JP6090201B2 - Mass spectrum data processing apparatus and mass spectrum data processing method - Google Patents

Description

  The present invention relates to a mass spectrum data processing apparatus and a mass spectrum data processing method for performing processing on mass spectrum data obtained by extracting signal intensities corresponding to physical quantities at a plurality of sampling points at a predetermined sampling interval. .

  As a method for reducing noise components included in data, a filtering method using various filters is known. For example, in a method using a smoothing filter, noise components can be reduced by using each data point as a point of interest and performing smoothing using data around the point of interest.

  When a Gaussian filter is used as an example of the smoothing filter, each peripheral point is weighted based on a Gaussian function. At this time, by changing the parameter of the Gaussian filter, it is possible to adjust the weighting ratio for each peripheral point and perform smoothing.

  7A and 7B are diagrams illustrating an example of a result obtained by performing smoothing on mass spectrum data using a Gaussian filter. 7A and 7B are the results of smoothing the same mass spectrum data, but the results are different because the parameters of the Gaussian filter are changed.

  In the example of FIG. 7A, as a result of performing smoothing with parameters that can maintain the shape of the peak P of the sample component, the effect of reducing the noise component N is insufficient. On the other hand, in the example of FIG. 7B, the shape of the peak P of the sample component is deteriorated as a result of performing smoothing with parameters that emphasize reduction of the noise component N. As described above, there is a limit to obtain the effect of reducing the noise component N while avoiding the deterioration of the shape of the peak P of the sample component. Smoothing methods such as Gaussian filters are based on frequency analysis based on the assumption that the latter is selectively reduced using the frequency characteristics of the peak and noise, and the noise reduction effect is adjusted by manipulating parameter values. It is possible to do this, but it cannot exceed the theoretical limit of frequency analysis.

  Furthermore, it is not easy to adjust the noise reduction effect by manipulating the parameter value. In software that handles mass spectrum data, the spectrum data is displayed as a waveform on a graph with the physical quantity as the horizontal axis, while the number of sampling points is often input as a parameter value for the smoothing method. For this reason, it is difficult for the operator to estimate an appropriate parameter value by looking at an actual data waveform.

  As an alternative to the frequency analysis, there is a time-frequency analysis capable of capturing the time change at the same time as the frequency characteristics of data, and a wavelet transform is known as a representative mathematical transformation. There is also a report example of a method of reducing a noise component included in mass spectrum data (for example, see Non-Patent Document 1 below). After the mass spectrum data is decomposed into multi-resolution components using wavelet transform, the noise components are reduced and then reconstructed to obtain noise reduction result data.

  However, since time-frequency analysis using wavelet transform treats frequencies as discrete units of frequency levels, it is normal for the operator to understand the relationship between the displayed data waveform and the processing parameters. It becomes even more difficult than the case of chemical method.

Kevin R. Coombes, et. Al., "Improved peak detection and quantification of mass spectrometry data acquired from surface-enhanced laser desorption and ionization by denoising spectra with the undecimated discrete wavelet transform", Proteomics 2005, 5, 4107-4117

  Since the noise reduction method based on frequency analysis can only distinguish the peak and noise based on the difference in frequency characteristics, the principle limit of occurrence of peak degradation at the time of noise reduction is low.

  Also, determine the processing parameter values, such as the number of sampling points, while looking at the spectrum data displayed on the graph with the physical quantity as the horizontal axis and referring to the analysis condition index such as the resolution indicated as the physical quantity. Is difficult. Mass spectrum data is obtained by extracting signal intensities corresponding to physical quantities (for example, mass-to-charge ratio m / z) at a plurality of sampling points at a predetermined sampling interval. However, a characteristic that data characteristics partially change is obtained. Have. Specifically, since the sampling interval is often changed such that the sampling interval is small in the region where the physical quantity is small and the sampling interval is large in the region where the physical quantity is large, it becomes more difficult to determine the optimum parameter.

  Due to such characteristics of the mass spectrum data, there was a case where the noise component could not be reduced satisfactorily while suppressing the deterioration of the peak shape even if it was attempted to reduce the noise component of the mass spectrum data by the conventional method. .

  The present invention has been made in view of the above circumstances, and provides a mass spectrum data processing apparatus and a mass spectrum data processing method capable of reducing noise components while suppressing peak shape deterioration of mass spectrum data as much as possible. The purpose is to provide.

