Analysis system and analysis methods

Abstract

Analysis system, comprehensive: an analyzer (510) configured to separate a sample (501) containing a plurality of components labelled with M types of fluorescent substances by chromatography and to obtain initial time series data of fluorescence signals acquired in N types of wavelength bands in a state in which at least some of the plurality of components are not completely separated, where M > N; a storage unit (1204) configured to store second time series data of individual model fluorescence signals of the multitude of components, and a computer (520) that is set up to compare the first time series data with the second time series data and to determine which type of fluorescent substance out of M types of fluorescent substances labels each of the multitude of components individually.

Description

Technical area

[0001] The present invention relates to an analysis system and an analysis method. State of the art

[0002] Analytical methods in which various types of components contained in samples, including biological samples such as DNA, proteins, and cells, are labeled with various types of fluorescent substances (the correspondence is not always a one-to-one correspondence), the fluorescence emitted by the various types of fluorescent substances is identified, and the various types of components are analyzed are widely used. Examples of such analytical methods include chromatography, DNA sequencing, DNA fragment analysis, flow cytometry, PCR, HPLC, Western blot, Northern blot, Southern blot, and microscopic examination.

[0003] In general, the fluorescence spectra of a variety of fluorescent substances exhibit overlaps with each other (hereinafter referred to as spectral overlaps). The fluorescence emissions of a variety of fluorescent substances also exhibit temporal or spatial overlaps with each other (hereinafter referred to as spatiotemporal overlaps). In situations where overlaps exist, there is a need for a technique that detects and identifies fluorescence emitted by a variety of fluorescent substances.

[0004] Next, this technique is described using an example of a DNA sequencer using electrophoresis. The first DNA sequencers using electrophoresis began in the 1980s, using plate electrophoresis, and in the 1990s and later, the process shifted to capillary electrophoresis. However, the techniques that solve these problems have not fundamentally changed. Fig. 3 of Non-Patent Literature 1 shows a method based on plate electrophoresis. Note that this technique is also used in capillary electrophoresis. The basic processes of the technique consist of the following processes (1) to (4) under the condition that M ≤ N. (1) A laser beam is irradiated onto a sample while separating it by electrophoresis, the sample containing M kinds of DNA fragments labeled with M kinds of fluorescent substances to cause the fluorescent substances to emit fluorescence, the fluorescence in N colors is detected in N kinds of wavelength bands, and thus time series data of fluorescence intensities of N colors are obtained. (2) Color conversion is applied for each time point of the time series data (1) to obtain time series data of concentrations of M kinds of fluorescent substances, that is, M kinds of DNA fragments are obtained. (3) A correction due to the difference in mobility of M kinds of fluorescent substances (hereinafter referred to as mobility correction) is applied for each time point of the time series data (2) to obtain mobility-corrected time series data of the concentrations of M kinds of fluorescent substances, that is, M kinds of DNA fragments are obtained. (4) Base calling is carried out on the basis of the time series data (3).

[0005] Based on processes (1) to (4), Fig. 3 of Non-Patent Literature 1. Here, M = 4, N = 4. The following describes the case where a single sample is analyzed using a single electrophoresis channel. In the case where a plurality of samples are analyzed using a plurality of electrophoresis channels, the following processes are performed in parallel.

[0006] First, (1) is described. When preparing a copy of DNA fragments of different lengths for a template DNA using the Sanger reaction, the DNA fragments are each labeled with one of four types of fluorescent substances corresponding to the terminal base species C, A, G, and T (hereinafter, these fluorescent substances are simply referred to as C, A, G, and T). While the DNA fragments labeled with the fluorescent substances are separated by length by electrophoresis, a laser beam is sequentially irradiated on them to cause the fluorescent substances to emit fluorescence. The emitted fluorescence is detected in four types of wavelength bands: b, g, y, and r (hereinafter referred to as four-color detection). The wavelength bands can correspond to the maximum emission wavelengths of C, A, G, and T.Thus, parts of the time series data of the fluorescence intensities I(b), I(g), I(y), and I(r) of these four colors are obtained. Fig. 3A of Non-Patent Literature 1, these are pieces of time series data (also called raw data) of these colors, and I(b), I(g), I(y), and I(r) are indicated by blue, green, black, and red, respectively.

[0007] Next, (2) is described. The fluorescence intensities of four colors at each time point are expressed by the concentrations D(C), D(A), D(G), and D(T) of four types of fluorescent substances C, A, G, and T at that time point as Equation (1). [Equation 1] (l(b)l(g)l(y)l(r))=(w(bC)w(bA)w(bG)w(bT)w(gC)w(gA)w(gG)w(gT)w(yC)w(yA)w(yG)w(yT)w(rC)w(rA)w(rG)w(rT))(D(C)D(A)D(G)D(T))

[0008] Here, elements w(XY) of a 4x4 matrix W express the intensity ratio based on spectral overlap at which a fluorescent substance Y (C, A, G, or T) is detected in a wavelength band X (b, g, y, or r). w(XY) is the fixed value determined only by the properties of the fluorescent substance Y (C, A, G, or T) and the wavelength band X (b, g, y, or r) and does not change during electrophoresis.

[0009] Therefore, the concentrations of four types of fluorescent substances at each time point are determined from the fluorescence intensities of four colors at that time point as Equation (2). [Equation 2] (D(C)D(A)D(G)D(T))=(w(bC)w(bA)w(bG)w(bT)w(gC)w(gA)w(gG)w(gT) w(yC)w(yA)w(yG)w(yT)w(rC)w(rA)w(rG)w(rT))−1(l(b)l(g)l(y)l(r))

[0010] As described above, the fluorescence intensities of four colors are calculated using the inverse matrix W -1multiplied, thus resolving the spectral overlap (hereinafter, this process is referred to as color conversion). Thus, the concentrations of four types of fluorescent substances, C, A, G, and T, i.e., portions of the time series data of the concentrations of DNA fragments with four types of bases at the ends, are obtained.

[0011] Fig. Figure 3B of Non-Patent Literature 1 shows this time series data. Color conversion is possible regardless of the presence or absence of space-time overlap. As shown in Fig. As shown in Figure 3B, the presence of concentrations of a variety of types of fluorescent substances at the same time indicates the presence of space-time overlap.

[0012] Fig. 3C of Non-Patent Literature 1 is a process that does not correspond to any of (1) to (4) and is not necessarily performed. The time series data in Fig. 3B are separated into individual peaks by deconvolution, i.e. each peak expresses the concentrations of DNA fragments of a certain length labeled with one of the four types of fluorescent substances C, A, G and T, and thus the space-time overlap is resolved.

[0013] Finally, (3) and (4) are described. Generally, DNA fragments of different base lengths are separated by electrophoresis at approximately regular intervals. However, the influence of the fluorescent substance labeled on the DNA fragment changes the mobility (electrophoresis speed), which sometimes makes the regular intervals uneven. Therefore, the magnitude correlation of the mobility according to the types of fluorescent substances to be labeled is checked beforehand, and the time series data are analyzed in Fig. 3B or Fig. 3C of Non-Patent Literature 1 are corrected based on size correlation information. This yields mobility-corrected time series data in which DNA fragments of different lengths are arranged at approximately regular base intervals. Fig. Figure 3D of Non-Patent Literature 1 shows these mobility-corrected time series data. Blue, green, black, and red peaks represent the concentrations of DNA fragments of a certain length with the terminal base species C, A, G, and T, respectively. Fig. In 3D, the DNA fragments are arranged in order of length at bases. Therefore, as in Fig. 3D, the terminal base species are arranged in order and thus the result of the base calling can be obtained.

[0014] In the above processes or before and after the processes, the time series data are sometimes subjected to appropriate processing such as smoothing, noise filtering, and baseline removal.

[0015] In the color conversion process in (2), as expressed by Equation (1), the unknowns of the concentrations D(C), D(A), D(G), and D(T) of four kinds of fluorescent substances C, A, G, and T are found by solving four simultaneous equations composed of the known quantities of four kinds of fluorescence detection intensities I(b), I(g), I(y), and I(r) in the wavelength bands b, g, y, and r for each of the time points. Since this is generally equivalent to solving M kinds of unknowns by N simultaneous equations, the condition M ≤ N is necessary, as described above. If M > N, no solution can be uniquely found (that is, a plurality of solutions may exist), and therefore the color conversion (Equation (2)) is not feasible.

[0016] Non-patent literature 2 deals with DNA sequencing by electrophoresis, but under the condition that M > N, where M = 4, N = 3. The fluorescence emission is detected in three wavelength bands, b, g, and r (hereinafter referred to as three-color detection), and then the fluorescence intensities in four colors at each time point in equation (1) are replaced by fluorescence intensities of three colors at each time point in equation (3). [Equation 3] (I(b)I(g)I(r))=(w(bC)w(bA)w(bG)w(bT)w(gC)w(gA)w(gG)w(gT)w(rC)w(rA)w[rG]w(rT))(D(C)D(A)D(G)D(T))

[0017] At this point, the matrix W has three rows and four columns and no inverse matrix. Thus, unlike equation (2), none of the concentrations of four types of fluorescent substances can be uniquely found. As described above, under the condition M > N, a solution cannot usually be found. However, a solution can be found through additional preconditions, as described below.

