JP2005214799A - Analyzer, method, and program for analyzing data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for selecting a specified peak in a plurality of elements, in analysis of an analytical data. <P>SOLUTION: This data analyzer 10 for selecting the specified peak resulting from the specified element out of the plurality of peaks, based on the analytical data of an objective sample containing the plurality of elements has a means for acquiring the analytical data of the objective sample, a means for acquiring an analytical data of a reference sample, a classifying means for classifying the marked peaks into the peaks having a preset threshold value or more of intensity and the peaks having intensity less than the preset threshold value, respectively, as to the analytical data of the objective sample and the analytical data of a reference sample, a calculation means for calculating mutual information volumes about the intensities of the marked peaks before and after the classification, and a selection means for selecting the specified peak, based on the respective mutual information volumes of the plurality of elements. A value of an average value or more determined based on the average value of the peak intensities of the marked peaks in the analytical data of the reference sample is set as the threshold value, in the classifying means. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a data analysis apparatus, a data analysis method, and a data analysis program for analyzing analysis data obtained by analyzing a target sample including a plurality of elements.

  In the field of natural science, social science, etc., when multiple peaks are mixed in the data obtained by analyzing a sample containing multiple elements, how to select specific peaks due to specific elements This is an important issue. In general, each element in a plurality of elements is used as a reference sample, and a specific peak due to a specific element is selected by comparison with a peak of analysis data obtained by analyzing the reference sample.

  However, when complex data analysis is required, such as proteins and polymer materials in scientific analysis, it may be difficult to select a specific peak by comparison with a simple reference sample.

  For example, in time-of-flight secondary ion mass spectrometry (TOF-SIMS), by scanning the surface of a material, the distribution of the distribution is based on a specific molecular ion peak for a specific chemical species distributed on the surface of the material. Imaging is possible. Even in the case of a macromolecule such as a protein or a polymer material, if a molecular ion peak specific to the macromolecule can be determined, the distribution can be imaged. However, in current ionization systems, such macromolecules are mostly fragment ions. The improvement of the primary ion source has enabled relatively large fragment ionization, but in the case of substances with similar structures such as proteins and polymer materials, the complex spectrum is composed of many fragment ions. Appropriate analysis methods are required to select specific fragment ions. In particular, in the measurement of a new substance, it is difficult to predict all the peaks with high specificity, and therefore, a technique capable of evaluating all the peaks existing in the hundreds to thousands is necessary.

  As an analysis method suitable for selection of specific fragment ions, it is known to analyze analysis data by multivariate analysis including principal component analysis (PCA) and discriminant analysis (LDA). However, in data analysis using multivariate analysis, especially principal component analysis and linear discriminant analysis undergo matrix calculation in the analysis process, in order to obtain accurate results, the number of peaks used for analysis is more than the number of data. It is desirable to be small. In particular, when measuring a new sample, it is difficult to predict fragment ions caused by the chemical species to be examined, so that principal component analysis and discriminant analysis are effectively performed for all possible peaks. Requires a large amount of data or repeated processing.

  As another analysis method suitable for selecting specific fragment ions, for example, there is an analysis method based on mutual information as shown in Non-Patent Document 1. This is a method for objectively quantifying the importance of each peak appearing in measurement data using an information entropy change amount (mutual information amount) generated by examining the intensity of a certain peak. Specifically, the mutual information amount before and after comparing the intensity of each fragment ion peak between the sample and the reference sample is calculated, and the possibility of being attributed to a specific substance (macromolecule) is quantified. At this time, a value that maximizes the mutual information amount is used as the threshold value in the comparison of peak intensities. Qualitative, quantitative, and distribution imaging of macromolecules determined from the results is possible.

S. Aoyagi, M. Hayama, U. Hasegawa, K. Sakai, M. Tozu, T. Hoshi and M. Kudo, e-Journal of Surface Science and Nanotechnology, 1, 67-71 (2003).

  However, the analysis method using the mutual information calculated by using the value that maximizes the mutual information as a threshold as shown in Non-Patent Document 1 can identify two types of sample groups from the peak intensity. Only. That is, since only the level of the intensity of the peak of each sample with respect to a certain threshold is used as a criterion for judgment, even if the mutual information amount is large even if the separation can be performed as shown in FIG. Higher may not be suitable for quantitative and distribution imaging. In other words, it is necessary to select again a fragment ion peak selected from the mutual information value with sufficiently low background intensity.

  The present invention is specified from a plurality of peaks included in analysis data obtained by analyzing a target sample including a plurality of elements in a simpler process with a minimum number of data sufficient for statistical science. A data analysis apparatus, a data analysis method, and a data analysis program capable of selecting a peak due to the element.