  The mass spectrum data processing apparatus according to the present invention performs noise reduction by data processing based on time-frequency analysis, which has a higher theoretical limit of occurrence of peak degradation during noise reduction. For this reason, the mass spectrum data processing apparatus includes an input receiving unit that receives data input, a decomposition processing unit that decomposes data into time-frequency components, a noise reduction processing unit that reduces noise components, and a mass from the reduction results. A reconstruction processing unit for reconstructing the spectrum.

  Further, the mass spectrum data processing apparatus according to the present invention can specify and input processing parameters such as physical quantities that are easy for an operator to understand. For this reason, the mass spectrum data processing apparatus calculates a constant used for noise reduction processing based on the width specification processing unit that can specify and input processing parameters that are easy for the operator to understand, such as physical quantities, and the input processing parameters. And a processing constant calculation unit.

  Thus, the mass spectrum data processing apparatus according to the present invention includes an input receiving unit, a decomposition processing unit, a width designation processing unit, a reduction processing constant calculation unit, a noise reduction processing unit, and a reconstruction processing unit. Prepare. The input receiving unit receives input of mass spectrum data obtained by extracting signal intensities corresponding to physical quantities at a plurality of sampling points at a predetermined sampling interval. The decomposition processing unit decomposes the mass spectrum data received by the input receiving unit into a plurality of frequency resolution multi-resolution components. The width designation processing unit designates a width regarded as a noise component in the mass spectrum data as a physical quantity. The reduction processing constant calculation unit converts the physical quantity specified by the width specification processing unit into the number of sampling points, and the plurality of frequency levels decomposed by the decomposition processing unit based on the number of sampling points Among the multi-resolution components, a constant for reducing the noise component in the multi-resolution component of the frequency level regarded as the noise reduction target is calculated. The noise reduction processing unit performs a process of reducing a noise component using a constant calculated by the reduction processing constant calculation unit with respect to a multi-resolution component of a frequency level regarded as a noise reduction target. The reconstruction processing unit reconstructs mass spectrum data based on the multi-resolution component subjected to the process of reducing the noise component and the remaining multi-resolution component.

  According to such a configuration, after the mass spectrum data is decomposed into a plurality of frequency level multi-resolution components and the noise level is reduced with respect to the frequency level multi-resolution components regarded as noise reduction targets. The mass spectrum data can be reconstructed based on the multi-resolution component subjected to the process of reducing the noise component and the remaining multi-resolution component. At this time, a width to be regarded as a noise component in the mass spectrum data is designated as a physical quantity, and a constant used when performing a process of reducing the noise component is calculated based on the number of sampling points converted from the width.

  As described above, instead of designating the number of sampling points in the mass spectrum data, it is possible to perform processing for designating a constant by designating the width of a physical quantity and reducing the noise component using the constant. As a result, even if the sampling interval in the mass spectrum data is partially changed, the constant can be calculated satisfactorily, and noise reduction processing based on time-frequency analysis is performed using the constant. Thus, it is possible to satisfactorily reduce the noise component while suppressing the peak shape deterioration of the mass spectrum data as much as possible.

  The reduction processing constant calculation unit may calculate a degree of reduction according to a frequency level. In this case, the noise reduction processing unit may reduce the noise component using the degree of reduction with respect to the multi-resolution component of the frequency level regarded as the noise reduction target.

  According to such a configuration, it is possible to reduce the noise component using a more powerful reduction degree with respect to the multi-resolution component of the frequency level that is highly likely to be a noise component. Since the noise component tends to have a relatively high frequency, the signal component (sample peak) of the mass spectrum data can be increased by reducing the noise component using a higher degree of reduction for the multi-resolution component having a higher frequency. The noise component can be satisfactorily reduced while being reliably held.

  The reduction processing constant calculation unit may include a reduction degree correction processing unit that corrects the reduction degree based on a signal intensity in a multi-resolution component of a frequency level that is regarded as a noise reduction target.

