JP6319716B2 - SPECTRUM PREDICTION METHOD, SPECTRUM PREDICTION DEVICE, AND PROGRAM - Google Patents

Description

  The present invention relates to a spectrum prediction method, a spectrum prediction apparatus, and a program. More specifically, the present invention relates to a spectrum prediction method, a spectrum prediction apparatus, and a program that predict a spectrum acquired in mass spectrometry by calculation.

A sample ionization method is known as one of mass spectrometry for determining the structure of a biomolecule. MALDI (Matrix-Assisted Laser Desorption / Ionization)-MS ⁿ (Multiple-stage Mass Spectrometry) is based on the characteristic fragmentation pattern in the MS ⁿ spectrum. It is an ionization method for samples that is indispensable for characterizing various classes of biomolecules including sugars.

MALDI-MS ⁿ uses the fact that fragmentation of sugar molecules by collision induced dissociation (CID) depends on the nature of chemical bonds in the molecule, and is obtained as a result of repeated fragmentation of sugar molecules by CID. By observing the resulting MS ⁿ spectrum, the isomer structure of the sugar molecule can be identified.

In the determination of the molecular structure by MS ⁿ spectrum, a public database of sugar chain structures is searched based on the actually measured spectrum data to find the corresponding structure. When there is no corresponding data, the composition is estimated from the m / z value (m = mass, z = number of charges) of the actually measured data. Next, the composition combination is estimated and the composition is narrowed down. The information recorded in the fragment database is then used to predict the spectrum of the estimated composition. Then, the structure is determined by judging the coincidence between the predicted spectrum and the actually measured spectrum.

  In recent years, various methods for predicting a spectrum from a combination of compositions have been studied.

  For example, as a method of analyzing a sugar chain structure by mass spectrometry, fragmentation patterns are obtained by actually fragmenting all kinds of sugar chains, and this is accumulated as data. There is a method for predicting a sugar chain structure by comparing with a fragmentation pattern of a test sugar chain (see, for example, Patent Document 1).

However, in order to predict the sugar chain structure of a sugar molecule from the observed MS ⁿ spectrum, a complicated and complicated process such as statistical analysis is required. And even if the molecular structure of the sugar molecule to which MALDI-MS ⁿ is applied is known, it is difficult to predict the MS ⁿ spectrum from the molecular structure of the known sugar molecule.

As a method for solving the difficulty of predicting such MS ⁿ spectrum, the MS of a given sugar molecule is used by simulating fragmentation of a sugar molecule by using the probability of breaking a chemical bond contained in the given sugar molecule. A method for predicting an ⁿ spectrum has been proposed (Patent Document 2).

International Publication No. WO2006 / 112343 Publication Pamphlet JP 2011-220733 A

Harvey DJ. “Fragmentation of negative ions from carbohydrates: part 2. Fragmentation of high-mannose N-linked glycans”, J Am Soc Mass Spectrom. 2005 May; 16 (5): 631-46

However, the conventional method of predicting the MS ⁿ spectrum using the probability of cleavage of a chemical bond contained in a given sugar molecule has a problem that the calculation is complicated and it is difficult to accurately determine the probability of cleavage. .

The present invention has been made in view of such problems, and an object of the present invention is to provide a spectrum prediction method, a spectrum prediction apparatus, and a program for accurately predicting an MS ⁿ spectrum by simple calculation. It is in.

The present inventor uses the probability of deprotonation of a hydroxyl group or amino group contained in a given sugar molecule and an empirical fragmentation pattern in the simulation of fragmentation of a given sugar molecule, and the MS of the given sugar molecule We have found that ⁿ spectra can be predicted accurately. Based on this knowledge, a spectrum prediction method, a spectrum prediction apparatus, and a program for predicting the MS ⁿ spectrum of a given sugar molecule were completed.

That is, in the present invention, rather than predicting MS ⁿ spectra from a database or the cutting probability, based on the deprotonation probability of sugar molecules, predicting MS ⁿ spectra.