[0018] In the first step, it is assumed that there is no space-time overlap for a large number of types of fluorescent substances, that is, it is assumed that only one type of fluorescent substance emits fluorescence at a time. At this time, Y (C, A, G, or T), where the ratio of three colors is fluorescence intensities (I(b) I(g) I(r)) T and the ratio of four columns (w(bY) w(gY) w(rY)) Tthe matrix W at each of the time points. In other words, if one type of fluorescent substance is selected in equation (3) and the concentrations of the remaining three types of fluorescent substances are zero, that is, if (D(C) D(A) D(G) D(T)) T (D(C) 0 0 0) T , (0 D (A) 0 0) T , (0 0 D (G) 0) T or (0 0 0 D(T)) T is, D (Y) (Y is C, A, G, or T) for which the difference between the left and right sides is the smallest is individually found. Thus, Y (C, A, G, or T) for which the difference between the left and right sides is the smallest at this time can be selected. If the difference between the left and right sides is not sufficiently small in any case, the process proceeds to the following second step.

[0019] In the second step, it is assumed that only two types of fluorescent substances emit fluorescence at the same time. Now, if two types of fluorescent substances are selected in equation (3) and the concentrations of the remaining two types of fluorescent substances are zero, that is, if (D(C) D(A) D(G) D(T)) T (D(C) D(A) 0 0) T , (D(C) 0 D(G) 0) T (D (C) 0 0 D(T)) T , (0 D (A) D (G) 0) T , (0 D (A) 0 D(T)) T or (0 0 D(G) D(T)) T , two types of D(Y) (Y is C, A, G, or T) for which the difference between the left and right sides is the smallest are individually found. Thus, two types of Y (C, A, G, or T) can be selected for which the difference between the left and right sides is the smallest at this time.

[0020] In this way, similar to process (2) in Non-Patent Literature 1, the time series data of the concentrations of four types of fluorescent substances, i.e., four types of DNA fragments, are obtained. Afterward, processes corresponding to processes (3) and (4) in Non-Patent Literature 1 are performed, and thus the base calling result can be obtained.

[0021] In order to comply with the conditions of performing the first and second steps in Non-Patent Literature 2, that is, assuming that only one or two kinds of fluorescent substances emit fluorescence at the same time, it is necessary to maintain a small space-time overlap, that is, the following two conditions. (a) In the time series data of the concentrations of four types of DNA fragments, a large number of peaks derived from the DNA fragments of certain lengths are arranged at approximately regular intervals. (b) In the time series data of the concentrations of four kinds of DNA fragments, two adjacent peaks derived from the DNA fragments with two kinds of base lengths differing by one base length are excellently separated.

[0022] In Non-Patent Literature 2, four types of primers used in the Sanger reaction are labeled with four different types of fluorescent substances (strictly speaking, one labeling method is modified for three types of fluorescent substances), and the difference in mobility due to the influence of the fluorescent substance with which the DNA fragment is labeled is sufficiently reduced. In addition, the condition that the separation efficiency of electrophoresis is sufficiently high is met. That is, the time series data of the fluorescence intensity of three colors (I(b) I(g) I(r)) T, where assumptions (a) and (b) are met, are obtained (the upper part of Fig. 2 in non-patent literature 2). Under these conditions, the first and second steps are carried out. Thus, the time series data of the concentrations of four types of fluorescent substances, i.e., four types of DNA fragments, are obtained, and the result of base calling is obtained (the lower part of Fig. 2 in non-patent literature 2). Furthermore, US 2011 / 256631 A1 (cf. patent literature 1) describes a method for improving the accuracy of base calling in pulse sequences, e.g., in DNA sequencing. An iterative gradient boosting algorithm is used to train scoring functions based on a set of features for each significant pulse. The scoring functions are refined by testing them against known sequences, enabling the identification of events such as insertions, deletions, and correct incorporations of bases. This approach enables accurate base calling even for unknown sequences. Literature listNon-patent literature Non-patent literature 1: Genom Res. pp. 644-65, 8(6), Jun. 1998 Non-patent literature 2: Electrophoresis, pp. 1403-14, 19(8-9), Jun. 1998 Patent literature

[0023] Patent literature 1: US 2011 / 256 631 A1 Brief description of the inventionTechnical task

[0024] An RGB color sensor is a two-dimensional sensor with three types of pixels arranged in rows. The sensor is designed to detect the wavelength bands of red (R), green (G), and blue (B), which correspond to the three primary colors that human eyes can identify. The RGB color sensor is used not only in digital SLR cameras and digital compact cameras, but also in digital cameras installed on smartphones these days, and its popularity is exploding worldwide. This significantly improves performance and also significantly reduces price. Therefore, it is extremely useful for the RGB color sensor to be applied to the analysis process of detecting the fluorescence emitted by the various types of fluorescent substances, thereby identifying and detecting a variety of components.However, since the RGB color sensor can only detect three colors, it is difficult to identify the fluorescence emissions of four or more types of fluorescent substances. As described above, to identify the fluorescence emissions of M types of fluorescent substances, when M types of fluorescent substances have spectral overlap and spatiotemporal overlap, detection of N colors in N types of wavelength bands is required, where the condition M ≤ N is satisfied.

[0025] Non-patent literature 2 solves the problem for the case where the fluorescence emissions of M = four types of fluorescent substances in N = three colors are detected on the DNA sequencer using electrophoresis (i.e., M > N) by providing preconditions for space-time overlap. These are the conditions under which only one or two types of fluorescent substances emit fluorescence at the same time, i.e., conditions (a) and (b) are satisfied.

[0026] In general, however, conditions (a) and (b) are not met in many cases. Also in the time series data of fluorescence intensities of three colors (top) and the time series data of the concentrations of four types of fluorescent substances (bottom) in Fig. 2 of Non-Patent Literature 2, in the areas marked by stars, despite the separation efficiency of electrophoresis, three or more peaks, each derived from DNA fragments of a specific length, are crowded due to a phenomenon called compression. Thus, condition (a) is not met, and thus correct base calling is not achieved. In the final phase of electrophoresis in Fig. 2 of Non-Patent Literature 2, the separation efficiency of electrophoresis is reduced. Since the separation of two adjacent peaks derived from DNA fragments with two types of base lengths differing by one base length is insufficient, condition (b) is not met, and thus correct base calling is not achieved.

[0027] In Non-Patent Literature 2, condition (a) is achieved by using a primer labeling method in which four kinds of primers used in the Sanger reaction are labeled with four kinds of fluorescent substances that are different from each other. However, nowadays, instead of the primer labeling method, a terminator labeling method in which four kinds of terminators used in the Sanger reaction are labeled with four kinds of different fluorescent substances is mainly used. In the primer labeling method, it is necessary to perform the Sanger reaction using four kinds of primers separately, that is, in four different sample tubes. On the other hand, in the terminator labeling method, the Sanger reaction using four kinds of terminators can be performed jointly, that is, in one sample vessel.Therefore, the terminator labeling method can significantly simplify the Sanger reaction.

[0028] However, in the primer labeling method, the difference in mobility of DNA fragments labeled with four types of fluorescent substances is small, while in the terminator labeling method, the difference in mobility of DNA fragments labeled with four types of fluorescent substances is large, thus condition (a) is not necessarily met. That is, not necessarily only one or two types of fluorescent substances emit fluorescence, and there is a possibility that sometimes three or more types of fluorescent substances emit fluorescence at the same time. Therefore, it is difficult to perform DNA sequencing by the method of Non-Patent Literature 2 when at least the terminator labeling method is used.

[0029] Therefore, the present invention is to provide an analysis technique for detecting M kinds of components by detecting N colors in N kinds (M > N) of wavelength bands in the state where the fluorescence emitted from M kinds of fluorescent substances has a spectral overlap and a space-time overlap. Solution to the task

[0030] To solve the problem, the configurations described in the claims are adopted, for example. The present application includes several schemes that solve the problem.For example, there is provided an analyzer configured to chromatographically separate a sample having a plurality of components labeled with the M types of fluorescent substances and acquire first time series data of fluorescent signals obtained in N types (M>N) of wavelength bands in a state where at least a part of the plurality of components is not completely separated; a storage unit configured to store second time series data of individual model fluorescent signals of the plurality of components; and a computer configured to compare the first time series data with the second time series data and determine which types of fluorescent substances of the M types of fluorescent substances individually label the plurality of components.