  According to such a configuration, it is possible to reduce the noise component by using a stronger reduction degree with respect to the multi-resolution component of the signal strength that is highly likely to be a noise component. Since noise components tend to be relatively small in intensity, the lower the multi-resolution components, the stronger the degree of reduction, so that the noise (mass peak) component of the mass spectrum data can be retained more reliably. Ingredients can be reduced even better. This makes it possible to selectively reduce the noise component by separating the signal component and the noise component that could not be separated by frequency analysis alone. This is a very high limit that takes advantage of the features of time-frequency analysis. A noise reduction method with performance can be realized.

  The reduction degree correction processing unit is based only on the signal strength in the multi-resolution component in the vicinity of the boundary with the frequency level that is not regarded as the noise reduction target among the multi-resolution components of the frequency level regarded as the noise reduction target. The degree of reduction may be corrected.

  According to such a configuration, when a multi-resolution component having a high frequency includes a sharp noise component having a narrow width and a large signal strength, the signal strength is high despite the noise component. It is possible to prevent the noise component from being sufficiently reduced due to the above. Therefore, a noise reduction method having a very high limit performance without side effects can be realized.

  The mass spectrum data processing apparatus may further include a reference width storage unit that associates physical quantities of a plurality of target sampling points in the mass spectrum data with a width regarded as a noise component in those physical quantities as a reference width. . In this case, the width designation processing unit may designate a width regarded as a noise component in an arbitrary physical quantity based on the reference width stored in the reference width storage unit.

  According to such a configuration, since the width regarded as a noise component in an arbitrary physical quantity can be specified based on the reference width, even if the characteristics of the mass spectrum data partially change, Appropriate width can be specified.

  The mass spectrum data processing method according to the present invention includes an input reception step, a decomposition processing step, a width designation processing step, a reduction processing constant calculation step, a noise reduction processing step, and a reconstruction processing step. In the input receiving step, input of mass spectrum data obtained by extracting signal intensities corresponding to physical quantities at a plurality of sampling points at a predetermined sampling interval is received. In the decomposition processing step, the mass spectrum data received in the input receiving step is decomposed into a plurality of frequency resolution multi-resolution components. In the width designation processing step, a width regarded as a noise component in the mass spectrum data is designated by a physical quantity. In the conversion processing step, the physical quantity specified in the width specification processing step is converted into the number of sampling points, and based on the number of sampling points, the multiple frequency levels decomposed in the decomposition processing step are multiplexed. Among the resolution components, a constant for reducing the noise component in the multi-resolution component of the frequency level regarded as the noise reduction target is calculated. In the noise reduction processing step, the noise component is reduced using the constant calculated in the reduction processing constant calculation step on the multi-resolution component of the frequency level regarded as the noise reduction target. In the reconstruction processing step, mass spectrum data is reconstructed based on the multi-resolution component subjected to the process of reducing the noise component and the remaining multi-resolution component.

  According to the present invention, even when the sampling interval in the mass spectrum data is partially changed, it is possible to satisfactorily calculate the constant used when performing the process of reducing the noise component. By performing the noise reduction processing based on the time-frequency analysis using, the noise component can be reduced satisfactorily while suppressing the peak shape deterioration of the mass spectrum data as much as possible.

It is a block diagram which shows the structural example of the mass spectrum data processing apparatus which concerns on one Embodiment of this invention. It is a figure which shows the aspect at the time of the mass spectrum data being decomposed | disassembled into a multi-resolution component, and has shown the mass spectrum data before decomposition | disassembly. It is a figure which shows the aspect at the time of mass spectrum data being decomposed | disassembled into a multi-resolution component, and has shown the multi-resolution component obtained by decomposition | disassembly. It is a figure which shows an example of the reduction degree according to each frequency level. It is a figure which shows an example of the aspect at the time of correct | amending a reduction degree based on signal strength. It is a figure which shows an example of the result of having performed the process which reduces a noise component with respect to mass spectrum data by the mass spectrum data processing apparatus which concerns on this embodiment. It is the flowchart which showed an example of the process by the control part at the time of performing the process which reduces a noise component with respect to mass spectrum data. It is a figure which shows an example of the result of having performed smoothing with respect to mass spectrum data using a Gaussian filter. It is a figure which shows an example of the result of having performed smoothing with respect to mass spectrum data using a Gaussian filter.

  FIG. 1 is a block diagram showing a configuration example of a mass spectrum data processing apparatus according to an embodiment of the present invention. This mass spectrum data processing apparatus is configured by, for example, a personal computer, and includes a control unit 1, a storage unit 2, a display unit 3, an operation unit 4, and the like.