  The present invention is based on target sample analysis data obtained by analyzing a target sample including a plurality of elements, and is caused by a specific element among the plurality of elements from a plurality of peaks included in the target sample analysis data. A data analysis apparatus for selecting a peak to be analyzed, comprising: means for acquiring the target sample analysis data; and reference sample analysis data obtained by analyzing a reference sample from which at least one of the plurality of elements is removed For each of the acquisition means, the target sample analysis data, and the reference sample analysis data, a certain peak of interest is classified into a peak having an intensity greater than or equal to a preset threshold and a peak having an intensity less than the threshold. The mutual information regarding the intensity of the peak of interest before and after the classification is calculated. And calculating means for selecting a specific peak derived from the specific element based on the mutual information obtained for each of the plurality of elements, and in the classification means, the reference sample analysis data A value equal to or higher than the average value determined based on the average value of the peak intensity of interest in is set as the threshold value.

  In the data analysis apparatus, when the mutual information amount calculated by the calculating means is less than a preset reference mutual information amount, the classification means and the calculating means until the reference mutual information amount is equal to or greater than the reference mutual information amount. Each is preferably classified and calculated.

  In the data analysis apparatus, it is preferable that a threshold value that gives a maximum mutual information amount in the calculation unit is set as a threshold value in the classification unit.

  Further, in the data analysis device, when the mutual information amount calculated by the calculating means does not reach the reference mutual information amount, the mutual information amount before and after the classification based on the threshold value that gives the maximum mutual information amount Thus, it is preferable to select the specific peak as the specific peak.

  In the data analysis apparatus, it is preferable that the selecting unit selects a peak having a large mutual information value as the specific peak.

  In the data analysis apparatus, it is preferable that the target sample analysis data and the reference sample analysis data are analyzed a plurality of times for each of the target sample and the reference sample.

  The data analysis apparatus preferably further includes means for imaging the distribution of the specific element in the target sample based on the specific peak.

  The data analysis apparatus preferably further includes means for quantifying the specific element in the target sample based on the specific peak.

  In the data analysis apparatus, the target sample analysis data and the reference sample analysis data are preferably data obtained by analysis through scientific analysis.

  In the data analysis apparatus, the scientific analysis is preferably secondary ion mass spectrometry or X-ray photoelectron spectroscopy.

  Further, the present invention provides a specific element of the plurality of elements from a plurality of peaks included in the target sample analysis data based on target sample analysis data obtained by analyzing a target sample including a plurality of elements. Analyzing a reference sample obtained by analyzing a reference sample from which at least one of the plurality of elements is removed, and a method of analyzing the reference sample from which the target sample analysis data is obtained A step of acquiring data, and for each of the target sample analysis data and the reference sample analysis data, a certain peak of interest is divided into a peak having an intensity greater than or equal to a preset threshold and a peak having an intensity less than the threshold. The classification step for classification and the intensity of the peak of interest before and after the classification. A calculation step for calculating a mutual information amount, and a selection step for selecting a specific peak derived from the specific element based on the mutual information amount obtained for each of the plurality of elements. A value equal to or higher than the average value determined based on the average value of the peak intensity of interest in the reference sample analysis data is set as the threshold value.

  In the data analysis method, when the mutual information amount calculated in the calculation step is less than a preset reference mutual information amount, the classification step and the calculation step are performed until the mutual mutual information amount is equal to or greater than the reference mutual information amount. Preferably it is repeated.

  Further, the present invention provides a specific element of the plurality of elements from a plurality of peaks included in the target sample analysis data based on target sample analysis data obtained by analyzing a target sample including a plurality of elements. A data analysis program for selecting a peak due to a computer obtained by analyzing a target sample analysis data and a reference sample from which at least one of the plurality of elements is removed A means for acquiring reference sample analysis data, and each of the target sample analysis data and the reference sample analysis data has a certain peak of interest, a peak having an intensity greater than or equal to a preset threshold, and an intensity less than the threshold Classification means for classifying into peaks, and the peak of interest before and after the classification Calculation means for calculating mutual information on intensity, and selection means for selecting a specific peak derived from the specific element based on the mutual information obtained for each of the plurality of elements, and the classification means The data analysis device functions as a threshold value that sets a value equal to or greater than the average value determined based on the average value of the peak intensity of interest in the reference sample analysis data.