The spectrum prediction method according to the present invention predicts the fragment type generated by deprotonation at each site of the sugar molecule using the deprotonation probability at each site of the sugar molecule and an empirical fragmentation pattern. And predicting an MS ⁿ spectrum of the sugar molecule based on the deprotonation probability and the predicted fragment species information.

  Here, using the information on the structure of the sugar molecule, the step of calculating the deprotonation probability in each part of the sugar molecule by performing deprotonation simulation in each part of the sugar molecule, The step of predicting the fragment type may use the calculated deprotonation probability.

  Here, the information on the structure of the sugar molecule may include positional information on a hydroxyl group or an amino group that can be deprotonated in the sugar molecule.

Further, the information on the structure of the sugar molecule may include information on the MS ⁿ spectrum experimentally observed for the sugar molecule.

  The deprotonation simulation may include a theoretical calculation based on information on the structure of the sugar molecule.

  Here, the theoretical calculation may include a quantum calculation.

  Moreover, the information regarding the structure of the sugar molecule may include a physical quantity regarding the sugar molecule.

  In addition, the physical quantity may include atomic interaction energy or structure stabilization energy before and after deprotonation at the hydroxyl group or amino group of the sugar molecule.

According to another aspect of the present invention, a spectrum prediction apparatus according to the present invention uses a deprotonation probability at each site of a sugar molecule and an empirical fragmentation pattern to deprotonate at each site of the sugar molecule. Means for predicting the fragment type generated by the conversion, and means for predicting the MS ⁿ spectrum of the sugar molecule based on the deprotonation probability and the information on the predicted fragment type.

  According to another aspect of the present invention, a program according to the present invention causes a computer to execute the above spectrum prediction method.

According to the present invention, it is possible to accurately predict an MS ⁿ spectrum from a molecular structure of a given sugar molecule with a simple calculation method.

It is a figure explaining the basic principle which estimates a spectrum from a given molecular structure. It is a figure which shows an example of the rule of thumb regarding the generation pattern of a fragment. It is a block diagram which shows the structure of the spectrum estimation apparatus which concerns on one Embodiment of this invention. It is a figure showing roughly an example of sugar chain structure concerning one embodiment of the present invention. It is a figure which shows the prediction simulation result of the fragment kind which concerns on one Embodiment of this invention. The flowchart of the process which the spectrum estimation apparatus which concerns on one Embodiment of this invention performs is shown. It is a figure which shows GUI which concerns on one Embodiment of this invention. It is a figure which shows the sugar chain structure after performing structure optimization calculation of Laminaripentaose-PBH. It is a figure which shows the sugar chain structure after performing structure optimization calculation of Laminaripentaose-PBH. It is a figure which shows the calculation result of the prediction simulation of the fragment kind by 1st Example of this invention. The sugar chain structure is shown when deprotonation is simulated for Laminaripentaose-PBH. (A) shows the experimental value of the MS ² spectrum of Laminaripentaose-PBH, and (b) shows the prediction result of the MS ² spectrum predicted by the first example of the present invention. It is a figure which shows the sugar_chain | carbohydrate structure of AHSG, (a) is the figure which symbolized each unit, (b) is the figure which showed each unit by the name. The calculation result by 2nd Example of this invention is shown. (A) Experimental value of MS ³ spectrum, (b) shows the predicted result of MS ³ spectrum predicted by the second embodiment of the present invention.

  Embodiments according to the present invention will be described below with reference to the drawings.

The basic principle of predicting a spectrum from a given molecular structure in MALDI-MS ⁿ will be described with reference to FIG. When a sample is irradiated with an ultraviolet laser, the physical phenomenon is that deprotonation occurs at the hydroxyl group or amino group in the unit, and then the electron at the deprotonated position moves, resulting in a more stable structure. To do. During this electron transfer, a part of the bond is cut to generate ions. The ions generated in this way are called fragments.

  In the conventional spectrum prediction, for each site to be cleaved, each of the three steps of deprotonation, bond cleavage, and electron transfer corresponds to the process of deprotonation → electron transfer → bond cleavage → fragment generation. Probability was estimated.