  The control unit 1 includes a CPU (Central Processing Unit), and when the CPU executes a program, the input reception unit 11, a decomposition processing unit 12, a width designation processing unit 13, a reduction processing constant calculation unit 14, a noise reduction process Functions as the unit 16 and the reconstruction processing unit 17. The storage unit 2 is configured by, for example, a hard disk or a RAM (Random Access Memory), and is assigned a mass spectrum data storage unit 21, a multi-resolution component storage unit 22, a reference width storage unit 23, and the like.

  The display unit 3 is configured by a liquid crystal display, for example, and can display mass spectrum data, an operation screen, and the like. The operation unit 4 is configured by a keyboard and a mouse, for example, and is operated when performing settings related to the operation of the mass spectrum data processing apparatus, such as an operation on an operation screen displayed on the display unit 3.

  This mass spectrum data processing apparatus performs a process for reducing noise components on mass spectrum data obtained by performing mass analysis of a sample with a mass spectrometer, for example. The mass spectrum data is obtained by extracting signal intensities corresponding to physical quantities (for example, mass to charge ratio m / z or mass m) at a plurality of sampling points at a predetermined sampling interval during mass analysis.

  In the present embodiment, a configuration will be described in which the mass spectrum data obtained as described above is stored in the mass spectrum data storage unit 21, and the mass spectrum data is read and processed. However, the configuration is not limited to such a configuration, and the mass spectrum data obtained by mass spectrometry may be input via another storage medium, or may be directly input from the mass spectrometer. It may be a simple configuration.

  The input receiving unit 11 receives an input of mass spectrum data stored in the mass spectrum data storage unit 21. That is, when performing processing on the mass spectrum data, the mass spectrum data to be processed is received by the input receiving unit 11, and the processing described below is performed on the mass spectrum data.

  The decomposition processing unit 12 decomposes the mass spectrum data received by the input receiving unit 11 into a plurality of frequency level multi-resolution components (time-frequency components) by data processing based on time-frequency analysis. Such decomposition into multi-resolution components can be performed using, for example, wavelet transform. This wavelet transform is a technique for analyzing a waveform using a localized wave packet (wavelet) as a basis function.

  When a basis function designed mathematically as an orthogonal basis is used, a discrete wave that is decomposed into multiple resolution components while changing its wave packet width by a factor of 2, 4, 8, etc. By performing the let transform, the mass spectrum data can be completely reconstructed from the decomposed components. In this case, since the localized wave packet is a basis function, the position accuracy of the analysis is high, and even when a side effect occurs, it is limited to the local area, and it is characterized in that a multi-resolution analysis result can be obtained at one time. .

  For example, a Coiflet that is an orthogonal basis can be used as the basis function, but not limited to this, other basis functions such as Daubechies and Symlet may be used. Since a specific processing method for wavelet conversion is known, a detailed description thereof is omitted here.

  2A and 2B are diagrams showing an aspect when mass spectrum data is decomposed into multi-resolution components. FIG. 2A shows mass spectrum data before decomposition, and FIG. 2B shows multi-resolution components obtained by decomposition. Show. In this example, the case where the mass spectrum data shown in FIG. 2A is decomposed into multi-resolution components of seven frequency levels L1 to L7 as shown in FIG. 2B will be described, but the number of frequency levels to be decomposed is six or less. It may be 8 or more.

  In the multi-resolution components of the frequency levels L1 to L7 shown in FIG. 2B, the frequency level L1 has the highest frequency, and the frequency level L7 has the lowest frequency. Based on the assumption that the noise component is higher in frequency than the peak of the sample component, after performing processing to reduce the noise component only to the multi-resolution component at the frequency level on the high frequency side, mass spectral data is obtained from each multi-resolution component. By reconfiguring, the noise component can be reduced satisfactorily.

  Referring to FIG. 1 again, the width designation processing unit 13 designates a width that is regarded as a noise component in the mass spectrum data as a physical quantity. Specifically, mass spectrum data as shown in FIG. 2A is displayed on the display unit 3, and the operator who confirms the operation operates the operation unit 4, so that the waveform width W (Fig. 2A) Enter the physical quantity. As a result, the operator can input with an easily understandable processing parameter called a physical quantity, and the input physical quantity is designated as a width to be regarded as a noise component.