  In the data analysis program, when the mutual information amount calculated by the calculating means is less than a preset reference mutual information amount, the classification means and the calculating means until the reference mutual information amount is equal to or greater than the reference mutual information amount. Each is preferably classified and calculated.

  According to the present invention, by using a threshold in consideration of the background in the calculation of mutual information, a plurality of elements are included in a simpler process with the minimum number of data sufficient for statistical science. It is possible to more accurately select a peak due to a specific element from a plurality of peaks included in analysis data obtained by analyzing a target sample.

  As shown in FIG. 2, the data analysis system according to an embodiment of the present invention can be configured to include a data analysis device 10 and at least one analysis device 20. The data analysis device 10 and the analysis device 20 are connected to be able to transmit information to each other according to a predetermined protocol via an information transmission medium such as a network. Analysis data obtained by performing a predetermined analysis in the analysis device 20 is transmitted to the data analysis device 10. The data analysis apparatus 10 performs analysis based on the acquired analysis data.

  The data analysis device 10 and the analysis device 20 may function as a single device. The data analysis device 10 may be incorporated in the analysis device 20, or the analysis device 20 may be incorporated in the data analysis device 10. Further, the data analysis device 10 and the analysis device 20 may not be connected via a network or the like. In this case, the analysis data and the like are stored in a recording medium such as a flexible disk, a magneto-optical disk, and an optical disk in the analysis device 20, and a recording medium data reading means such as a flexible disk drive, a magneto-optical disk drive, and an optical disk drive in the data analysis device 10. Read by etc. Further, the data analysis system in the present embodiment does not necessarily require the analysis device 20. When analyzing the data obtained without going through the analysis device 20, analysis data or the like may be directly input in the data analysis device 10.

  As shown in FIG. 2, the data analysis apparatus 10 can be configured to include a control unit 100, a storage unit 120, a display unit 140, an input unit 160, a data interface 180, and the like.

  The control unit 100 includes a control element such as a CPU, and has a function of controlling the data analysis apparatus 10 in an integrated manner.

  The storage unit 120 includes a semiconductor memory, a hard disk, and the like, and stores and holds electronic information such as programs processed by the control unit 100, analysis data, and calculation data. When a large amount of data is stored and held, it is also preferable to use a mass storage device such as a hard disk device, a magnetic tape device, a magnetic disk device, an optical disk device, or a magneto-optical disk device as an auxiliary storage device.

  The display unit 140 includes a monitor and the like, and is used for presenting information from the data analysis apparatus 10 to the user.

  The input unit 160 includes a keyboard, a mouse, a touch panel, and the like, and is used to acquire various types of information used for processing from the user. The display unit 140 and the input unit 160 may be an integrated user interface such as a touch panel.

  The data interface 180 is connected to the data interface in the analysis apparatus 20 through a network or the like so as to be able to transmit information, and is used for transmitting and receiving analysis data and the like.

  Further, the data analysis device 10 has been described as including the control unit 100, the storage unit 120, the display unit 140, the input unit 160, and the data interface 180. However, all of these configurations need to be included in one device. There is no. For example, the data analysis apparatus 10 may be appropriately combined with a configuration including the control unit 100, the display unit 140, and the data interface 180, with the storage unit 120 and the input unit 160 as independent devices.

  The data analysis apparatus 10 executes a data analysis program according to the present embodiment, which is stored in the storage unit 120, so that the analysis apparatus 20 analyzes a target sample including a plurality of elements and obtains the target sample analysis Analysis based on the data. Alternatively, the data analysis device 10 executes the data analysis program according to the present embodiment stored in the storage unit 120, and is obtained without passing through the analysis device 20, and includes a target sample including a plurality of elements. The analysis is performed based on the target sample analysis data obtained by analyzing the above.

  Here, the target sample analysis data in the present embodiment is not particularly limited as long as it is data obtained by investigating the target sample in the fields of natural science and social science and can be plotted with two-dimensional coordinates. For example, statistical data, measurement data, calculation data, map data, spectrum data, and the like can be given. Further, the target sample analysis data is preferably data obtained by analysis through scientific analysis, which is a technical technique for identifying a substance, detecting a substance, and determining a chemical composition of the substance. The target sample analysis data may be data obtained without going through the analysis device 20, for example, data obtained by manually collecting questionnaires.

  The analysis device 20 is not particularly limited as long as it is a device that analyzes target samples including a plurality of elements and obtains target sample analysis data, but may be a scientific analysis device capable of performing scientific analysis. preferable.