  Here, in a series of steps of electron transfer → bond cleavage → fragment generation, there is a certain pattern depending on the site to be deprotonated. That is, the site of the unit to be deprotonated is associated with the fragment type to be generated according to the determined fragment generation pattern. Therefore, in the present invention, the deprotonation probability is estimated, while the step of electron transfer → bond cleavage → fragment generation determines the fragment generation pattern based on empirical rules. The deprotonation probability represents the probability that deprotonation occurs at a predetermined position in the molecular structure of the sugar molecule.

FIG. 2 is a diagram showing an example of an empirical rule regarding a fragment generation pattern (see Non-Patent Document 1). In the example shown in the figure, deprotonation occurs at the amino group (—NHOCH ₃ —NH). Then, when the change to a more stable chemical structure electrons move, glycosidic linkages (R-O-R ₁₎ is cut, R ₁ -O ^- ions (Y ions) are generated. Thus, for deprotonation in the position of the figure N, R ₁ -O as a fragment species produced ^- ions is associated.

  Such a fragment generation pattern can be defined based on what is known from paper information, measurement data analysis results, and the like. A plurality of fragment generation patterns are stored in advance in the fragment prediction apparatus.

  Next, the configuration of the spectrum prediction apparatus according to the present embodiment will be described with reference to FIG.

In the present embodiment, the spectrum prediction apparatus 400 predicts an MS ⁿ spectrum from the molecular structure of a given sugar molecule by executing a process shown in FIG. 6 described later using a parameter called deprotonation probability.

  The CPU 402 reads the program from the storage unit 408 and executes it.

  The programs executed here include a deprotonation simulation program, a fragment prediction program, and a spectrum prediction program.

  The deprotonation simulation program is a program for calculating a deprotonation probability by simulating deprotonation in a sugar molecule using sugar chain structure data.

  The fragment prediction program is a program for predicting fragment species generated by deprotonation using sugar chain structure data and fragmentation patterns. Specifically, the fragment prediction program determines the position of a bond cleaved as a result of deprotonation in a sugar molecule based on the fragmentation pattern, and predicts the fragment type generated by the cleavage.

  Here, the sugar chain structure data is included in an input file described later. On the other hand, the fragmentation pattern is data in which a site to be deprotonated, a generated fragment type, and a cleavage position are associated with each other, and is written in the fragment prediction program.

The spectrum prediction program is a program that predicts the MS ⁿ spectrum of a sugar molecule based on the deprotonation probability and information on the predicted fragment type.

  When the CPU 402 executes the above program, the input file stored in the storage unit 408 is used. The input file is input from the input unit 404 and stored in the storage unit 408 in advance.

The input file contains the parameters required for simulation to predict the MS ⁿ spectrum. For example, the input file includes sugar chain structure data.

  The sugar chain structure data is data on the structure of the sugar molecule. For example, the types of monosaccharides (units such as Gal, GlcNAc, and Man), how they are connected (order), and binding information (what number carbon is connected). Can be included).

  Further, the sugar chain structure data may include positional information of a hydroxyl group or an amino group that can be deprotonated in the sugar molecule.

In addition, the sugar chain structure data may include information on MS ⁿ spectra experimentally observed for sugar molecules.

  Further, the sugar chain structure data may include physical quantities related to sugar molecules.

  Here, the physical quantity may include atomic interaction energy before and after deprotonation at the hydroxyl group or amino group of the sugar molecule.

  Further, the physical quantity may include the structure stabilization energy before and after deprotonation at the hydroxyl group or amino group of the sugar molecule.

  By executing the above program, the spectrum predicting apparatus 400 performs deprotonation simulation on the sugar molecule using the sugar chain structure data, and removes each hydroxyl group or amino group position of the given sugar molecule. Calculate protonation probability.

  Next, the fragment species generated by deprotonation are predicted using the sugar chain structure data and the fragmentation pattern.

Then, the MS ⁿ spectrum of the sugar molecule is predicted based on the deprotonation probability and the predicted fragment type information.