The reduction processing constant calculation unit 14 converts the physical quantity designated by the width designation processing unit 13 into the number of sampling points. Specifically, the number of sampling points can be calculated by the following equation (1). For example, when the operator operates the operation unit 4 to specify 0.4 (dalton) as the physical quantity (mass m) and the sample interval is 0.01 (dalton / pt), the following formula (1) Thus, the number of sampling points is calculated as 40 (pt).
Number of sampling points = physical quantity / sampling interval (1)

  However, in the present embodiment, the width designation processing unit 13 can also designate a width regarded as a noise component based on the reference width stored in the reference width storage unit 23. In this case, a width regarded as a noise component in an arbitrary physical quantity X can be designated (see FIG. 2A).

  The reference width storage unit 23 stores, as reference widths, widths regarded as noise components in the physical quantities in association with physical quantities of a plurality of sampling points of interest in the mass spectrum data. Such a reference width corresponding to each target sampling point can be set as a priori information when it is known in advance that the characteristic (for example, sampling interval) of the mass spectrum data changes at each target sampling point. .

For example, as shown in Table 1 below, a case where the widths regarded as noise components in the physical quantities corresponding to the physical quantities (mass m) of the three target sampling points are stored in the reference width storage unit 23 as reference widths will be described. To do. Such a reference width may be stored in advance in the reference width storage unit 23, or may be stored in the reference width storage unit 23 by an operator performing an input operation.

In this case, an arbitrary physical quantity between each sampling point of interest (physical quantity X _i , width W _i regarded as a noise component: i = 1, 2, 3) is regarded as a noise component in the physical quantity by the following equation (2). The width W can be specified. For example, when the physical quantity X is 800 (dalton), the a priori information (X _i = 700, X _{i + 1} = 1500, W _i = 0.2, W _{i + 1} = 0.5) obtained from Table 1 above is used. Thus, the width W regarded as a noise component in the physical quantity X can be specified as 0.2375 (dalton).
W = W _i + (X−X _i ) × (W _{i + 1} −W _i ) / (X _{i + 1} −X _i ) (2)

Thus, since the width W regarded as a noise component in an arbitrary physical quantity X can be specified based on the reference width _Wi , even if the characteristics of the mass spectrum data partially change, The assumed width W can be appropriately specified.

Further, the reduction processing constant calculation unit 14 is a frequency level regarded as a noise reduction target among the multi-resolution components of the plurality of frequency levels L1 to L7 decomposed by the decomposition processing unit 12 based on the converted number of sampling points. As an example of a constant for reducing the noise component in the multi-resolution component, a degree of reduction corresponding to the frequency level is calculated. Here, the frequency level can be expressed by the following equation (3).
Frequency level = log ₂ (number of sampling points per cycle) (3)

That is, the multi-resolution component at the frequency level L1 indicates a change in signal intensity when the number of sampling points per cycle is 2 ¹ = 2. The multi-resolution component at the frequency level L2 indicates a change in signal intensity when the number of sampling points per cycle is 2 ² = 4. Similarly, the multi-resolution components of the frequency levels L3, L4, L5, L6, and L7 indicate changes in signal strength when the number of sampling points per cycle is 8, 16, 32, 64, and 128, respectively.

  For example, when the converted number of sampling points is substituted into the above equation (3) and, as a result, the frequency level (boundary frequency level) regarded as a noise component is calculated as 3.5, the boundary frequency level is A multi-resolution component of a frequency level that is regarded as a noise reduction target is determined as a reference. In this case, for example, multi-resolution components of frequency levels L3 and L4 above and below the boundary frequency level 3.5 and multi-resolution components of frequency levels L1 and L2 higher than these are determined as noise reduction targets. .

  The noise reduction processing unit 16 performs a process of reducing the noise component with respect to the multi-resolution component of the frequency level regarded as the noise reduction target. In the present embodiment, the noise component is reduced by using the degree of reduction calculated by the reduction processing constant calculation unit 14. For example, when the multi-resolution components of the frequency levels L1 to L4 are regarded as noise reduction targets as described above, the degree of reduction corresponding to each frequency level L1 to L4 is used for those multi-resolution components. Noise component is reduced.