  Scientific analysis includes, for example, nuclear magnetic resonance spectroscopy (NMR), infrared spectroscopy (IR), Raman spectroscopy, ultraviolet-visible spectroscopy, atomic absorption spectroscopy, electron spin resonance spectroscopy (ESR), fluorescent X-ray Spectroscopic analysis (XRF), X-ray photoelectron spectroscopy (XPS), etc .; secondary ion mass spectrometry (SIMS), mass analysis (MS) such as matrix-assisted laser desorption ionization (MALDI); X-ray diffraction analysis Separation analysis of high performance liquid chromatograph (HPLC), gas chromatograph (GC), gel filtration chromatograph (GPC), etc .; differential scanning calorimetry (DSC), thermogravimetric / differential thermal analysis (TG-DTA), thermogravimetric Thermal analysis such as mass spectrometry (TG-MS); Moreover, the analysis which combined each of these analysis methods may be sufficient. The scientific analysis device may be a device that combines devices that perform these analysis methods.

  In this embodiment, it is preferable that the scientific analysis is an analysis method in which complicated spectra, chromatograms, and the like are often obtained when a sample including a plurality of components is measured. Of these, secondary ion mass spectrometry (SIMS) and X-ray photoelectron spectroscopic analysis (XPS) are more preferable, and more complex spectra can be obtained when analyzing samples containing various elements including proteins and polymer materials. More preferably, it is time-type secondary ion mass spectrometry.

  One embodiment of the data analysis method of the present invention is realized by executing each step of the flowchart shown in FIG. Each step of the data analysis method is converted into a data analysis program that can be processed by a computer, and stored and held in the storage unit 120. The control unit 100 reads a data analysis program held in the storage unit 120 and sequentially executes each process. In this embodiment, each process is not an essential requirement and can be executed in combination as appropriate.

  First, in step S10, the analysis device 20 performs a predetermined analysis on a target sample including a plurality of elements. The analysis apparatus 20 transmits the analyzed target sample analysis data to the data analysis apparatus 10 via a network or the like using a data interface or the like. In the data analysis apparatus 10, target sample analysis data is acquired using the data interface 180 and stored in the storage unit 120. This analysis is preferably performed multiple times for one target sample. More preferably, it is performed 3 times or more.

  In step S12, the analysis device 20 performs the same analysis as the target sample on the reference sample from which at least one of the plurality of elements included in the target sample is removed. The number of reference samples is preferably equal to or greater than the number of elements to be analyzed included in the target sample. The analysis conditions for the reference sample are preferably the same as those for the analysis of the target sample performed in step S10. The analysis device 20 transmits reference sample analysis data measured via a network or the like using a data interface or the like to the data analysis device 10. In the data analysis apparatus 10, reference sample analysis data is acquired using the data interface 180 and stored in the storage unit 120. This analysis is preferably performed multiple times for one target sample. More preferably, it is performed 3 times or more.

  Here, the target sample analysis data and the reference sample analysis data obtained by the analyzer 20 include spectrum data or chromatograms corresponding to the measurement method performed by the analyzer 20.

  In step S14, for the spectrum data or chromatogram included in the target sample analysis data and the reference sample analysis data, the intensity of a certain peak of interest is set to an intensity equal to or higher than the threshold V based on a threshold V preset for the peak of interest. And a peak having an intensity less than the threshold value V. Then, the information entropy before and after the classification and the mutual information amount that is the amount of change in the information entropy before and after the classification are calculated. The peak intensity generally refers to the peak intensity or peak height in data obtained by a certain analysis, but in the present embodiment also includes the peak area intensity. It also includes peak frequency.

Here, the information entropy is an index representing the ambiguity of the state, similar to the entropy of statistical mechanics. An index that objectively represents the state from the probabilities of all the events that can occur in a certain state, and is represented by S = −Σpi (log ₂ pi) (S: information entropy, pi: probability of occurrence of an event i) . The mutual information amount (mutual entropy) is obtained for the event system A after obtaining information about the event system B from the information entropy S (A) of the event system A when there are two event systems A and B. It is obtained by subtracting information entropy S (A / B), and is represented by I (A; B) = S (A) −S (A / B) (I: mutual information amount). The initial ambiguity is the index that indicates whether or not the ambiguity after examining something is subtracted from the initial ambiguity.

  First, a priori entropy is calculated. The control unit 100 analyzes the number of data analyzed by the analyzer 20 from the target sample analysis data and the reference sample analysis data held in the storage unit 120 in steps S10 and S12, and the target sample was analyzed by the analyzer 20. The prior entropy S (A) is calculated based on the equation (1), where n (a1) is the number of data and n (a2) is the number of data analyzed by the analyzer 20 for the reference sample (i = 1, 2).

