JP2018116027A - Peak detection method free from influence of negative peak - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a peak detection method that is able to correctly detect a peak even in a chromatogram in which a negative peak is present near a peak generated in a positive direction.SOLUTION: A peak detection method comprises: a step of selecting two or more data points, serving as base points of a base line, for an unprocessed chromatogram, and formulating a base line from the data points by mathematical processing using approximation calculation; a step of removing the base line from the unprocessed chromatogram and acquiring a chromatogram free from base line change; a step of multiplying -1 only when an output value is negative or multiplying -1 only when the output value is positive, such that the chromatogram free from base line change is matched in symbol to the output value; a step of acquiring a primary differential value for the chromatogram matched in symbol to the output value, and acquiring respective pieces of time information at a peak start point and a peak end point; and a step of applying the pieces of time information at the peak start point and the peak end point in the unprocessed chromatogram, and determining a peak.SELECTED DRAWING: None

Description

  The present invention relates to a method that enables accurate peak detection even in a chromatogram having a negative peak.

  Chromatography is a method for separating and quantifying a sample containing multiple components on a column. Generally, qualitative analysis is performed based on the elution time of each component peak, and quantification is performed based on the output level of the detector. In order to perform quantitative analysis, it is important to accurately detect a peak and calculate the area or height of the peak. The detection of the peak is generally performed based on the amount of change in the output of the detector per time, that is, the change in the differential value of the chromatogram. A peak start point and a peak end point are determined using a predetermined value (peak start sensitivity, peak end sensitivity) as a threshold value, and a peak height, an area, and the like are calculated.

  Generally, when the peak start sensitivity (S) and peak end sensitivity (-E) are set, the point where the detector output exceeds S is the peak start point, and the detector output exceeds -E after the peak starts. Peak detection is performed using the detected point as a peak end point (see FIG. 1).

  However, when there are a positive peak and a negative peak as shown in FIG. 2 and they are close to each other, the peak may not be detected correctly. FIG. 2a shows the chromatogram and FIG. 2b shows the differential curve. In such a chromatogram, the peak detection result is as shown in FIG. A straight line is drawn from the lowest point of the negative peak to the peak end of the positive peak, and is vertically cut at the peak start of the positive peak, and is quantitatively calculated as two peaks.

  For such chromatograms that cause erroneous peak detection, there are many cases where peak start sensitivity and peak end sensitivity alone cannot be used, and correct results can be obtained using other special parameters. In some cases, the peak start point and the peak end point are corrected manually.

  An object of the present invention is to provide a peak detection method capable of accurately detecting a peak even in a chromatogram in which a negative peak exists in the vicinity of a peak generated in the positive direction.

  Common peak detection is performed based on the change of the first derivative value of the chromatogram. At the peak in the positive direction, when a peak occurs, the differential value rises from zero, reaches a maximum near the peak top, rapidly decreases, and approaches a zero through a negative minimum value. On the other hand, in the negative peak, when a peak occurs, the differential value decreases from zero, reaches a minimum near the peak top, increases rapidly, and approaches a zero through a positive maximum. Thus, a positive peak and a negative peak show completely opposite changes (see FIG. 2b). For this reason, when a positive peak and a negative peak are present in the vicinity, the peak start and peak end cannot be accurately determined, and erroneous peak detection often occurs.

  The present invention is a peak detection method for solving the above-mentioned problems, and it has been found that the peak start and the peak end can be accurately determined by matching the signs of the output values of the chromatogram.

That is, the present invention
As a first step, with respect to an unprocessed chromatogram, two or more data points serving as a base point of a baseline are selected, and a baseline is established by mathematical processing based on approximate calculation from the data points;
As a second step, subtracting the baseline from the untreated chromatogram to obtain a chromatogram without baseline fluctuations;
As a third step, the sign of the output value is obtained by multiplying the chromatogram without the baseline fluctuation by −1 only when the output value is negative or by −1 only when the output value is positive. And the process of combining
As a fourth step, for the chromatogram combined with the sign of the output value, obtaining a primary differential value, obtaining time information of the peak start point and peak end point,
As a fifth step, the time information of the peak start point and the peak end point is adapted to the unprocessed chromatogram to determine the peak,
The present invention relates to a chromatogram peak detection method comprising:

  Moreover, this invention may include the process of performing the smoothing process with respect to the chromatogram without the said baseline fluctuation | variation between a 2nd process and a 3rd process.