  The multi-resolution component that has been subjected to the process of reducing the noise component by the noise reduction processing unit 16 is input to the reconstruction processing unit 17 together with the remaining multi-resolution component that has not been subjected to the process of reducing the noise component. Mass spectrum data is reconstructed by the processing unit 17. The mass spectrum data reconstructed in this way is displayed on the display unit 3.

  FIG. 3 is a diagram illustrating an example of the degree of reduction corresponding to each frequency level. As shown in FIG. 3, the degree of reduction is closer to 0 as the frequency level is lower (higher frequency) and closer to 1 as the frequency level is higher (lower frequency). This corresponds to the general frequency characteristics of data in which noise components are distributed on the high frequency side and signal components are distributed on the low frequency side.

The degree of reduction determined in this way is substituted as the constant Damp (L) in the following equation (4), thereby reducing the noise component with respect to the multi-resolution component d (L) (X) at each frequency level. As a result, d (L) (X) ′ is obtained. Note that Damp (L) means the degree of reduction at the frequency level L.
d (L) (X) ′ = d (L) (X) × Damp (L) (4)

  As described above, in the present embodiment, mass spectrum data is decomposed into multi-resolution components of a plurality of frequency levels L1 to L7, and noise components are compared with multi-resolution components of frequency levels L1 to L4 that are regarded as noise reduction targets. After performing the process of reducing the noise, the mass spectrum data is reconstructed based on the multi-resolution component and the remaining multi-resolution component on which the process of reducing the noise component has been performed. At this time, a width W to be regarded as a noise component in the mass spectrum data is designated as a physical quantity, and a constant (degree of reduction) used when performing a process of reducing the noise component based on the number of sampling points converted from the width W. Is calculated.

  In this way, instead of specifying the number of sampling points in the mass spectrum data, it is possible to calculate the degree of reduction by designating the width W of the physical quantity, and to perform processing for reducing the noise component using the degree of reduction. . As a result, even if the sampling interval in the mass spectrum data is partially changed, the constant can be calculated satisfactorily, and noise reduction processing based on time-frequency analysis is performed using the constant. Thus, it is possible to satisfactorily reduce the noise component while suppressing the peak shape deterioration of the mass spectrum data as much as possible.

  Further, in the present embodiment, as shown in FIG. 3, a stronger reduction degree (which can reduce the noise component) is used for the multi-resolution components of the frequency levels L1 to L4 that are likely to be noise components. Noise components can be reduced. Since the noise component tends to have a relatively high frequency, the multi-resolution component having a higher frequency reduces the noise component using a stronger reduction degree (a reduction degree closer to 0), thereby reducing the mass spectrum data. It is possible to satisfactorily reduce the noise component while holding the signal (sample peak) component more reliably.

  As shown in FIG. 1, in this embodiment, the reduction processing constant calculation unit 14 includes a reduction degree correction processing unit 15 that corrects the reduction degree. The degree-of-reduction correction processing unit 15 corrects the degree of reduction based on the signal strength in the multi-resolution component of the frequency level considered as the noise reduction target. This is based on the assumption that the signal intensity of the noise component is smaller than the signal intensity at the peak of the sample component.

  FIG. 4 is a diagram illustrating an example of a mode for correcting the reduction degree based on the signal strength. The signal intensity (absolute amplitude value) in the multi-resolution component at each frequency level is compared with the first threshold value V1 and the second threshold value V2. The first threshold value V1 and the second threshold value V2 are preset values, and the second threshold value V2 is set to a value larger than the first threshold value V1.

  As a result of the comparison, the frequency level having a signal intensity lower than the first threshold value V1 is regarded as a noise component, and the reduction degree Damp (L) at the frequency level determined in the manner as shown in FIG. 3 is used as it is. The noise component is reduced. Further, a frequency level having a signal intensity greater than the second threshold value V2 is regarded as not a noise component, regardless of the value of the degree of reduction Damp (L) at the frequency level determined in the manner shown in FIG. The degree of reduction is corrected to 1.0. Note that a reduction degree of 1.0 is equivalent to the fact that the process of reducing the noise component is not performed on the multi-resolution component at that frequency level.

  When the signal intensity is between the first threshold value V1 and the second threshold value V2, the degree of reduction is corrected according to the signal intensity value. In this example, between the first threshold value V1 and the second threshold value V2, the reduction degree Damp (L) is set so as to be corrected to a value with a higher reduction degree as the signal intensity value increases from 1.0 to 1.0. ing.