S (A) = − Σ [p (ai) log ₂ p (ai)] (1)

  Here, the establishment p (ai) is expressed by Expression (2).

  p (ai) = n (ai) / N (2)

  Next, a posteriori entropy is calculated. First, the control unit 100 sets a threshold value V1 for the intensity of a certain peak of interest. And it classify | categorizes into the peak which has an intensity | strength more than threshold value V1, and the peak which has an intensity | strength less than threshold value V1 in the analyzed series of data. Among the number of data n (a1) analyzed for the target sample, n (a1, b1) is the number of data greater than or equal to the threshold value V1, and n (a1, b2) is the number of data less than the threshold value V1. In the number n (a2), the number of data greater than or equal to the threshold value V1 is n (a2, b1), the number of data less than the threshold value V1 is n (a2, b2), and the a posteriori entropy S (A / B) is expressed by Equation (3). Calculate based on (j = 1, 2).

S (A / B) = − Σ [n (ai) / N] · Σ [p (ai, bj) log ₂ p (ai, bj)] (3)

  Here, the establishment p (ai, bj) is expressed by Expression (4).

  p (ai, bj) = n (ai, bj) / N (4)

  And the control part 100 calculates the mutual information amount I (A; B) represented by Formula (5).

  I (A; B) = S (A) -S (A / B) (5)

  For example, as shown in FIG. 1 (a), each intensity of a peak of interest in 5 data analyzed for sample 1 (target sample here) and data analyzed for sample 2 (here reference sample) When the intensities of the same peak of interest as Sample 1 among 5 points cannot be distinguished at all by the set threshold value V, the mutual information amount I (A; B) becomes the minimum value 0. As shown in FIG. 1B, when the peaks of sample 1 and sample 2 can be completely distinguished with threshold V, mutual information I (A; B) has a maximum value of 1.

When the mutual information amount I (A; B) calculated by the equation (5) becomes equal to or larger than the predetermined reference mutual information amount I ₀ , the calculation of the mutual information amount is ended there. If the value does not exceed I _{0, the} threshold value V1 is slightly lowered and reset as the threshold value V2, and the mutual information amount I (A; B) is calculated. That is, step S14 is repeated until the mutual information amount becomes I ₀ or more. However, the calculation may be terminated when the threshold value V is equal to or less than a predetermined value V ₀ . In this case, the peak is not selected as a specific peak. The setting of I ₀ can be, for example, S (A), 0.9 × S (A), or the like.

In the present embodiment, when comparing peak intensities in step S14, it is not compared whether the peak intensities can be distinguished between samples, but a threshold considering the background, that is, the peak of interest in the reference sample analysis data. As the intensity value of the intensity, a value equal to or higher than the average value determined based on the average value of the peak intensity of interest in the series of data is used. For example, the mutual information amount is preferably calculated by setting the average value of the peak intensity of interest to 1 to 20 times, preferably 1 to 10 times, as the first set threshold V1. If the unit is not to be I ₀ or more, until I ₀ or more, for example, it is preferable to lower by 0.1-fold to 0.5-fold increments of the average value of the target peak intensity thresholds from V1.

  Furthermore, the mutual information amount is calculated for both the case where the maximum mutual information amount is given as a threshold value and the case where a threshold value considering the background (a value equal to or higher than the average value of the peak intensity in the reference sample) is used. It is preferable. In this case, it is preferable that the calculation result of the mutual information amount by the threshold considering the background is prioritized. When an ion peak having a sufficiently large mutual information amount cannot be obtained in consideration of the background, the value of the mutual information amount based on a threshold value that maximizes the conventional mutual information amount is used.

  The process of step S14 may be performed for each peak in the data analyzed for the target sample and all the peaks in the data analyzed for the reference sample, or may be selected in advance from each peak. May be performed on the peak.

  In step S16, the control unit 100 selects a specific peak based on the mutual information amount obtained in step S14 and the threshold value when the mutual information amount is calculated. For each target sample and reference sample, an arbitrary number of peaks are generated in order from the largest mutual information amount, from the largest threshold value, or from the largest mutual information amount and the higher threshold value. It is preferable to be elected. In addition, the selected peaks are preferably displayed using the display unit 140 in this order. As for the number of specific peaks to be selected, one is sufficient for each sample when the mutual information is close to 1 and the peak intensity is sufficiently high, but when the peak intensity is low, two or more specific peaks are selected. It is preferable that a peak is selected. For example, in the case of imaging by time-of-flight secondary ion mass spectrometry (TOF-SIMS), which will be described later, the number of specific peaks to be selected is determined for each sample when the peak intensity is sufficiently high to obtain a clear image. It is preferable that two or more specific peaks are selected when the intensity is low enough to obtain only a blurred image that cannot be distinguished from the background.