[0031] According to another example, there is provided an analysis method comprising: chromatographically separating a sample having a plurality of components labeled with the M types of fluorescent substances, and acquiring first time series data of fluorescence signals obtained in N types (M>N) of wavelength bands in a state where at least a part of the plurality of components is not completely separated; and determining which types of fluorescent substances of the M types of fluorescent substances individually label the plurality of components by comparing the first time series data with second time series data of individual model fluorescence signals of the plurality of components. Advantageous effects of the invention

[0032] According to the present invention, even when M>N, M types of components can be detected in the state where the fluorescence emitted by M types of fluorescent substances has spectral overlap and space-time overlap. Note that other features related to the present invention will become apparent from the description of the present specification and the accompanying drawings. Objects, configurations, and effects other than those described above will become apparent from the description of the embodiments below. Short description of the figures Fig. 1 is a diagram showing a DNA sequencing method according to Non-Patent Literature 1 using model data. Fig. Figure 2 is a diagram showing a DNA sequencing method according to Non-Patent Literature 2 using model data. Fig. 3 is a diagram showing an example of a DNA sequencing method using model data according to a first embodiment. Fig. 4 is a diagram showing another example of a DNA sequencing method using model data according to a first embodiment. Fig. 5 is a diagram showing process steps and the configuration of a system according to the first embodiment. Fig. 6 is a diagram showing process steps and the configuration of a system according to the first embodiment. Fig. 7 is a diagram showing process steps and the configuration of a system according to a second embodiment. Fig. 8 is a diagram showing process steps and the configuration of a system according to a third embodiment (M = 4, N = 3). Fig. 9 is a diagram showing process steps and the configuration of a system according to a fourth embodiment (M = 4, N = 2). Fig. Figure 10 is a block diagram of a multicapillary electrophoresis apparatus. Fig. Figure 11 is a block diagram of a multicolor detection system of a multicapillary electrophoresis apparatus. Fig. 12 is a block diagram of a computer. Fig. Figure 13 shows a fifth embodiment in which DNA sequencing is performed using the process steps and configurations of the Fig. 9 to 12 shown system. Fig. 14 is a graph showing time series data of two color fluorescence intensities of model peaks of DNA fragments of certain length labeled with four kinds of fluorescent substances according to the fifth embodiment. Fig. 15 is a diagram illustrating the process of DNA sequencing according to the fifth embodiment. Fig. 16 is a diagram illustrating the process of DNA sequencing according to the fifth embodiment. Fig. Figure 17 is a graph showing the terminal base species, accuracy of fit and QV (Quality Value) for model peaks in Fig. 15 (5) in the chronological order of electrophoresis. Fig. Figure 18 is a graph showing the terminal base species, accuracy of fit and QV (Quality Value) for model peaks in Fig. 15 (6) in the chronological order of electrophoresis after correction. Description of the embodiments

[0033] Embodiments of the present invention will be described below with reference to the accompanying drawings. It should be noted that the accompanying drawings illustrate specific embodiments according to the principle of the present invention. However, these drawings are provided for the purpose of understanding the present invention and are in no way intended to limit the interpretation of the present invention.

[0034] The following embodiments relate to a device that detects fluorescence to be identified from a sample containing a plurality of components labeled with a plurality of fluorescent substances, and thus analyzes the components. The embodiments below are applicable, for example, to the fields of chromatography, DNA sequencing, DNA fragment analysis, flow cytometry, PCR, HPLC, Western blot, Northern blot, Southern blot, microscopic examination, and any other method.

[0035] If DNA sequencing is performed by electrophoresis, the contents of Non-Patent Literatures 1 and 2 are checked based on the Fig. 1 to 4 are described in more detail.

[0036] Fig. 1 shows a method of Non-Patent Literature 1. Fig. Figure 1 (1) shows time series data of fluorescence intensities of four colors I(b), I(g), I(y), and I(r) obtained by four-color detection of the fluorescence emissions of four kinds of fluorescent substances C, A, G, and T in four kinds of wavelength bands b, g, y, and r. The horizontal axis in Fig. 1 expresses time, and the vertical axis represents fluorescence intensity.

[0037] Fig. 1 (W) shows the elements w(XY) in four rows and four columns of a matrix W. For example, four black bar charts from the left express w(bC), w(gC), w(yC), and w(rC). Similarly, horizontally striped bar charts express w(XA) (X is b, g, y, and r), diagonally striped bar charts w(XG) (X is b, g, y, and r), and checkered bar charts express w(XT) (X is b, g, y, and r). The charts w(bY), w(gY), w(yY), and w(rY) are normalized so that w(bY) + w(gY) + w(yY) + w(yY) + w(rY) = 1 (Y is C, A, G, or T).

[0038] Specifically, the matrix W and the inverse matrix W -1 the matrix W as follows. [Equation 4] W=(w(bC)w(bA)w(bG)w(bT)w(gC)w(gA)w(gG)w(gT)w(yC)w(yA)w(yG)w(yT)w(rC)w(rA)w( rG)w(rT))=(0.560.260.160.090.280,430.240.180.110,220.400.270.060.090.200.45) W−1=(w(bC)w(bA)w(bG)w(bT)w(gC)w(gA)w(gG)w(gT)w(yC)w(yA)w(yG)w(yT)w(rC)w(rA)w(rG)w (rT))−1=(0.55−1.440.270.23−1.764.27−1.68−0.340.33−2.124.642.00−0.120.29−1.693.12)

[0039] I(b), I(g), I(y) and I(r) at any time in Fig. 1 (1) are combined with the inverse matrix W -1 the matrix W (i.e. by color conversion) and thus the time series data of the concentrations D(C), D(A), D(G) and D(T) of four kinds of fluorescent substances C, A, G and T, i.e. four kinds of DNA fragments with the base species C, A, G and T at the ends, as shown in Fig. 1 (2) shown.

[0040] In Fig. 1 (2), four peaks of C, A, G and T are obtained. The heights of the peaks, that is, the concentrations (arbitrary unit), are D (C) = 100, D (A) = 80, D (G) = 90 and D (T) = 80. The times (arbitrary unit) of the peaks are 30, 55, 60 and 75. Although the peak of A and the peak of G have a large space-time overlap, the color conversion can correctly find each of the concentrations even in this case.

[0041] Fig. 1 (3) shows mobility-corrected time series data of the concentrations D(C), D(A), D(G) and D(T) of four kinds of fluorescent substances or DNA fragments, which were obtained by compensating for the differences in mobility due to the kinds of fluorescent substances to be labeled, which were calculated in advance in the results in Fig. 1 (2). In particular, since it is known that the mobility of the DNA fragments labeled with the fluorescent substance A relative to the mobilities of the DNA fragments labeled with the other fluorescent substances is reduced to such an extent that the peak detection time of the DNA fragments labeled with the fluorescent substance A is delayed by a duration (arbitrary unit) of 10 relative to the peak detection time of the DNA fragments labeled with the other fluorescent substances. Therefore, a mobility correction is carried out in which the peak detection time of the DNA fragments labeled with the fluorescent substance A is preceded by 10. That is, the detection time of the peak of A is Fig. 1 (2) from 55 to 45. As a result of this mobility correction, time series data are obtained in which DNA fragments of different lengths are arranged at approximately regular base intervals.

[0042] Fig. 1 (4) shows the results of base calling based on Fig. 1 (3). It is nice that the labeling fluorescent substance species to which the peaks belong, or the terminal base species of the DNA fragments, are read out in chronological order.

[0043] Fig. 2 shows a method of Non-Patent Literature 2. Fig. Figure 2 (1) shows time series data of fluorescence intensities of three colors I(b), I(g) and I(r) obtained by three-color detection of the fluorescence emissions of four kinds of fluorescent substances C, A, G and T in three kinds of wavelength bands b, g and r.

[0044] In Fig. 2 (1) only the time series data of I(y) are from Fig. 1 (1) is removed, and the time series data of I(b), I(g) and I(r) are the same in both figures.

[0045] Fig. Figure 2(W) shows the elements w(XY) in three rows and four columns of a matrix W. For example, three black bar charts from the left express w(bC), w(gC), and w(rC). Similarly, horizontally striped bar charts express w(XA) (X is b, g, and r), diagonally striped bar charts w(XG) (X is b, g, and r), and checkered bar charts express w(XT) (X is b, g, and r). w(bY), w(gY), and w(rY) are normalized such that w(bY) + w(gY) + w(rY) = 1 (Y is C, A, G, or T).

[0046] Specifically, the matrix W looks like this: [Equation 5] W=(w(bC)w(bA)w(bG)w(bT)w(gC)w(gA)w(gG)w(gT)w(rC)w(rA)w(rG)w(rT))=(0.630.330.270.130.310,560.400.250.060,110.400.63) As a first step, Fig. 2 (1) using Fig. 2 (W) considers the case where only one type of fluorescent substance emits fluorescence at a time. In Fig. 2 (1) two peaks observed on the left and right side show ratios of the fluorescence intensity of the colors (I(b) I(g) I(r)) T close to the ratios of (0.63 0.31 0.06) T or (0.13 0.25 0.63) T Thus, these two peaks can be determined as individual peaks of C and T. The heights of the peaks of these C and T, i.e., the concentrations (arbitrary unit), are D(C) = 100 and D(T) = 80, and the times (arbitrary unit) of the peaks of these C and T are 30 and 75, respectively.

[0047] On the other hand, the one in the middle of Fig. 2 (1) observed peak a ratio of the fluorescence intensity of the three colors (I(b) I(g) I(r)) T which does not have a ratio of (w(bY) w(gY) w(rY)) T (Y is C, A, G, or T). Therefore, as a second step, using Fig. 2 (W) considers the case where only two kinds of fluorescent substances emit fluorescence. Here, the solution can be derived that the peaks of A and G are detected simultaneously. In particular, when the heights of the peaks, i.e., the concentrations (arbitrary unit) D(A) = 80 and D(G) = 90 and the times (arbitrary unit) of the peaks are 55 and 55, the difference between the left and right sides in equation (3) is reduced, and thus the peaks in the middle of Fig. 2 (1). From the above results, time series data of the concentrations D(C), D(A), D(G) and D(T) of four kinds of fluorescent substances C, A, G and T, i.e. four kinds of DNA fragments with the base species C, A, G and T at the ends, are obtained, as shown in Fig. 2 (2).

[0048] Similar to Fig. 1 (3) shows Fig. 2 (3) Mobility-corrected time series data of the concentrations D(C), D(A), D(G), and D(T) of four types of fluorescent substances or DNA fragments. The acquisition time (arbitrary unit) of the peak of A is preceded by 10 and corrected from 55 to 45. To equalize the peak intervals, the acquisition time (arbitrary unit) of the peak of G is shifted back by five and corrected from 55 to 60. As a result, Fig. 2 (3) and Fig. 1 (3) equal. Fig. 2 (4) shows the results of a base calling similar to Fig. 1 and equal to Fig. 1 (4).

[0049] On the other hand, Fig. 2, that processes of Fig. 2 (1) to Fig. 2 (2) cannot be determined in a specific way. Although the first step is the same as the above processes, in the second step a different solution is derived and Fig. 2 (2)' instead of Fig. 2 (2). Namely, the solution can be derived that the peaks of A and T are detected simultaneously. In particular, when the heights of the peaks, ie the concentrations (arbitrary unit) D(A) = 107 and D(T[[G]]) = 34 and the times (arbitrary unit) of the peaks are 55 and 55, the difference between the left and right sides in equation (3) is reduced, and thus the peaks in the middle of Fig. 2 (1). From the above results, time series data of the concentrations D(C), D(A), D(G) and D(T) of four kinds of fluorescent substances C, A, G and T, i.e. four kinds of DNA fragments with the base species C, A, G and T at the ends, are obtained, as shown in Fig. 2 (2)'.

[0050] Because in Fig. 2 (2)' but G is not detected, D(G) = 0. Similar to Fig. 1 (3) shows Fig. 2 (3)' Mobility-corrected time series data of the concentrations D(C), D(A), D(G), and D(T) of four types of fluorescent substances or DNA fragments. The acquisition time (arbitrary unit) of the peak of A is prefixed with 10 and corrected from 55 to 45. Fig. 2 (4)' shows the results of the base calling based on Fig. 2 (3)' was carried out. Although the same Fig. 2 (1) and Fig. 2 (W) are used, differ Fig. 2 (2), Fig. 2 (3) and Fig. 2 (4) completely from Fig. 2 (2)', Fig. 2 (3)' and Fig. 2 (4)', and different base calling results are derived. For the model data used here, Fig. 2 (4) correct base calling results, since the Fig. 1 (4) presented base calling results are a positive solution, but Fig. 2 (4)' shows incorrect base calling results.

[0051] Therefore, the method of Non-Patent Literature 2 may be able to derive a plurality of solutions from the same measurement results, and there is a risk of deriving incorrect base-calling results or generally incorrect analysis results. First embodiment

[0052] Fig. 3 is a diagram showing a DNA sequencing method using model data according to a first embodiment. Fig. 3 (1) shows time series data of fluorescence intensities of three colors I(b), I(g), I(y) and I(r) obtained by three-color detection of the fluorescence emissions of four kinds of fluorescent substances C, A, G and T in three kinds of wavelength bands b, g and r, which is equal to Fig. 2 (1) is.

[0053] Fig. Figure 3 (5) shows time series data of the fluorescence intensities of three colors I(b), I(g), and I(r) of a model peak obtained when DNA fragments of a single length labeled with the fluorescent substance C are detected in three colors. The vertical axis in Fig. 3 (5) expresses the fluorescence intensity, and the horizontal axis is time. The data in Fig. 3 (5) expresses a temporal change in the fluorescence intensity ratio between the model peaks of three fluorescent colors (b, g, and r) when detecting DNA fragments of a single length labeled with fluorescent substance C. Here, the fluorescence intensity ratio of the three colors at any time is (w(bC) w(gC) w(rC)) T = (0.63 0.31 0.06) T of the matrix W in equation (6). Likewise, the Fig. 3 (6), (7) and (8) Time series data of the fluorescence intensities of three colors I(b), I(g) and I(r) of model peaks when DNA fragments of a single length labeled with the fluorescent substances A, G and T are detected in three colors. The fluorescence intensity ratios of the three colors in the Fig. 3 (6), (7) and (8) are each (0.33 0.56 0.11) T , (0.27 0.40 0.33) T , (0.13 0.25 0.63) T .

[0054] The shape of each of the model peaks here is Gaussian, and the dispersion of the Gaussian distribution corresponds to the spatial dispersion of DNA fragments of a single length observed in experiments. Note that the shapes of the model peaks are not restrictive for this example, and the shapes of the model peaks can have other configurations. Here, the model peaks are shown in the Fig. 3 (5), (6), (7) and (8) only by changing the heights and times of the model peaks to the time series data of Fig. 3 (1). An example of the adaptation process is given with reference to Fig. 3. For example, the fitting process is carried out step by step from the left end of the data of Fig. 3 (1). For example, the height (fluorescence intensity) and the median (electrophoresis time) of the Gaussian distribution are Fig. 3 (5) varies by the difference between the fitted curve (the varied Gaussian distribution) and the time series data of Fig. 3(1). Then, the fit is achieved when the difference is smaller than a predetermined error. That is, here, the fit is determined when the fluorescence intensity and electrophoresis time of the fitted curve best match those of the time series data of Fig. 3(1). Note that the fitting can be performed while also changing the width of the Gaussian distribution. In the case that in Fig. 3 (5) no fitting is achieved by the Gaussian distribution, it is determined whether the fitting can be carried out using other data (Gaussian distribution in Fig. 3 (6), (7) and (8)) or the combination of other data. As described above, the adjustment to the data of Fig. 3 (1) using any data of the Fig. 3 (5), (6), (7) and (8) or this combination of these data parts. The error in the fitting process can be based, for example, on the difference between the peak shapes of Fig. 3 (1) and the shapes of the fitted model peaks. Various publicly known methods can be used to calculate the error. It should be noted that, for efficiency, the fitting should preferably be carried out stepwise, for example, from the end of the data of Fig. 3 (1). The reason for this is that since the tail of a neighboring fluorescence peak runs into a particular fluorescence peak, performing the fitting from the end of the data efficiently allows for the fitting taking into account the run-in of the neighboring peak.

[0055] Fig. Figure 3 (2) shows the results of performing the fitting process. At this time, the peak shape of C represents time series data where the height of the time series data of I(b) is multiplied by 1 / w(gC) = 1 / 0.63 = 1.59. The peak shape of A represents time series data where the height of the time series data of I(g) is multiplied by 1 / w(gA) = 1 / 0.056 = 1.79. The peak shape of G represents time series data where the height of the time series data of I(g) is multiplied by 1 / w(gG) = 1 / 0.04 = 2.50. The peak shape of T represents time series data where the height of the time series data of I(r) is multiplied by 1 / w(gT) = 1 / 0.63 = 1.59.

[0056] Based on these results, Fig. 3 (2) the same as Fig. 1 (2). A procedure for further obtaining the Fig. 3 (3) and (4) is further similar to the procedure for obtaining the Fig. 1 (3) and (4). Since the process of Fig. 3 (2) from Fig. 3 (1), correct base calling results can be obtained according to the embodiment as shown in Fig. 3 (4).

[0057] In the case of non-patent literature 2 in Fig. 2, in which Fig. 2 (2) or Fig. 2 (2)' from Fig. 2 (1), only the matrix W shown in equation (6) is used, ie the fluorescence intensity ratios of the detection of the three colors of the fluorescence emissions of the fluorescent substances. On the other hand, in the Fig. 3 shown embodiment in the derivation of Fig. 3 (2) from Fig. 3 (1) In addition to the matrix W shown in equation (6), that is, the fluorescence intensity ratios of the detection of the three colors of the fluorescence emissions from the fluorescent substances, the peak shapes of the fluorescence emissions from the fluorescent substances are used, that is, information about the temporal change. These differences lead to the difference between whether the solution can be clearly derived, that is, whether correct base calling results can be obtained.

[0058] Fig. 4 shows an example of the case that the Fig. 3 is further restricted to a two-color capture. Fig. 4 (1) shows time series data of fluorescence intensities of two colors I(b) and I(r) obtained by two-color detection of the fluorescence emissions of four kinds of fluorescent substances C, A, G and T in two kinds of wavelength bands b and r, and it is the time series data of Fig. 3 (1) from which the time series data of I(g) were removed.

[0059] Fig. 4 (5) shows time series data of the fluorescence intensities I(b) and I(r) of two colors of a model peak when DNA fragments of a certain length labeled with the fluorescent substance C are detected in two colors, and presents the time series data in Fig. 3 (5), from which the time series data of I(g) are removed. Likewise, Fig. 4 (6), (7) and (8) show time series data of the fluorescence intensities I(b) and I(r) of two colors of a model peak when DNA fragments of a certain length labeled with fluorescent substances A, G and T are detected in two colors, and represent the time series data of Fig. 3 (6), (7) and (8), from which I(g) are removed.

[0060] Here the model peaks are shown in the Fig. 4 (5), (6), (7) and (8) only by changing the heights and times of the model peaks to the time series data of Fig. 4 (1) adapted. Fig. Figure 4 (2) shows the results of performing the fitting process. The peak shape of C at this time represents time series data where the height of the time series data of I(b) is multiplied by 1 / w(bC) = 1 / 0.63 = 1.59. The peak shape of A represents time series data where the height of the time series data of I(b) is multiplied by 1 / w(bA) = 1 / 0.33 = 3.03. The peak shape of G represents time series data where the height of the time series data of I(r) is multiplied by 1 / w(rG) = 1 / 0.33 = 3.03. The peak shape of T represents time series data where the height of the time series data of I(r) is multiplied by 1 / w(rT) = 1 / 0.63 = 1.59.

[0061] Based on these results, Fig. 4 (2) equal Fig. 1 (2). A procedure for further obtaining the Fig. 4 (3) and (4) is similar to the procedure for obtaining the Fig. 1 (3) and (4). Since the process of Fig. 4 (2) from Fig. 4 (1), according to the embodiment is uniquely determined, correct base calling results can be obtained as in Fig. 4 (4).

[0062] Fig. 5 shows process steps and a configuration of a system according to the embodiment. The system according to the first embodiment includes an analyzer 510, a computer 520, and a display device 530. The analyzer 510 is, for example, a liquid chromatography device. The computer 520 can be achieved using, for example, a standard computer. The processing unit of the computer 520 can be achieved as functions of programs executed on a computer. The computer includes at least a processor such as a CPU (Central Processing Unit) and a storage unit such as a memory. The process of the computer 520 can be achieved by storing the processes corresponding to the program codes in the memory, and the processor executing the program codes.

[0063] In this configuration, the analyzer 510 separates a sample having a plurality of components labeled with M types of fluorescent substances (sample 501 labeled with M fluorescent substances) by chromatography and obtains first time series data of fluorescence signals acquired in N types (M>N) of wavelength bands in a state where at least a part of the plurality of components is not completely separated. The first time series data of the fluorescence signal corresponds to the time series data of the detection of N colors 513 of a sample 501 labeled with M fluorescent substances, as described below. The computer 520 includes a storage unit (e.g., a memory and a hard disk). The storage unit prestores second time series data of individual model fluorescence signals of the plurality of components.The second time series data of the individual model fluorescent signals of the plurality of components corresponds to the time series data of detecting N colors 541 of a single peak of each of the components labeled with M fluorescent substances, as described below. The computer 520 compares the first time series data with the second time series data and determines which type of fluorescent substances of M types of fluorescent substances individually label each of the plurality of components. The display device 530 displays third time series data of concentrations of M types of fluorescent substances that contribute to the fluorescent signals. The third time series data of the concentrations of the fluorescent substances corresponds to the time series data of the concentration of the M fluorescent substances 523 described below. The processes are described in more detail below.

[0064] First, the sample 501 labeled with M fluorescent substances, which contains the plurality of components labeled with M types of fluorescent substances, is injected into the analyzer 510. Then, in the analyzer 510, a separation analysis process 511 of the plurality of components contained in the sample 501 is performed. The analyzer 510 detects fluorescence emissions from M types of fluorescent substances in N types (M>N) of wavelength bands (detection of N colors) and acquires the detection time series data of the N colors (fluorescence detection time series data) 513 of the sample labeled with M fluorescent substances. At this time, the plurality of components are not always excellently separated. That is, the fluorescence of a portion of different components labeled with different fluorescent substances is detected in a state of space-time overlap.The analyzer 510 outputs the time series data from the acquisition of the N colors 513 to the computer 520.

[0065] Subsequently, the computer 520 receives as input information the time series data of the detection of N colors 513 and the time series data of the detection of N colors 541 of the individual peak of each of the components labeled with M fluorescent substances. The time series data of the detection of N colors 541 of the individual peak of each of the components labeled with M fluorescent substances are data that, for example, Fig. 3 (5), (6), (7), and (8). Note that the time series data of the acquisition of N colors 541 of the individual peak of each of the components labeled with M fluorescent substance are stored in advance in a first database 540.

[0066] Subsequently, the computer 520 performs a comparison analysis process 521 between the time series data of detecting N colors 513 and the time series data of detecting N colors 541 of the individual peak of each component labeled with M fluorescent substances. As a result, the computer 520 obtains the time series data of the concentration of M fluorescent substances 523, that is, the time series data of the detected concentrations of M types of fluorescent substances, that is, the concentrations of components labeled with M types of fluorescent substances. The time series data of the concentration of M fluorescent substances 523 is data that, for example, Fig. 3 (2). Finally, the display device 530 performs a display process 531 of the time series data of the concentration of the M fluorescent substances 523.

[0067] Fig. 6 is a diagram showing the comparison analysis on the computer 520 in Fig. 5. The computer 520 performs a fitting analysis process 522 on the N-color detection time series data 513 using the N-color detection time series data 541 of the individual peak of each of the components labeled with M fluorescent substances as a comparative analysis. As a result, the computer 520 obtains fitting error data (or fitting accuracy data) 524 based on the difference between the N-color detection time series data 513 [

[541] ] and their fitting result together with the M fluorescent substance concentration time series data 523. The display device 530 performs the display process 531 for the M fluorescent substance concentration time series data 523 and / or the fitting error data 524. Second embodiment

[0068] Fig. Figure 7 shows the process steps and a system configuration for the case where the present invention is applied to electrophoresis analysis of DNA fragments. A variety of components that are analytical targets may be nucleic acid fragments of different lengths or compositions, and the chromatography may be electrophoresis.

[0069] An analyzer 510 is an electrophoresis device. First, the DNA sample 502 labeled with M fluorescent substances, which contains the plurality of components labeled with M types of fluorescent substances, is injected into the analyzer 510. Then, a separation analysis process 512 of the plurality of types of DNA fragments contained in the DNA sample is performed in the analyzer 510. The analyzer 510 detects the fluorescence emissions of M types of fluorescent substances in N types (M>N) of wavelength bands (detecting N colors) and obtains time series data of the detection of N colors 513. In this case, the plurality of types of DNA fragments are not always excellently separated. That is, the fluorescence of a portion of different types of DNA fragments labeled with different fluorescent substances is detected in a state of space-time overlap.The analyzer 510 outputs the time series data from the acquisition of N colors 513 to the computer 520.

[0070] Next, the computer 520 receives as input information the time series data of detecting N colors 513 and the time series data of detecting N colors 541 of a single peak of each of the components labeled with M fluorescent substances, that is, the time series data of detecting N colors of a single component labeled with any one of M types of fluorescent substances. Next, the computer 520 performs a comparison analysis process 521 between the time series data of detecting N colors 513 and the time series data of detecting N colors 541 of the single peak of each of the components labeled with M fluorescent substances. More specifically, the computer 520 performs a fitting analysis process 522 on the time series data of detecting N colors 513 and the time series data of detecting N colors 541 of the single peak of each of the components labeled with M fluorescent substances.As a result, the computer 520 receives the time series data of the concentration of M fluorescent substances 523, that is, the time series data of the detected concentrations of M types of fluorescent substances, that is, the concentrations of DNA fragments labeled with M types of fluorescent substances. At the same time, the computer 520 receives fitting error data (or fitting accuracy data) 524.

[0071] In this case, the mobility of DNA fragments is influenced by M types of fluorescent substances to be labeled by electrophoresis. Therefore, to reduce this influence, the computer 520 performs a process using mobility difference data 551 of the components labeled with M fluorescent substances, which indicates the difference in mobility due to M types of fluorescent substances to be labeled. The mobility difference data 551 of the components labeled with M fluorescent substances is stored in advance in a second database 550.The computer 520 performs a mobility correction process 525 on the time series data of the concentration of M fluorescent substances 523 using the mobility difference data 551 of the components labeled with M fluorescent substances, and obtains corrected time series data of the concentration of M fluorescent substances 526 (hereinafter referred to as corrected data). The corrected data 526 is data that, for example, Fig. 3 (3).

[0072] Finally, the display device 530 performs a display process 531 for part or all of the time series data of the concentration of M fluorescent substances 523, the fitting error data (or fitting accuracy data) 524, and the corrected time series data of the concentration of M fluorescent substances 526. Third embodiment

[0073] Fig. Figure 8 shows the process steps and a system configuration when the present invention is applied to DNA sequencing by electrophoresis. Fig. 8, N = 3 and M = 4. An analyzer 510 is a DNA sequencer.

[0074] First, a DNA sequencing sample 503 labeled with four fluorescent substances is prepared. The DNA sequencing sample 503 labeled with four fluorescent substances contains four types of DNA fragments prepared by a Sanger method using a target DNA as a template and labeled with four types of fluorescent substances corresponding to four types of terminal base species. The DNA sequencing sample 503 labeled with four fluorescent substances is injected into the analyzer 510. Then, in the analyzer 510, an electrophoresis separation analysis process 512 is performed on four types of DNA fragments contained in the DNA sequencing sample. The analyzer 510 detects fluorescence emissions from four types of fluorescent substances in three types of wavelength bands and obtains time-series data of the three-color detection 513. In this process, four types of DNA fragments are not always excellently separated.That is, the fluorescence of a portion of DNA fragments of different lengths labeled with different fluorescent substances is detected in a state of space-time overlap. Analyzer 510 outputs the time-series data of three-color acquisition 513 to computer 520.

[0075] Next, the computer 520 receives as input information the time series data of three-color detection 513 and the time series data of N-color detection 541 of a single peak of each of the components labeled with four fluorescent substances, that is, the time series data of three-color detection of the DNA fragments of a single length labeled with the four types of fluorescent substances. The time series data of three-color detection 541 of the single peak of each of the components labeled with four fluorescent substances is stored in advance in a first database 540. The computer 520 performs a comparison analysis between the time series data of three-color detection 513 and the time series data of three-color detection 541 of the single peak of each of the components labeled with four fluorescent substances.Specifically, the computer 520 performs a fitting analysis process 522 on the three-color acquisition time series data 513 of the DNA sequencing sample labeled with four fluorescent substances, using the three-color acquisition time series data 541 of the individual peak of each of the components labeled with four fluorescent substances. As a result, the computer 520 obtains the time series data of the concentration of M fluorescent substances 523, that is, the time series data of the detected concentrations of four types of fluorescent substances, that is, the concentrations of four types of DNA fragments that have different terminal base species and that are labeled with four types of fluorescent substances. At the same time, the computer 520 obtains the fitting error data (or fitting accuracy data) 524.

[0076] In this case, the mobility of DNA fragments through electrophoresis is influenced by four types of fluorescent substances to be labeled. Therefore, to reduce this influence, a process is performed that uses the mobility difference data 551 of the components labeled with four fluorescent substances, which indicate the difference in mobility due to the four types of fluorescent substances to be labeled. The mobility difference data 551 of the components labeled with four fluorescent substances is stored in advance in a second database 550.The computer 520 performs a mobility correction process 525 on the time series data of the concentration of the four fluorescent substances 523 using the mobility difference data 551 of the components labeled with four fluorescent substances, and obtains corrected time series data of the concentration of the four fluorescent substances 526 (hereinafter referred to as corrected data).

[0077] The computer 520 performs a DNA base sequence determination process 528 using the corrected time series data of the concentration of the four fluorescent substances 526. On the other hand, the computer 520 obtains base sequence determination error data (or base sequence determination accuracy data) 527 for each base of the determined DNA base sequence using the fitting error data (or fitting accuracy data) 524.

[0078] Finally, the display device 530 performs a display process 531 for part or all of the time series data of the concentration of the four fluorescent substances 523, the fitting error data (or fitting accuracy data) 524, the corrected time series data of the concentration of the four fluorescent substances 526, the results of the DNA base sequence determination, and the base sequence determination error data (or base sequence determination accuracy data) 527. Fourth embodiment

[0079] Fig. 9 shows the process steps and a configuration of a system in case the configuration of Fig. 8 under the conditions N = 2 and M = 4. The process steps and the configuration of the system are similar to those in Fig. 8, so the description is omitted.

[0080] Fig. 10 is a block diagram of a multi-capillary electrophoresis apparatus, which is an example of the analyzer 510. A capillary electrophoresis apparatus 100 is used, for example, as a DNA sequencer and a DNA fragment analysis device. The inside of a capillary 1 is filled with an electrophoretic separation medium containing electrolyte, and a sample injection end 2 and a sample elution end 3 of the capillary 1 are immersed in a cathode-side electrolyte solution 4 and an anode-side electrolyte solution 5, respectively. A negative electrode 6 is immersed in the cathode-side electrolyte solution 4, and a positive electrode 7 is immersed in the anode-side electrolyte solution 5. A high-voltage power supply 8 applies a high voltage to the negative electrode 6 and the positive electrode 7, thereby performing electrophoresis.

[0081] Sample injection into capillary 1 occurs by immersing sample injection end 2 and negative electrode 6 in a sample solution 9, and the high-voltage power supply 8 briefly applies a high voltage to negative electrode 6 and positive electrode 7. Sample solution 9 contains various types of components labeled with various types of fluorescent substances. After sample injection, sample injection end 2 and negative electrode 6 are immersed again in the cathode-side electrolyte solution 4, a high voltage is applied across negative electrode 6 and positive electrode 7, and electrophoresis is performed.

[0082] Negatively charged components contained in the sample solution 9, such as DNA fragments, are electrophoretically migrated from the sample injection end 2 to the sample elution end 3 in the capillary 1 in an electrophoresis direction 10 indicated by an arrow. Due to the difference in mobility due to electrophoresis, the plurality of types of components contained in the sample solution 9 are successively separated. At a position (a laser beam irradiation position 15) where the components have electrophoretically migrated by a certain distance in the capillary 1, a laser beam 12 emitted from a laser light source 11 is irradiated. As the components pass the laser beam irradiation position 15, fluorescence emission 13 is induced from the plurality of types of fluorescent substances labeled on the components.The fluorescence 13, which changes over time during electrophoresis, is measured by a multicolor detection system 14 that performs optical detection in a variety of wavelength bands. Although in . Fig. 10 only one capillary 1 is shown, an electrophoresis device with several capillaries can be used, which carries out the electrophoresis analysis in parallel using a plurality of capillaries 1.

[0083] Fig. Figure 11 shows an example of a multi-color detection system of a multi-capillary electrophoresis device. The laser beam irradiation positions 15 of a plurality of capillaries 1 are arranged at regular intervals on the same plane. The left figure in Fig. 11 is a cross-sectional view perpendicular to the main axis of the plurality of capillaries 1, and the right figure in Fig. 11 is a cross-sectional view parallel to the major axis of a given capillary 1.

[0084] The laser beam 12 is irradiated along the arrangement plane of the plurality of capillaries 1. Thus, the laser beam 12 is irradiated simultaneously onto the plurality of capillaries 1. The fluorescence 13 emitted by each of the capillaries 1 is condensed in parallel by separate lenses 16. The condensed beams directly enter a two-dimensional color sensor 17. The two-dimensional color sensor 17 is an RGB color sensor capable of three-color detection in three types of wavelength bands. The fluorescence 13 emitted by the capillaries 1 forms points at different positions on the two-dimensional color sensor 17, respectively, and thus the fluorescence 13 can be independently detected in three colors.

[0085] Fig. 12 shows an exemplary configuration of a computer 520. As shown in the Fig. 5 to 9, the computer 520 is connected to the analyzer 510. The computer 520 can not only process the Fig. 5 to 9, but also the analyzer 510. In the Fig. 5 through 9, a display device 530 and databases 540 and 550 are shown on the outside of computer 520. However, databases 540 and 550 may be contained within computer 520.

[0086] The computer 520 includes a CPU (processor) 1201, a memory 1202, a display unit 1203, a hard disk 1204, an input unit 1205, and a network interface (NIF) 1206. The display unit 1203 is, for example, a display and can be used as the display device 530. The input unit 1205 is an input device, such as a keyboard and a mouse. A user can set the data analysis conditions and the conditions for controlling the analyzer 510 via the input unit 1205. The N-color acquisition time series data 513 output from the analyzer 510 are sequentially stored in the memory 1202.

[0087] The HDD 1204 may contain the databases 540 and 550. The HDD 1204 may contain programs that perform the adaptation analysis process, the mobility correction process, the DNA base sequence determination process, and any other process of the computer 520. The process of the computer 520 may be achieved by storing program codes corresponding to the processes in the memory 1202 and having the CPU 1201 execute the program codes.

[0088] For example, the N-color acquisition time series data 541 of the single peak of each of the components labeled with M fluorescent substances stored on the HDD 1204 is stored in the memory 1202, and the CPU 1201 performs the comparison analysis process using the N-color acquisition time series data 513 and the N-color acquisition time series data 541 of the single peak of each of the components labeled with M fluorescent substances. The display unit 1203 displays the analyzed results. Note that the analyzed results can be verified against information on a network via the NIF 1206. Fifth embodiment

[0089] Fig. 13 to 18 show a fifth embodiment in which DNA sequencing is carried out using the process steps and configurations of the Fig. 9 to 12 shown system.

[0090] As a DNA sequencing sample labeled with four fluorescent substances, a sample was prepared by dissolving 3500 / 3500xL sequencing standards, BigDye Terminator v3.1 (Thermo Fisher Scientific) in 300 µL of formamide. This sample contains four types of DNA fragments with terminal base species C, A, G, and T, which are labeled with four types of fluorescent substances: dROX (maximum emission wavelength of 618 nm), dR6G (maximum emission wavelength of 568 nm), dR110 (541 nm), and dTAMRA (maximum emission wavelength of 595 nm), respectively.

[0091] There are four capillaries 1 with an outer diameter of 360 µm, an inner diameter of 50 µm, a total length of 56 cm, and an effective length of 36 cm. POP-7 (Thermo Fisher Scientific), a polymer solution, was used as the electrophoretic separation medium. During electrophoresis, capillary 1 was set to a temperature of 60 °C and the electric field strength to 182 V / cm. Sample injection was performed by electrokinetic injection with an electric field strength of 27 V / cm for eight seconds. The laser beam 12 had a wavelength of 505 nm and a power of 20 mW. A long-pass filter was used between the lens 16 and the two-dimensional color sensor to block the laser beam 12.

[0092] Fig. 13 (1) corresponds Fig. 4 (1) and represents time series data of two-color detection obtained during electrophoresis by the RGB color sensor by detecting the fluorescence 13 emitted by one of four capillaries 1. The RGB color sensor used in the embodiment can perform three-color detection with three kinds of wavelength bands r, g, and b corresponding to R (red), G (green), and B (blue). However, the fluorescence emissions of four kinds of fluorescent substances were rarely detected in the wavelength band b and only in the wavelength bands r and g. Thus, Fig. 13 (1) shows a time series of fluorescence intensity I(r) and I(g) in these two colors. The horizontal axis represents the time elapsed (electrophoresis time) from the start of electrophoresis in units of seconds, and the vertical axis represents the fluorescence intensity in arbitrary units.