  In the present embodiment, it is possible to reduce a noise component using a stronger reduction degree with respect to a multi-resolution component having a signal strength that is likely to be a noise component (a signal strength smaller than the first threshold value V1). . Since the noise component tends to have a relatively low signal strength, the signal (sample peak) component of the mass spectrum data can be more reliably retained by making the multiresolution component with a lower signal strength a stronger reduction degree. In addition, the noise component can be further satisfactorily reduced. This makes it possible to selectively reduce the noise component by separating the signal component and the noise component that could not be separated by frequency analysis alone. This is a very high limit that takes advantage of the features of time-frequency analysis. A noise reduction method with performance can be realized.

  However, when the degree of reduction is simply corrected in the manner shown in FIG. 4, a sharp multi-resolution component having a high frequency (for example, multi-resolution component of frequency levels L1 to L4) is narrow and has a high signal strength. When a noise component is included, the noise component may not be sufficiently reduced due to the high signal intensity even though the noise component is included. That is, due to the presence of a sharp noise component having a large amplitude, the signal intensity may become larger than the second threshold value V2, and the degree of reduction may be corrected to 1.0.

  Therefore, the reduction degree correction may not be performed for the specific frequency level L, and the reduction degree Damp (L) itself may be used regardless of the signal intensity. For example, among the multi-resolution components of the frequency levels LV1 to LV4 that are regarded as noise reduction targets, only the signal intensity in the multi-resolution components near the boundary with the frequency levels LV5 to LV7 that are not regarded as noise reduction targets, It is good also as a structure which correct | amends a reduction degree. In this case, for example, only the multi-resolution components of the upper and lower frequency levels L3 and L4 with respect to the boundary frequency level 3.5 are configured to correct the degree of reduction in the manner shown in FIG. As a result, it is possible to prevent a sharp noise component from being sufficiently reduced due to the above-described causes, and therefore, it is possible to realize a noise reduction method having a very high marginal performance without side effects.

  FIG. 5 is a diagram illustrating an example of a result of performing a process of reducing noise components on mass spectrum data by the mass spectrum data processing apparatus according to the present embodiment. As apparent from FIG. 5, according to this embodiment, the noise component N can be reduced more than the result of the conventional method shown in FIG. 7A, and more than the result of the conventional method shown in FIG. 7B. Deterioration of the shape of the peak P of the sample component can be avoided.

  FIG. 6 is a flowchart illustrating an example of processing performed by the control unit 1 when performing processing for reducing noise components on mass spectrum data. When an input of mass spectrum data is accepted (Yes in step S101: input accepting step), the mass spectrum data is first decomposed into multi-resolution components of a plurality of frequency levels (step S102: decomposition processing step).

  Thereafter, based on the input of the physical quantity by the operator (Yes in step S103), the inputted physical quantity is designated as a width regarded as a noise component (step S104: width designation processing step). The designated width is converted into the number of sampling points (step S105), and based on the converted number of sampling points, among the multiple resolution components of a plurality of frequency levels, the multi-resolution of the frequency level regarded as the noise reduction target Components are determined (step S106). Then, for example, the degree of reduction is calculated in the manner described with reference to FIGS. 3 and 4 (step S107). These steps S105 to S107 constitute a reduction processing constant calculation step, and may include a reduction degree correction processing step in which the reduction degree correction processing unit 15 corrects the reduction degree.

  A process for reducing the noise component is performed on the multi-resolution component of the frequency level regarded as the noise reduction target by using the degree of reduction calculated in the reduction processing constant calculation step (step S108: noise reduction processing step).

  After the process for reducing the noise component is performed in this way, the mass spectrum data is reconstructed based on the multi-resolution component for which the process has been performed and the remaining multi-resolution component (step S109: Configuration processing step). Then, the reconstructed mass spectrum data after the noise reduction processing is displayed on the display unit 3 in a manner as shown in FIG. 5, for example (step S110).

  In the above embodiment, the mass spectrum data processing apparatus in which the storage unit 2, the display unit 3, and the operation unit 4 are configured integrally with the control unit 1 has been described. However, the configuration is not limited to such a configuration, and a configuration in which at least one of the storage unit 2, the display unit 3, and the operation unit 4 is provided separately from the mass spectrum data processing apparatus may be used.

  The mass spectrum data processing device is not limited to a dedicated data processing device for processing mass spectrum data, and a data processing device composed of a general-purpose personal computer functions as the mass spectrum data processing device. May be.

  Furthermore, in the above-described embodiment, the configuration in which the mass spectrum data processing apparatus is provided separately from the mass spectrometer has been described. However, the mass spectrum data processing apparatus according to the present invention is configured integrally with the mass spectrometer. It may be.

DESCRIPTION OF SYMBOLS 1 Control part 2 Memory | storage part 3 Display part 4 Operation part 11 Input reception part 12 Decomposition processing part 13 Width designation | designated processing part 14 Reduction process constant calculation part 15 Reduction degree correction process part 16 Noise reduction process part 17 Reconstruction process part 21 Mass spectrum Data storage unit 22 Multi-resolution component storage unit 23 Reference width storage unit

Claims (6)

An input receiving unit that receives input of mass spectrum data obtained by extracting signal intensities corresponding to physical quantities at a plurality of sampling points at a predetermined sampling interval;
A decomposition processing unit that decomposes the mass spectrum data received by the input receiving unit into multi-resolution components of a plurality of frequency levels;
A width designation processing unit that designates, as a physical quantity, a width regarded as a noise component in mass spectrum data;
The physical quantity designated by the width designation processing unit is converted into the number of sampling points, and based on the number of sampling points, noise among the multiple resolution components of the plurality of frequency levels decomposed by the decomposition processing unit A reduction processing constant calculation unit for calculating a constant for reducing a noise component in a multi-resolution component of a frequency level regarded as a reduction target;
A noise reduction processing unit that performs processing for reducing a noise component using a constant calculated by the reduction processing constant calculation unit for a multi-resolution component of a frequency level regarded as a noise reduction target;
A mass spectrum data processing apparatus comprising: a reconstruction processing unit configured to reconstruct mass spectrum data based on a multi-resolution component subjected to a process for reducing a noise component and a remaining multi-resolution component. The reduction processing constant calculation unit calculates a reduction degree according to the frequency level,
2. The mass spectrum data processing according to claim 1, wherein the noise reduction processing unit reduces the noise component using the degree of reduction with respect to a multi-resolution component of a frequency level regarded as a noise reduction target. apparatus.   The reduction processing constant calculation unit includes a reduction degree correction processing unit that corrects the reduction degree based on a signal intensity in a multi-resolution component of a frequency level regarded as a noise reduction target. The mass spectrum data processing apparatus described.   The reduction degree correction processing unit is based only on the signal strength in the multi-resolution component in the vicinity of the boundary with the frequency level that is not regarded as the noise reduction target among the multi-resolution components of the frequency level regarded as the noise reduction target. The mass spectrum data processing apparatus according to claim 3, wherein the degree of reduction is corrected. A reference width storage unit that stores a width that is regarded as a noise component in these physical quantities as a reference width in association with physical quantities of a plurality of sampling points of interest in the mass spectrum data,
The width specification processing unit specifies a width to be regarded as a noise component in an arbitrary physical quantity based on a reference width stored in the reference width storage unit. Mass spectrum data processing device. An input receiving step for receiving an input of mass spectrum data obtained by extracting signal intensities corresponding to physical quantities at a plurality of sampling points at a predetermined sampling interval;
A decomposing step of decomposing the mass spectrum data received in the input receiving step into multi-resolution components of a plurality of frequency levels;
A width designation processing step for designating a width regarded as a noise component in mass spectrum data by a physical quantity;
The physical quantity designated in the width designation processing step is converted into the number of the sampling points, and based on the number of sampling points, among the multi-resolution components of the plurality of frequency levels decomposed in the decomposition processing step, noise A reduction processing constant calculation step for calculating a constant for reducing a noise component in a multi-resolution component of a frequency level regarded as a reduction target;
A noise reduction processing step for performing processing for reducing the noise component using the constant calculated in the reduction processing constant calculation step for the multi-resolution component of the frequency level regarded as the noise reduction target;
A mass spectrum data processing method comprising: a reconstruction processing step for reconstructing mass spectrum data based on a multi-resolution component subjected to a process for reducing a noise component and a remaining multi-resolution component.

Images