  As described above, a peak caused by a specific element can be selected from a plurality of peaks included in analysis data obtained by analyzing a target sample including a plurality of elements. This embodiment is an example in the case where two types of elements are included in the target sample. However, when three or more types of elements are included, a peak due to a specific element is selected as shown in FIG. Also good. For example, when four types of elements are included, in Stage 1, the substance 4 is classified as either the substance 4 or the other two types. Stage 2 is classified as substance 1 and others, stage 3 is classified as substance 2 and substance 3, and a specific peak is selected for each of substances 1 to 4.

  Moreover, it is preferable that qualitative, quantitative, distribution imaging and the like of the specific element are performed based on the selected peak of the specific element. When a plurality of specific peaks are selected, it is preferable that qualitative, quantitative, distribution imaging, etc. of the specific elements are performed based on the plurality of specific peaks. In this case, qualitative, quantitative, distribution imaging, etc. may be performed based on all of the selected plurality of specific peaks, or qualitative based on any number of peaks selected from the selected plurality of specific peaks, Quantification, distribution imaging, etc. may be performed.

  Qualitative results, quantitative results, distribution imaging results, and the like are preferably displayed using the display unit 140. It is preferable that the qualitative result, the quantitative result, the distribution imaging result, and the like displayed using the display unit 140 change according to which peak is selected from the selected plurality of specific peaks.

  The data analysis method of this embodiment uses mutual information by comparison with a reference sample using information entropy for the rapid detection of characteristic peaks derived from complex chemical species not in the database. This is particularly effective when it is difficult to predict patterns such as mass spectra and XPS spectra. Moreover, the importance of each fragment ion can be collectively evaluated with the minimum number of data required in statistical science regardless of the number of peaks.

  Hereinafter, although an example and a comparative example are given and the present invention is explained more concretely in detail, the present invention is not limited to the following examples.

  The following examples show how specific time-of-flight secondary ion mass spectrometry (TOF-SIMS) can be applied to a specific substance from a mass spectrum composed of complex fragment ions obtained for a target sample containing proteins and polymeric materials. The resulting peak is selected based on the data analysis method of one embodiment of the present invention, and the distribution of the specific substance on the sample is obtained based on the peak.

(Example)
The result of investigating the distribution of protein bovine serum albumin (BSA) adsorbed on the inner and outer surfaces and pores of a hollow fiber type polymer dialysis membrane containing polysulfone as a main component is shown.

  Bovine serum albumin (BSA; Sigma) in 0.1 M phosphate buffered saline (PBS, pH = 7.4) on a hollow fiber type polymer dialysis membrane APS-E (manufactured by Asahi Medical) mainly composed of polysulfone Chemical solution (3.75 g / dL) was dialyzed for 7 hours at a flow rate of 1.2 mL / hr. After that, PBS was allowed to flow for 55 minutes at the same speed, and further washed with distilled water. After washing, the polymer membrane was dried in a desiccator.

Using a 12 keV Ga ⁺ ion source as a primary ion source, a mass spectrum was obtained for a target sample prepared by a time-of-flight secondary ion mass spectrometer TFS-2100 (manufactured by Physical electrons). The measurement was performed in the same manner using a washed APS-E polymer membrane that had no action as a reference sample.

  Since fragment ions attributed to membrane constituents are similar to fragment ions attributed to BSA, a simple ion comparison cannot identify a specific ion peak. Therefore, mass spectra obtained by time-of-flight secondary ion mass spectrometry of a washed membrane (sample A) that has not been acted on and a membrane (sample B) that has been dialyzed with an aqueous BSA solution are shown in one embodiment of the present invention. Analysis was performed according to the method.

  Within the measurement range of the mass spectrum, 40 to 210 amu were regarded as effective fragment ion peaks, and the intensity of each peak (normalized by the total ion intensity) was compared between Sample A and Sample B. When examining the sample A, the average value of the peak intensity in the sample B was used as the lowest standard, and compared with the peak intensity in the sample A, the mutual information amount was calculated. For all the effective fragment ion peaks from 40 to 210 amu, the case for sample A and the case for sample B were performed respectively.