FIG. 4 schematically shows an example of a sugar chain structure according to this embodiment. The sugar chain structure shown in the figure is labeled with PBH (1-pyrenebutanoic acid hydrazide). Symbols shaded circles (●), shaded squares (squares) and white circles (○) indicate units constituting sugar chains, ● indicates Mannose, squares indicate GlcNAc (N-acetyl-D-glucosamine), ○ indicates Galactose Show. Also, unique identification information (Y ₅ , ² , ⁴ A ₅ etc.) is assigned to each cleavage position of the sugar chain structure.

FIG. 5 shows the result of a simulation of fragment species prediction for deprotonation in an example of a sugar chain structure. The table shown in the figure shows, from left to right, the predicted fragment species m / z (indicated by the symbol mz :), the peak intensity in the predicted value of the MS ⁿ spectrum (indicated by the symbol level :), and possible cuts Represents a position (indicated by ions :). In this case, information is divided by a symbol “/” in multiple cuttings.

  The execution result of the process based on the spectrum prediction program may be displayed on the display unit 406 according to a command from the CPU 402.

FIG. 6 shows a flowchart 500 of processing executed when the spectrum prediction apparatus 400 according to the present invention predicts an MS ⁿ spectrum.

  First, in step S502, the spectrum prediction apparatus 400 reads an input file from the storage unit 408 when the deprotonation simulation program is executed.

  Next, in step S504, the spectrum prediction apparatus 400 performs deprotonation simulation of each site in the sugar molecule based on the sugar chain structure data included in the input file read in step S502. As a result of this simulation, the deprotonation probability is calculated.

  Here, the deprotonation probability is the position of the hydroxyl group or amino group based on, for example, positional information (1st position, 2nd position, etc.) of the hydroxyl group or amino group binding site contained in the sugar molecule where deprotonation can occur. You may calculate for every.

Further, the deprotonation probability may be calculated based on information of MS ⁿ spectrum experimentally observed for sugar molecules. For example, the deprotonation probability may be determined based on the relative intensity of the spectrum in the actually measured value for the fragment type predicted to be generated by the deprotonation simulation.

  Further, the deprotonation probability may be calculated by theoretical calculation based on information on the structure of the sugar molecule.

  Here, as the theoretical calculation, for example, a quantum calculation such as a density functional (DFT) method may be used.

  Further, the deprotonation probability may be calculated based on the physical quantity related to the sugar molecule.

  Here, the interaction energy between atoms before and after deprotonation at the hydroxyl group or amino group of the sugar molecule may be used as a physical quantity.

  Further, the structure stabilization energy before and after deprotonation at the hydroxyl group or amino group of the sugar molecule may be used as a physical quantity. That is, for the deprotonated molecular structure, the deprotonation probability is calculated using the interatomic interaction energy after the structure optimization calculation or the interatomic interaction energy before the structure optimization calculation. It is also good to do.

  Furthermore, the deprotonation probability may be calculated using information such as the type of sugar chain adjacent to the monosaccharide having a hydroxyl group and the activation energy required for bond cleavage as a physical quantity.

  Next, when the spectrum prediction apparatus 400 executes the fragment prediction program in step S506, the spectrum prediction apparatus 400 is generated by the demutation that is muted from the sugar chain structure in the input file read in step S502. Predict fragment types. Specifically, depending on the deprotonated site, the bond cleavage position predicted based on the fragment generation pattern, the fragment species generated by the bond cleavage, and the m / z of the fragment species Can be predicted.

  Next, in step S <b> 508, the spectrum prediction device 400 stores the information on the predicted fragment type of the sugar molecule in the storage unit 408. For example, the spectrum prediction apparatus 400 may store the molecular weight and the number of charges of the generated fragment species in the storage unit 408.