  Hereinafter, the peak detection method of the present invention will be described in detail, and FIG. 3 shows changes in the chromatogram after the process.

(First step)
A step of subtracting the baseline from the unprocessed chromatogram is performed.
A plurality of points are selected as base points from a region where no peak exists in an unprocessed chromatogram. Next, the selected base points are mathematically connected to obtain a virtual baseline. Mathematical processing means calculating an approximate expression of the plurality of base points (Xn: time, Yn: detector output) by the least square method or the like, or calculating a spline curve.

  When a linear expression is adopted as the approximate expression, it is preferable to select at least two base points, preferably five or more base points. When a quadratic expression is adopted as an approximate expression, it is preferable to select at least three base points, preferably five or more base points. When a cubic expression is adopted as the approximate expression, it is preferable to select at least 4 base points, preferably 6 or more base points. When using a spline curve, it is desirable to select as many base points as possible (see FIG. 4).

  The base point can be selected / set arbitrarily by the operator looking at the chromatogram, but the region where there is little change in the primary differential value can be automatically narrowed down and the base point can be selected / set automatically. The selection / setting is not limited.

(Second step)
The baseline (approximate expression) obtained in the above step is subtracted from the unprocessed chromatogram to obtain a chromatogram having no baseline fluctuation.

(Smoothing process)
A smoothing process is performed on the chromatogram having no baseline fluctuation. Although this step may be performed as necessary, it is not an indispensable step, but it is preferable to carry out this step because false detection of peaks due to the influence of noise or the like can be suppressed. In general, the smoothing process for the chromatogram is performed so that the shape of the target component peak does not change, but in the present invention, even if the peak shape slightly changes, the final quantitative calculation is hardly affected. Therefore, there is no problem even if a strong smoothing process is performed. There are various smoothing processes such as a moving average method and a low-pass filter, but the method is not limited.

(Third process)
Next, for a chromatogram with no baseline fluctuation (a smoothed chromatogram when a smoothing process is performed), only when the output value is negative, -1 is multiplied or only when the output value is positive. Multiply -1 to match the sign of the output value.

(Fourth process)
Next, a first derivative is taken with respect to the chromatogram obtained by combining the signs of the output values. Similar to a general peak detection method, the time information of the peak start point and peak end point is obtained by comparing with a preset threshold value. In FIG. 3d, the peak start sensitivity S is set to 50 (mABU / min) and the peak end sensitivity E is set to −50 (mABU / min) as thresholds.

(Fifth process)
Next, the time information of the peak start point and peak end point obtained in the above step is applied to the untreated chromatogram, and the peak start point (time, output) and peak end point in the untreated chromatogram are applied. (Time, output) is acquired, and quantitative calculation (height, area, etc.) is performed based on it. As described above, only the time information of the peak start point and peak end point is acquired by performing various treatments, and the values are applied to an unprocessed chromatogram to perform quantitative calculation (height, area, etc.) Therefore, there is no influence on the base variation or peak shape of the untreated chromatogram. Moreover, it turns out that the detection of the negative peak which is the objective of this invention is also performed reliably.

  For comparison, a first-order derivative is obtained for a general unprocessed chromatogram, compared with a preset threshold value, and the time when peak start point and peak end point time information is obtained is obtained. As shown in FIG. As described above, the threshold is set such that the peak start sensitivity S is set to 50 (mABU / min) and the peak end sensitivity E is set to -50 (mABU / min). As can be seen from FIG. 5d, the peak detection is abnormal in the negative peak region and the positive peak region appearing thereafter, which causes a problem in the quantitative calculation. Moreover, the peak start and peak end cannot be detected in the negative peak and positive peak regions after 7.5 minutes. This is because the chromatogram drifts in the positive direction.