  From the result of mutual information, a fragment ion derived from sample A, that is, this membrane element, and a fragment ion derived from sample B, that is, BSA, are selected. Based on the selected fragment ions (specifically BSA: m / z = 45, 60, 72, 88 amu, APS-E: m / z = 50, 82, 86, 112 amu), the membrane, and the distribution of BSA The figure was obtained as a secondary ion image (imaging).

  FIG. 5 shows a distribution image of BSA adsorbed on the hollow fiber type polymer membrane based on the fragment ion peak of BSA obtained in Examples. Further, FIG. 6 shows an image diagram of an untreated polymer film not adsorbed with BSA obtained in the same manner. In FIGS. 5 and 6, the white (bright) part is considered to be a part where a lot of BSA exists. In FIG. 5, it can be seen that a large amount of BSA exists on the inner and outer surfaces of the polymer film, but hardly exists in the cross section. In FIG. 6, almost no white portion is seen, and it can be seen that BSA is not present in the untreated polymer film and a correct distribution is obtained.

(Comparative example)
A fragment ion derived from BSA was selected in the same manner as in the example except that the calculation was performed using a threshold value that gives the maximum amount of mutual information as in the past (specifically, m / z = 45, 46). 58, 59, 60, 72, 88 amu).

  FIG. 7 shows a distribution image of BSA adsorbed on the hollow fiber type polymer membrane based on the BSA fragment ion peak obtained in the comparative example. Further, FIG. 8 shows an image diagram of an untreated polymer film not adsorbed with BSA obtained in the same manner. FIG. 7 shows that the inner and outer surfaces of the polymer film are white, but the cross section is also slightly white. Also in FIG. 8, white portions are seen on the inner and outer surfaces, and it can be seen that a correct distribution is not obtained as compared with the example.

  TOF-SIMS is a technique capable of identifying, quantifying, and imaging the distribution of chemical species on the material surface, but in the case of macromolecules such as proteins, analysis from fragment ions is required. Since proteins are composed of the same 20 kinds of amino acids, it is difficult to select fragment ions derived from a specific protein because fragment ions are similar even in different proteins. In this example, by analyzing data based on information entropy, it was possible to simply select fragment ions originating from proteins and polymer membranes, and to image protein adsorption on the membranes.

  According to this example, in particular, in a secondary ion mass spectrum of a sample containing substances composed of macromolecules having structures similar to each other, a secondary ion that can be determined to be derived only from each substance is a simple reference sample. Even if it is difficult to make a decision based on the comparison or the estimation from the chemical structure, it can be selected quickly. In particular, in samples containing multiple proteins, even when cluster ions with high ionization efficiency of macromolecules were used as the primary ion source, complex analysis and a huge amount of data were required to identify each protein. By this example, the efficiency of determination of fragment ions caused by each protein can be further improved.

  According to the present invention, by using a threshold in consideration of background in the calculation of mutual information, a plurality of elements are included in a simpler process with a minimum number of data sufficient for statistical science. It is possible to more accurately select a peak due to a specific element from a plurality of peaks included in target sample analysis data obtained by measuring a sample. Complex data analysis can be performed easily and quickly in a wide range of natural and social sciences.

(A) It is a figure which shows the relationship between the focused peak intensity | strength and threshold value in a series of data about the samples A and B in case of mutual information = 0. (B) It is a figure which shows the relationship between the focused peak intensity | strength and threshold value in a series of data about the samples A and B in case of mutual information = 1. It is a block diagram which shows the data analyzer in one Embodiment of this invention. It is a chart figure which shows the data analysis method in one Embodiment of this invention. It is a figure which shows the selection method of the specific peak in the case of including 3 or more types of elements in one Embodiment of this invention. It is a figure which shows the distribution image of the protein which adsorb | sucked to the polymer film surface in the Example of this invention. It is a figure which shows the distribution image of the protein of the surface of the non-processed polymer film which is a reference sample in the Example of this invention. It is a figure which shows the distribution image of the protein adsorb | sucked to the polymer film surface in the comparative example of this invention. It is a figure which shows the distribution image of the protein of the surface of the non-processed polymer film which is a reference sample in the comparative example of this invention.

Explanation of symbols

  10 data analysis device, 20 analysis device, 100 control unit, 120 storage unit, 140 display unit, 160 input unit, 180 data interface.

Claims (14)

Based on target sample analysis data obtained by analyzing a target sample including a plurality of elements, a peak due to a specific element among the plurality of elements is selected from a plurality of peaks included in the target sample analysis data A data analysis device for
Means for obtaining the target sample analysis data;
Means for obtaining reference sample analysis data obtained by analyzing a reference sample from which at least one of the plurality of elements is removed;
For each of the target sample analysis data and the reference sample analysis data, a classification means for classifying a certain peak into a peak having an intensity equal to or higher than a preset threshold and a peak having an intensity less than the threshold;
Calculation means for calculating mutual information regarding the intensity of the peak of interest before and after the classification;
Selection means for selecting a specific peak derived from the specific element based on the mutual information obtained for each of the plurality of elements;
Have
The data analysis apparatus characterized in that, in the classification means, a value equal to or greater than an average value determined based on an average value of the peak intensity of interest in the reference sample analysis data is set as the threshold value. The data analysis apparatus according to claim 1,
If the mutual information amount calculated by the calculating means is less than a preset reference mutual information amount, the classification means and the calculating means repeat the classification and calculation until the reference mutual information amount is greater than or equal to the reference mutual information amount. Characteristic data analysis device. The data analysis device according to claim 1 or 2,
Furthermore, in the classification means, a threshold value that gives the maximum mutual information amount in the calculation means is set as a threshold value. The data analysis apparatus according to claim 3,
If the mutual information amount calculated by the calculating means does not reach the reference mutual information amount, the selecting means determines the specific information based on the mutual information amount before and after the classification based on the threshold value that gives the maximum mutual information amount. A data analysis device characterized by being selected as a peak. The data analysis device according to any one of claims 1 to 4,
In the selecting means, the specific peak selects a peak having a large mutual information value as the specific peak. A data analysis device according to any one of claims 1 to 5,
The target sample analysis data and the reference sample analysis data are analyzed a plurality of times for each of the target sample and the reference sample. The data analysis device according to any one of claims 1 to 6,
Means for imaging the distribution of the particular element in the target sample based on the particular peak;
A data analysis apparatus characterized by comprising: The data analysis device according to any one of claims 1 to 6,
Further, means for quantifying the specific element in the target sample based on the specific peak,
A data analysis apparatus characterized by comprising: A data analysis device according to any one of claims 1 to 8,
The target sample analysis data and the reference sample analysis data are data obtained by analyzing by scientific analysis. The data analysis device according to claim 9, wherein
The data analysis apparatus characterized in that the scientific analysis is secondary ion mass spectrometry or X-ray photoelectron spectroscopy. Based on target sample analysis data obtained by analyzing a target sample including a plurality of elements, a peak due to a specific element among the plurality of elements is selected from a plurality of peaks included in the target sample analysis data A data analysis method for
Obtaining the target sample analysis data;
Obtaining reference sample analysis data obtained by analyzing a reference sample from which at least one of the plurality of elements is removed;
For each of the target sample analysis data and the reference sample analysis data, a classification step of classifying a certain peak into a peak having an intensity equal to or higher than a preset threshold and a peak having an intensity less than the threshold;
A calculation step of calculating mutual information regarding the intensity of the peak of interest before and after the classification;
A selection step of selecting a specific peak derived from the specific element based on the mutual information obtained for each of the plurality of elements;
Have
A data analysis method characterized in that, in the classification step, a value equal to or greater than an average value determined based on an average value of the peak intensity of interest in the reference sample analysis data is set as the threshold value. The data analysis method according to claim 11, comprising:
If the mutual information amount calculated in the calculation step is less than a preset reference mutual information amount, the classification step and the calculation step are repeated until the mutual mutual information amount is equal to or greater than the reference mutual information amount. analysis method. Based on target sample analysis data obtained by analyzing a target sample including a plurality of elements, a peak due to a specific element among the plurality of elements is selected from a plurality of peaks included in the target sample analysis data A data analysis program for
Computer
Means for obtaining the target sample analysis data;
Means for obtaining reference sample analysis data obtained by analyzing a reference sample from which at least one of the plurality of elements is removed;
For each of the target sample analysis data and the reference sample analysis data, a classification means for classifying a certain peak into a peak having an intensity equal to or higher than a preset threshold and a peak having an intensity less than the threshold;
Calculating means for calculating mutual information regarding the intensity of the peak of interest before and after the classification;
Selection means for selecting a specific peak derived from the specific element based on the mutual information obtained for each of the plurality of elements;
Have
A data analysis program that causes the classification means to function as a data analysis apparatus that sets a value equal to or greater than an average value determined based on an average value of the peak intensity of interest in the reference sample analysis data as the threshold value. A data analysis program according to claim 13, comprising:
If the mutual information amount calculated by the calculating means is less than a preset reference mutual information amount, the classification means and the calculating means repeat the classification and calculation until the reference mutual information amount is greater than or equal to the reference mutual information amount. A featured data analysis program.