In step S510, the spectrum prediction apparatus 400 executes the spectrum prediction program and uses the fragment type information generated by the simulation stored in the storage unit 408, for each given m / z of the fragment type. Predict MS ⁿ spectra of sugar molecules. Specifically, the intensity of the MS ⁿ spectrum at the m / z value of the predicted fragment species is determined according to the deprotonation probability. Here, the spectrum prediction apparatus 400 may display the predicted MS ⁿ spectrum on the display unit 406.

  Thereafter, the spectrum prediction apparatus 400 ends the process 500.

  FIG. 7 shows an example of a GUI for creating sugar chain structure data in the spectrum prediction apparatus 400. The GUI 700 may be displayed on the display unit 406 according to an instruction from the CPU 402 in the spectrum prediction apparatus 400. When the user operates the input unit 404 and inputs the sugar chain structure for which the spectrum is predicted while visualizing the file, a text format file describing the sugar chain structure data described above is created. At this time, the spectrum prediction apparatus 400 can specify the position of the hydroxyl group or amino group in the sugar molecule where deprotonation can occur. This information is stored in the created file.

Example 1
FIG. 8 shows a sugar chain structure of a sugar molecule that simulates deprotonation according to the first embodiment of the present invention. The figure shows the sugar chain structure after structural optimization calculation for Laminaripentaose-PBH, and is named C1, C2, C3, C4 and C5 in order from the left unit. Each unit part is given a number as shown in FIG. The site to be deprotonated is identified by notation based on these information. For example, the hydroxyl group at the 2-position of the unit C4 shown in FIG. 9 is represented as C4_2.

  The deprotonation probability is databased in the format shown in FIG. 10, for example. This figure shows data when Latonaripentaose-PBH is subjected to deprotonation simulation by quantum calculation.

  In the example shown in the figure, the position of the hydroxyl group to be deprotonated in order from the left (“O-position”), the value of the deprotonation bond energy calculated by DFT calculation (“DFT”, unit: kcal / mol), a relative value (“dH”) based on the DFT calculation result of C2_2, a deprotonation probability, and a generated fragment type.

  Here, the value of dH is the absolute value (16362.7) of the DFT calculation result at the position of C2_2, and the other values are the difference from the value of C2_2. For example, in the deprotonation of C5_4, dH = 16362.7−16027.6 = 335.1.

  The deprotonation probability can be calculated based on the value of dH. For example, since the fragment is considered to exist more stably as the value of dH is smaller, the deprotonation probability can be set higher. In this embodiment, the deprotonation probability corresponding to the smallest value in dH is defined as 100, the deprotonation probability corresponding to the largest value is defined as 0, and the intermediate value is proportionally deprotonated. The probability of conversion was determined.

  In the measured value of Laminaripentaose-PBH, three peaks of C ions are observed in the order of the intensity of fragment C4> fragment C3> fragment C2. Therefore, attention is focused on the calculation results for fragment C2 to fragment C4.

  According to Non-Patent Document 1, C ions such as fragment C2, fragment C3, and fragment C4 are generated only when deprotonated at the 3-position. However, as is apparent from FIGS. 8 and 9, Laminaripentaose-PBH is a 1-3 bond (that is, the first position of one unit and the third position of the other unit are bonded via O). Therefore, there is no hydroxyl group or amino group at the 3-position of the unit, and deprotonation does not occur theoretically.

  Therefore, it is presumed that the 2nd and 4th positions of the unit are deprotonated and C ions are generated as a result of electron transfer, and the calculation results at the 2nd and 4th positions were also considered.

  Referring to FIG. 10 again, the deprotonation probability of fragment C2, fragment C3, and fragment C4 is the highest at 95% for deprotonation at position C4_2.

  According to the figure, the fragment species expected to be generated in response to deprotonation at C4_2 is C3.

  Next, the stabilized structure is judged against deprotonation at C4_2.

Referring to FIG. 9, it is considered that deprotonation at the 4-position forms a hydrogen bond with the hydroxyl group at the 6-position and is easy to stabilize. In the deprotonation of C4_4, the distance between C4_4 O and C3_6 CH ₃ H is 3.639 mm, and the distance between C4_4 O and C4_6 hydroxyl group H is 2.590 mm. In contrast, the distance between O in C4_2 and H of the hydroxyl group in C5_2 is 3.916３. Therefore, the structure in which C4_2 O is deprotonated is considered to be more unstable than the structure in which C4_4 O is deprotonated.

  FIG. 11 is a diagram showing a sugar chain structure when the deprotonation at the position of C4_2 is simulated.

  The “initial structure” in FIG. 8A shows the initial structure in the deprotonated state in C4_2 of the structure in FIG. 8, and the “stabilized structure” in FIG. The structure after the optimization calculation (DFT calculation) is shown.

  In the initial structure, the hydroxyl group closest to O at the position of C4_2 is the 2nd position of the unit C5, that is, the position of C5_2.

  On the other hand, in the structure after stabilization, the distance between O at the position of C4_2 and H of C5_2 is as short as 1.077 mm. With this distance, it is determined that H of C5_2 has moved in the structure after stabilization.

  Therefore, the simulation result of deprotonation has the highest deprotonation probability at the position of C4_2. However, in actuality, C5_2 H moves and C5_2 deprotonation occurs, and as a result, C4 ions are generated. Can be judged.

FIG. 12 (a) shows the measured values of the MS ² spectrum observed for Laminaripentaose-PBH fragments C2, C3 and C4. In the figure, the fact that the peak of C ions in the measured value is in the order of the intensity of C4>C3> C2 supports the above determination.

FIG. 12 (b) shows the predicted results of the MS ² spectra predicted by the simulation according to the present embodiment. The predicted results of the MS ² spectrum and the experimental values agree well, indicating that the prediction using the deprotonation probability is appropriate.

Therefore, predicting the deprotonation probability by quantum calculation and accurately predicting the experimental value of the MS ⁿ spectrum for the structure of a given sugar molecule using the pattern of the process of electron transfer → bond cleavage → fragment generation. I was able to.

(Example 2)
FIG. 13 is a diagram showing the sugar chain structure of AHSG (Alpha-2-HS-Glycoprotein) according to the second embodiment of the present invention, and FIG. 13 (a) is a diagram symbolizing each unit, and FIG. 13 (b). Is a diagram showing each unit by name.

  FIG. 14 shows the result of calculating the deprotonation probability according to the present example for AHSG. In this embodiment, the upper part of the branch shown in FIG. 13B is expressed as α chain (C6, C5, C4), and the lower part is expressed as β chain (C1, C2, C3).

  The third column of the table ("Precursor") indicates the position to be deprotonated. The arrow “→” in the same row indicates that further deprotonation occurs, for example, “C5 — 3 → α chain C6 — 4” indicates that deprotonation occurs at C5 — 3 and then deprotonation occurs at the α chain C6 — 4.

  The fourth to eighth columns show the result of calculating the binding energy when deprotonation occurs at the position described in the “Precursor” column by DFT calculation after inputting the sugar chain structure information of FIG. In order, binding energy (“BE”, unit: kcal / mol), relative binding energy (“dH”) with −20043.00 as the reference (base), dH + 40, dH + 40 after normalized values (“ Deprotonation probability ”), and a value after normalization (“ deprotonation probability (after correction) ”) in consideration of the case where deprotonation has occurred twice or more.

Looking at the “precursor” column in the table shown in FIG. 14, some fragments, such as “C5 — 3 → α chain C6 — 4”, are deprotonated two or more times. It is considered that the probability of the fragment decreases as the number of protonations increases. Therefore, for example, when “precursor” includes one “→”, the deprotonation probability is multiplied by 0.1,
Deprotonation probability (after correction) = (first deprotonation probability) × (second deprotonation probability) × 0.1
And calculate. For example, for deprotonation where “precursor” is represented by “C5 — 3 → α chain C6 — 4”,
C5_3 dH + 40.0 (after specification) = 0.74
C6_4 dH + 40.0 (after specification) = 0.85
Deprotonation probability (after correction) = 0.74 × 0.85 × 0.1 = 0.06
It can be calculated as follows.

  Similarly for the third and subsequent deprotonations, the deprotonation probability is multiplied by 0.1.

  The first and second columns of the table shown in FIG. 14 are predicted to be generated on the basis of a fragmentation pattern based on empirical rules when deprotonation occurs at the position described in the “precursor” column. Indicates / z and fragment.

FIG. 15A shows an actual measurement value of the MS ³ spectrum, and FIG. 15B shows a predicted value. Specifically, FIG. 15B is a graph of values in the first column from the right of the table shown in FIG. 14 (values after normalization considering deprotonation at two or more locations). The relationship between the relative strengths of the predicted values in the plurality of spectra approximates the measured values with high accuracy.

<Other embodiments>
In addition, you may implement this invention by making a computer perform the program which performs the process which the spectrum estimation apparatus 400 which concerns on embodiment of this invention performs.

400 Spectrum prediction device 402 CPU
404 Input unit 406 Display unit 408 Storage unit

Claims (17)

Predicting the fragment species generated by deprotonation at each site of the sugar molecule using deprotonation probabilities at each site of the sugar molecule and an empirical fragmentation pattern;
A spectrum prediction method comprising: predicting an MS ⁿ spectrum of the sugar molecule based on the deprotonation probability and information on the predicted fragment type.   Calculating the deprotonation probability at each part of the sugar molecule by simulating deprotonation at each part of the sugar molecule using information on the structure of the sugar molecule, The spectrum prediction method according to claim 1, wherein the predicting step uses the calculated deprotonation probability.   3. The spectrum prediction method according to claim 2, wherein the information on the structure of the sugar molecule includes positional information on a hydroxyl group or an amino group that can be deprotonated in the sugar molecule. The spectrum prediction method according to claim 2 or 3, wherein the information on the structure of the sugar molecule includes information on an MS ⁿ spectrum experimentally observed for the sugar molecule.   5. The spectrum prediction method according to claim 2, wherein the deprotonation simulation includes a theoretical calculation based on information on a structure of the sugar molecule.   The spectrum prediction method according to claim 5, wherein the theoretical calculation includes quantum calculation.   The spectrum prediction method according to claim 2, wherein the information related to the structure of the sugar molecule includes a physical quantity related to the sugar molecule.   The spectrum prediction method according to claim 7, wherein the physical quantity includes an atomic interaction energy or a structure stabilization energy before and after deprotonation in the hydroxyl group or amino group of the sugar molecule. Means for predicting the fragment species generated by deprotonation at each site of the sugar molecule using the deprotonation probability at each site of the sugar molecule and an empirical fragmentation pattern;
A spectrum prediction apparatus comprising: means for predicting an MS ⁿ spectrum of the sugar molecule based on the deprotonation probability and information on the predicted fragment type.   Means for calculating the deprotonation probability at each part of the sugar molecule by simulating deprotonation at each part of the sugar molecule using information on the structure of the sugar molecule, The spectrum prediction apparatus according to claim 9, wherein the predicting unit uses the calculated deprotonation probability.   The spectrum prediction apparatus according to claim 10, wherein the information related to the structure of the sugar molecule includes positional information of a hydroxyl group or an amino group that can be deprotonated in the sugar molecule. The spectrum prediction apparatus according to claim 10 or 11, wherein the information on the structure of the sugar molecule includes information of an MS ⁿ spectrum experimentally observed for the sugar molecule.   13. The spectrum prediction apparatus according to claim 10, wherein the deprotonation simulation includes a theoretical calculation based on information related to a structure of the sugar molecule.   The spectrum prediction apparatus according to claim 13, wherein the theoretical calculation includes quantum calculation.   The spectrum prediction apparatus according to claim 10, wherein the information related to the structure of the sugar molecule includes a physical quantity related to the sugar molecule.   The spectrum prediction apparatus according to claim 15, wherein the physical quantity includes atomic interaction energy or structure stabilization energy before and after deprotonation in the hydroxyl group or amino group of the sugar molecule.   A program for causing a computer to execute the spectrum prediction method according to claim 1.