[0093] Fig. 14 (1) corresponds Fig. 4 (5) and represents time series data of fluorescence intensities of two colors I(g) and I(r) of a model peak when DNA fragments with terminal base species C and a certain length, labeled with the fluorescent substance dROX, were acquired in two colors. The ratio of the fluorescence intensities of the two colors at each time point is (w(gC) w(rC)). T = (0.02 0.98) T . Likewise, Fig. 14 (2) Fig. 4 (6) and represents time series data of fluorescence intensities of two colors, I(g) and I(r), of a model peak when DNA fragments with terminal base species A and a certain length, labeled with the fluorescent substance dR6G, were detected in two colors. The ratio of the fluorescence intensities of the two colors is (w(gA) w(rA)). T = (0.50 0.50) T . Fig. 14 (3) corresponds Fig. 4 (7) and represents time-series data of fluorescence intensities of two colors, I(g) and I(r), of a model peak when DNA fragments with terminal base species G and a specific length, labeled with the fluorescent substance dR110, were acquired in two colors. The ratio of the fluorescence intensities of the two colors is (w(gG) w(rG)). T = (0.71 0.29) T . Fig. 14 (4) corresponds Fig. 4 (8) and represents time-series data of fluorescence intensities of two colors, I(g) and I(r), of a model peak when DNA fragments with terminal base species T and a specific length, labeled with the fluorescent substance dTAMRA, were acquired in two colors. The ratio of the fluorescence intensities of the two colors is (w(gT) w(rT)). T = (0.16 0.84) TThe shapes, offsets, and standard deviations of the model peaks were Gaussian, zero, and one second, respectively. These values ​​are consistent with those measured experimentally under the current electrophoresis conditions.

[0094] The time series data of the two-color acquisition of four types of model peaks that were recorded in the Fig. 14 (1) to (4) were successively fitted to the time series data of the two-color acquisition obtained by electrophoresis analysis in Fig. 13 (1). In the fitting, only medians (electrophoresis times) and heights (fluorescence intensities) of the Fig. 14 (1) to (4) were varied, while the two fluorescence intensity ratios of the two colors and standard deviations of the Gaussian distributions were preserved and the medians and heights were determined so that the errors from the time series data of the two-color acquisition in Fig. 13 (1) were minimized.

[0095] Fig. 13 (2) is the result that the model peak of G (dR110) in Fig. 14 (3) was fitted exactly to the peak of g, which is on the far left in Fig. 13 (1) was observed. Fig. 13 (3) is the result that the model peak of T (dTAMRA) in Fig. 14 (4) exactly to the peak of r on the right side of the above peak of g in Fig. 13 (1) was adapted. Fig. 13 (4) is the result that the model peak of C (dROX) in Fig. 14 (1) exactly to the peak of r on the right side of the above peak of r in Fig. 13 (1) was adapted. Fig. 13 (5) are the time series data of the two-color acquisition, which Fig. 13 (2), (3) and (4), ie time series data containing I(g) and I(r) in Fig. 13 (2), (3) and (4) added together. As in Fig. 13 (5), the time series data of the two-colour acquisition were faithfully mapped to the corresponding sections of the time series data of the two-colour acquisition in Fig. 13 (1).

[0096] The fitting error and fitting accuracy were determined as follows. The fitting error was found by dividing the standard deviation of the difference between a fitting model peak and the corresponding measured time series data of the two-color acquisition in a two-second period (twice the standard deviation of the Gaussian distribution), centered at the time of the peak of the fitting model peak, by a larger value of the measured time series data of the two-color acquisition at the time of the peak of the fitting model peak. The fitting accuracy was obtained by subtracting the corresponding fitting error from one. The fitting accuracy is 100% when a fitting model peak perfectly matches the corresponding time series data of the two-color acquisition.Then, the fitting accuracy decreases with the deviation between the fitting model peak and the measured time series data of the two-color acquisition, and becomes 0% when the deviation is greater than or equal to the larger value of the measured fluorescence intensities of the two colors. In this embodiment, fitting error and accuracy are defined as described above. However, definitions other than these are of course acceptable.

[0097] The fitting accuracy of the fitting model peak of T in Fig. 13 (3) alone was only 43.41%. However, the fitting accuracy of the fitting model peak of T when the fitting model peaks of G, T, and C were summed was 95.56%, as shown in Fig. 13 (5). This means that in the presence of space-time overlap, it is important to fit neighboring model peaks together and not just one model peak alone.

[0098] In a similar way, the Fig. 14 (1) to (4) shown time series data of two-color detection of four types of model peaks successively to all peaks of g and r in Fig. 13 (1) amended. As in Fig. 13 (6), the time series data of the two-color acquisition are faithfully mapped to the time series data of the two-color acquisition in Fig. 13 (1) were reproduced by summing all fitting model peaks.

[0099] Fig. 15 (1) shows time series data of the concentration of the fluorescent substance dR110, ie the concentration of DNA fragments of the terminal base species G, obtained by summing I(g) and I(r) in Fig. 13 (2). Fig. 15 (2) shows time series data of the concentration of the fluorescent substance dTAMRA, ie the concentration of DNA fragments of the terminal base species T, obtained by summing I(g) and I(r) in Fig. 13 (3). Fig. 15 (3) shows time series data of the concentration of the fluorescent substance dROX, ie the concentration of DNA fragments of the terminal base species C, which is calculated by summing I(g) and I(r) in Fig. 13 (4). Likewise, Fig. 15 (5) Time series data of concentrations of DNA fragments of the terminal base species C, A, G and T, which are dispersed by all fitting model peaks in Fig. 13 (6) were obtained.

[0100] Fig. Figure 17 shows a list containing the terminal base species, the accuracy of fit and the QV (Quality Value) of all model peaks in Fig. 15 (5) summarizes the time sequence of electrophoresis. QV is given by QV = -10*Log(1-S), where the fit accuracy is S. In 200 seconds from the electrophoresis time of 1100 seconds to 1300 seconds, 76 types of DNA fragments with different base lengths are measured, and their terminal base species are individually identified.

[0101] Since the fitting accuracy is on average 94.72% and the QV is on average 13.95, a highly accurate fitting is achieved. It should be noted that the computer stores 520 data in Fig. 17 and the display device 530 can calculate data in Fig. 17 can be displayed.

[0102] Fig. 15 (6) shows corrected time series data of concentrations of DNA fragments of the terminal base species C, A, G and T, obtained by applying a mobility correction to Fig. 15 (5) based on the difference in mobility due to the differences between four types of fluorescent substances. Specifically, the electrophoresis times of the medians of the fitting model peaks of the terminal base species G in Fig. 15 (5) shifted back by 1.6 seconds, the electrophoresis times of the medians of the fitting model peaks of the terminal base species T in Fig. 15 (5) were shifted forward by 1.1 seconds, and the model peaks of terminal base species C and A were not corrected. As a result, the peaks of DNA fragments with different base lengths were arranged at approximately regular intervals. In particular, the measurement order of the DNA fragments of terminal base species G and terminal base species T are reversed due to the influence of the mobility difference (i.e., long DNA fragments overtake short DNA fragments in electrophoresis), as shown in Fig. 15 (5), while it is excellently corrected, as in Fig. 15 (6).

[0103] Fig. 18 shows a list containing the terminal base species, the accuracy of fit and the QV (Quality Value) of all model peaks in Fig. 15 (6) in the corrected chronological order of electrophoresis. Fig. 18 is the list that Fig. 17, and the numerical values ​​used are the same. The arrangement of the terminal base species in Fig. 18 provides DNA sequencing results (base calling results). Base fit accuracy and base QV are indices that correlate with base calling accuracy, but are not identical to base calling accuracy. Generally, base calling accuracy is expected to be greater than the fit accuracy. In fact, the accuracy of DNA sequencing results in Fig. 18 100%. It should be noted that the computer 520 stores the data in Fig. 18 and the display device 530 displays the data in Fig. 18 can be displayed.

[0104] Fig. 16 summarizes the processes of DNA sequencing according to the embodiment. Fig. 16 (1) is the same as Fig. 13 [

[15] ] (1) and represents the time series data of the two-color acquisition obtained by electrophoresis analysis according to Fig. 4 (1) were obtained. Fig. 16 (2) is the same as Fig. 15 (5) and represents the time series data of concentrations of DNA fragments of terminal base species C, A, G and T, corresponding to Fig. 4 (2). Fig. 16 (3) is the same as Fig. 15 (6) and represents the corrected time series data of concentrations of DNA fragments of terminal base types C, A, G and T, which were determined by mobility correction in Fig. 16 (2) were received, corresponding Fig. 4 (3). Finally, Fig. 16 (4) the results of performing base calling based on the results in Fig. 16 (3), accordingly Fig. 4 (4). The base calling results in Fig. 16 (4) match the base sequences of the target DNA.

[0105] The schemes and effects of the first to fifth embodiments are summarized. According to the above embodiments, analysis methods can be provided in which M types of components are identified and detected by detecting N colors in N types (M>N) of wavelength bands in the state where the fluorescence emitted by M types of fluorescent substances has spectral overlaps and space-time overlaps. Following Non-Patent Literature 2, a scheme on a DNA sequencer using electrophoresis that detects fluorescence emissions from M = four types of fluorescent substances in N = three colors will be described below.

[0106] The analyzer 510 detects the fluorescence emissions of four types of fluorescent substances C, A, G, and T in three colors in three types of wavelength bands b, g, and r. The process is similar to that in Non-Patent Literature 2 except for the process of obtaining the fluorescence intensities of the three colors in Equation (3) at each time point. Here, the hard disk (storage unit) 1204 of the computer 520 stores model peak data, that is, time series data of fluorescence intensities of three colors of model peaks when DNA fragments of a certain length labeled with the four types of fluorescent substances C, A, G, and T are detected in three colors. The fluorescence intensity ratio of the three colors of the model peaks of the DNA fragments labeled with the fluorescent substance Y (C, A, G, or T) is (w(bY) w(gY) w(rY)) T ) T. Therefore, the model peak data contain information corresponding to the matrix W. In addition, the model peak data contain information about the shapes of the model peaks, i.e., time series information.

[0107] In Non-Patent Literature 2, one or two types of fluorescent substances emitting fluorescence are selected at each time point, and their concentrations are determined using the matrix W. On the other hand, in the aforementioned embodiments, the computer 520 performs the fitting analysis process on the time series data of the fluorescence intensities of the three colors expressed by Equation (3) using the model peak data of four types of fluorescent substances. Even if the fluorescence emitted by four types of fluorescent substances has spectral and spatiotemporal overlap, the computer 520 can perform the fitting analysis process. For example, it is not a problem if three or more types of fluorescent substances emit fluorescence at the same time.The fitting results composed of the model peak data of C, A, G, and T express the time series data of the concentrations D(C), D(A), D(G), and D(T) of C, A, G, and T. That is, although no color conversion is performed, the concentrations of four types of fluorescent substances, that is, the time series data of the concentration of four base species corresponding to Equation (2) of Non-Patent Literature 1, can be obtained with the time series data of the fluorescence intensities of the three colors. Unlike Non-Patent Literature 2, the aforementioned embodiments have significant features that utilize time series information on the concentrations of four types of fluorescent substances. After that, the computer 520 performs processes corresponding to processes (3) and (4) in Non-Patent Literature 1, and thus the computer 520 can obtain the base calling results.

[0108] According to the above embodiments, the fitting is performed on the time series data of fluorescence intensities of N colors obtained in N kinds (M>N) of wavelength bands by detecting N colors in the state where the fluorescence emitted from M kinds of fluorescent substances has spectral overlaps and space-time overlaps, using the model peak data of M kinds of fluorescent substances, and thus the time series data of the concentrations of M kinds of fluorescent substances, that is, M kinds of components, can be analyzed.

[0109] In order to perform an analysis in which M kinds of components are identified and detected by detecting N colors in N kinds of wavelength bands in the state where the fluorescence emitted by M kinds of fluorescent substances has a spectral overlap and a space-time overlap, the necessary conditions are usually M ≤ N. According to the above embodiments, an analysis can also be performed in which M kinds of components are identified and detected even when M > N. That is, the effect is exerted in which a similar analysis can be achieved by a much simpler, smaller, and lower-cost device configuration.For example, with N = three-color detection using an RGB color sensor, whose performance is improved and cost is rapidly reduced, analysis is feasible in which M = four types or more components labeled with M = four types or more fluorescent substances are identified and detected. From the above results, analysis by highly accurate and low-cost multi-color detection is possible. For example, N = three-color detection is performed using a low-cost RGB color sensor while M = four types of DNA fragments labeled with M = four types of fluorescent substances are subjected to electrophoresis by the Sanger reaction.Therefore, although the DNA fragments of different lengths labeled with M = four kinds of fluorescent substances are measured in a mixed state, the time series data of the concentrations of M = four kinds of DNA fragments can be obtained, and thus DNA sequencing can be excellently performed.

[0110] The present invention is not limited to the aforementioned embodiments, including various exemplary modifications. The aforementioned embodiments are described in detail for a better understanding of the present invention and are not necessarily limited to those having all configurations. Part of the configuration of one embodiment may be replaced with the configuration of another embodiment. The configuration of another embodiment may be added to the configuration of one embodiment. The other configurations may be added, removed, or replaced with part of the configuration of the embodiments.

[0111] The configurations, functions, processing units, and processing schemes of, for example, computer 520 may be achieved through hardware, for example, by designing part or all of it using an integrated circuit. The configurations, functions, and any other components may be achieved through software by a processor that interprets and executes programs that implement the functions. Information, such as programs, spreadsheets, and files, that achieve the functions may be stored on various types of computer-readable media. Examples of the non-transitory computer-readable media used include a flexible floppy disk, CD-ROM, DVD-ROM, hard disk, optical disk, magneto-optical disk, CD-R, magnetic tape, non-volatile memory card, and ROM.

[0112] The above descriptions depict control and information lines that are considered necessary for the description. Not all control and information lines of products are necessarily depicted. All configurations can be interconnected. List of reference symbols c Fluorescence signal detected in wavelength band c g Fluorescence signal detected in wavelength band g y Fluorescence signal detected in wavelength band y r fluorescence signal detected in wavelength band r C fluorescent substance C or base species C A fluorescent substance A or base species A G fluorescent substance G or base species G T fluorescent substance T or base species T 1 capillary 2 Sample injection end 3 End of sample elution 4 cathode-side electrolyte solution 5 anode-side electrolyte solution 6 negative electrode 7 positive electrode 8 High-voltage power supply 9 Sample solution 10 Electrophoresis direction 11 Laser light source 12 laser beam 13 Fluorescence 14 Multicolor detection system 15 Laser beam irradiation position 16 lens 17 two-dimensional color sensor 510 Analyzer 520 computers 530 display device 540, 550 database

Claims

[1] Analysis system comprising: an analyzer (510) configured to separate a sample (501) having a plurality of components labeled with M kinds of fluorescent substances by chromatography and to obtain first time series data of fluorescent signals detected in N kinds of wavelength bands in a state where at least a part of the plurality of components is not completely separated, where M > N; a storage unit (1204) configured to store second time series data of individual model fluorescence signals of the plurality of components, and a computer (520) configured to compare the first time series data with the second time series data and to determine which type of fluorescent substance of M types of fluorescent substances individually labels each of the plurality of components. [2] The analysis system according to claim 1, wherein the second time series data is data expressing a temporal change in fluorescence intensity ratios between N kinds of fluorescence of model peaks when a component labeled with each of the fluorescent substances is detected. [3] The analysis system according to claim 1, wherein the computer (520) outputs third time series data of concentrations of M kinds of fluorescent substances contributing to the fluorescence signals by fitting the second time series data to the first time series data. [4] Analysis system according to claim 3, wherein the storage unit (1204) stores mobility difference data (551) relating to differences in mobility due to differences in the M types of fluorescent substances, and the computer (520) corrects the differences in mobility in the third time series data based on the mobility difference data (551). [5] The analysis system of claim 3, wherein the computer (520) outputs fitting error data (524) or fitting accuracy data related to a difference between the first time series data and a result of fitting on each of the plurality of components. [6] The analysis system of claim 5, further comprising: a display (530) configured to display at least one of the following: the result of the adjustment with respect to each of the multitude of components, Adjustment error data (524) or adjustment accuracy data with respect to each of the plurality of components and the third time series data. [7] The analysis system of claim 1, wherein the plurality of components are nucleic acid fragments of different length or composition and the chromatography is electrophoresis. [8] Analysis system according to claim 7, wherein the plurality of components are DNA fragments terminally labeled with M = four kinds of fluorescent substances according to the terminal base species and prepared by a Sanger method using a target DNA as a template; the first time series data are time series data of fluorescence signals acquired in N = three types or two types of wavelength bands, and the computer (520) determines a base sequence of the target DNA. [9] Analytical methods, including: Separating a sample (501) having a plurality of components labeled with M kinds of fluorescent substances by chromatography and obtaining first time series data of fluorescent signals detected in N kinds of wavelength bands in a state in which at least a part of the plurality of components is not completely separated, where M > N, and Determining which type of fluorescent substance of M types of fluorescent substances individually labels each of the plurality of components by comparing the first time series data with second time series data of individual model fluorescence signals of the plurality of components. [10] The analysis method according to claim 9, wherein the second time series data is data expressing fluorescence intensity ratios between N kinds of fluorescence of model peaks when each of the fluorescent substances emits fluorescence alone. [11] The analysis method according to claim 9, wherein the determining includes outputting third time series data of concentrations of M kinds of fluorescent substances contributing to the fluorescence signals by fitting the second time series data to the first time series data. [12] The analysis method according to claim 11, wherein the determining includes correcting the differences in mobility in the third time series data based on mobility difference data (551) relating to differences in mobility due to differences in the M types of fluorescent substances. [13] The analysis method of claim 11, wherein determining includes outputting fitting error data (524) or fitting accuracy data related to a difference between the first time series data and a result of fitting each of the plurality of components. [14] The analysis method according to claim 13, further comprising: displaying at least one of the following: the result of the adjustment with respect to each of the multitude of components, Adjustment error data (524) or adjustment accuracy data with respect to each of the plurality of components and the third time series data. [15] The analysis method of claim 9, wherein the plurality of components are nucleic acid fragments of different length or different composition and the chromatography is electrophoresis. [16] The analysis method according to claim 15, wherein the plurality of components are DNA fragments terminally labeled with M = four kinds of fluorescent substances according to the terminal base species and prepared by a Sanger method using a target DNA as a template; the first time series data is time series data of fluorescence signals acquired in N = three kinds or two kinds of wavelength bands, and the determining includes determining a base sequence of the target DNA.

Images