  As described above, the peak detection method of the present invention can accurately detect a peak even in a chromatogram in which a negative peak and a positive peak exist in the vicinity by matching the sign of the output value of the chromatogram. Furthermore, since the baseline suppression process is incorporated, the peak can be accurately detected even when the baseline is drifting or wavy, which shows the superiority of the present invention.

It is the figure which showed an example of the peak detection result by general peak detection (only a positive peak). It is the figure which showed an example of the peak detection result by general peak detection (positive and negative peak mixed). It is the figure which showed an example of the peak detection result by the peak detection in this invention. 3a is an unprocessed chromatogram, FIG. 3b is a base point designation and baseline determination process, FIG. 3c is a baseline suppression, smoothing process, and a process of matching the sign of the signal output value, FIG. 3d is a first derivative result, FIG. 3e shows the peak start / end position, and FIG. 3f shows the final peak detection result. It is the figure which showed an example of the designation | designated of the base point in this invention, and a baseline determination process. It is the figure which showed an example of the peak detection result by general peak detection. FIG. 5a shows an unprocessed chromatogram, FIG. 5b shows a first derivative result, FIG. 5c shows a peak start / end position, and FIG. 5d shows a final peak detection result. 1 is a diagram showing a system configuration used in Example 1 and Comparative Example 1. FIG. It is the figure which showed the peak detection result in this invention in Example 1. FIG. FIG. 7a is an unprocessed chromatogram, and FIG. 7b is a diagram showing peak start / end positions after base point designation, baseline, baseline suppression, smoothing processing, and signal output value code matching. FIG. 7c shows a first-order differential curve after the above processing, and FIG. 7d shows a final peak detection result. It is the figure which showed the conventional peak detection result with respect to the chromatogram used in the comparative example 1. FIG. 8a shows an unprocessed chromatogram, FIG. 8b shows a first derivative result, and FIG. 8c shows a final peak detection result. It is the figure which showed the system configuration used in Example 2 and Comparative Example 2. It is the figure which showed the peak detection result in this invention in Example 2. FIG. FIG. 10a is an unprocessed chromatogram, and FIG. 10b is a diagram showing peak start / end positions after base point designation, baseline, baseline suppression, smoothing processing, and signal output value code matching. FIG. 10c shows a first-order differential curve after the above processing, and FIG. 10d shows a final peak detection result. It is the figure which showed the conventional peak detection result with respect to the chromatogram used in the comparative example 2. FIG. 11a shows an unprocessed chromatogram, FIG. 11b shows a first derivative result, and FIG. 11c shows a final peak detection result.

  The effect of the present invention was verified using an actual chromatogram.

Example 1
Verification was carried out using a GPC (Molecular Exclusion Chromatography) system. The actual measurement was performed using the liquid chromatogram system shown in FIG. The system is composed of a solvent degassing device (SD-8020) 2, a sample side liquid feeding pump (DP-8020) 3, a reference side liquid feeding pump (DP-8020) 10, a sample injection device (AS-8020) 4, a column oven. (CO-8020) 6, differential refractometer (RI-8020) 12, and data processor (LC-8020II) 9 (all manufactured by Tosoh Corporation). As the analytical column 5, SuperMultipore PW-N (4 μm 6.0 mm ID × 15 cm × 2) manufactured by Tosoh Corporation was used, and Polyvinylpyrrolidone K 15 (MW: 10,000, 25 mg / mL, Tokyo Chemical Industry ( ))). Other conditions are as follows.
Injection volume: 50 uL, column temperature: 40 ° C., sample side flow rate: 0.600 mL / min, reference side flow rate: 0.3 mL / min, eluent: 200 mM NaNO ₃ + 20% acetonitrile

  FIG. 7 shows a peak detection process and result according to the peak detection method of the present invention. FIG. 7a shows an unprocessed chromatogram and a base point. