from the respective nearest mass corresponding to the theoretical peptide mass model is determined, wherein, assuming that the measured signals can be traced back to peptides, for every m / z value m the nominal mass mN is determined, for which the modulus of the deviation between m and the mass expected according to the theoretical peptide mass model mP(mN) is minimized, wherein the minimum deviation δP(m), which is called peptide model distance, can assume values from -0.5 to 0.5, and further wherein the positions of all local maxima are entered into a diagram whose horizontal axis corresponds to the mass or the m / z value, and on whose vertical axis the deviation from the peptide mass model determined above is plotted.

9. Method according to Claim 8, characterized in that interquartile ranges of the ascertained discrepancies are determined over partial intervals of the mass axis in order to determine a variance of the mass defect discrepancies, and that the data are rejected as defective particularly when a limit value of the variance is exceeded.

10. Method according to Claim 8 or 9, characterized in that an interval for acceptable variances is formed by limit values d p 1,2 m = ± μ v m where v m = σ P 2 m + Δ m m 2 12 and σ P m N = σ 0 + s P m N , where σ 0 ≈ 0.02 and s P ≈ 2.0 × 10 − 5 , and Δm(m) designates the width of m / z bins at the mass position m, and m / z bins represent the intervals resulting from the discretization of the mass axis, and µ > 0 provides a scaling factor, preferably µ = 2, and that, in particular, the data are rejected as defective when the variance of the mass defect discrepancies is outside the interval thus specified.

11. Method according to Claim 8 or any one of the further Claims, characterized in that the mass range with recognizable peptide signal is determined as the totality of all partial intervals of the mass axis for which the quotient of the actual variance and the maximum permissible variance dP(m) does not exceed a specified threshold value t, preferably setting t = 1.2, and that the data are rejected as defective particularly when a limit value for the lower or upper limit of this mass range or its extent is exceeded or not reached.

12. Mass spectrometric method for the analysis of peptides from biological samples, particularly using a MALDI TOF mass spectrometer, comprising the following steps: a) carrying out one or more mass spectrometric analyses on the biological sample and providing data which result from the mass spectrometric analyses; b) carrying out the method according to any one of the preceding Claims.

13. Mass spectrometric method for handling an individual spectrum of peptides from biological samples, comprising the following steps: a) measuring m / z values for different peptides using a measuring device, particularly by using a MALDI-TOF mass spectrometer, b) assigning the measured m / z values to corresponding m / z bins of a 2D histogram, c) plotting spectral intensities of the m / z bins in the 2D histogram, where the m / z bins are plotted on an abscissa axis and the discrepancy between measured and expected mass defect is plotted on an ordinate axis, d) where the diagram area defined by the two axes is subdivided into a plurality, preferably 20 to 50, of rectangles, e) where the measured values are interpolated to an m / z resolution which corresponds to the selected subdivision of the ordinate axis, and f) where, for each rectangle, those intensity values of the interpolated spectrum are summed, whose mass defect discrepancy falls within the respective partial intervals of the axes, g) where different intensity values are labelled differently and equal intensities values are labelled equally.

14. Mass spectrometric method according to Claim 13, characterized in that for each partial interval of the abscissa axis, especially the horizontal mass axis, the corresponding intensity values of the ordinate axis are statistically evaluated in order to determine cluster points and / or variance values.

15. Mass spectrometric method according to Claim 14, characterized in that circular statistics are used to describe the distribution of the intensity values in vertical direction, in particular a first circular moment Z can be used as (complex-valued) statistics.

16. Mass spectrometric method according to Claim 15, characterized in that the steps for the formation of the histogram and the calculation of the circular moments are combined in order to determine a mass shift profile (vector of all circular moments Z for all partial intervals of the abscissa axis, especially the horizontal mass axis) and are expressed according to the equation Z k = 1 ∫ I k S ˜ t dt ∫ I k S ˜ t e iωt dt , with ω = 2 π 1 + r P as Fourier integrals of the continuously interpolated spectrum S̃ over the partial intervals Ik of the abscissa axis.

17. Mass spectrometric method to handle an individual spectrum of peptides from biological samples, especially according to any one of the Claims 13 to 16, characterized in that for a mass shift normalization for an ensemble, for each spectrum a) a mass shift profile is determined, b) a common average reference profile is formed from all the individual mass shift profiles by forming the arithmetic mean, element by element, c) each spectrum is modified so that the mass shift profile of the modified spectrum corresponds to the reference profile.

18. Mass spectrometric method according to Claim 17, characterized in that for the normalization of an individual spectrum to the reference profile, relative shift values are determined and interpolated over the complete mass axis, for each of the individual partial intervals for which the mass shift profiles are calculated, and the measured values of the spectrum are corrected by these interpolated shift values.

19. MALDI-TOF mass spectrometer with a control unit for the analysis of peptides from biological samples using a method according to any one of the Claims 1 to 18.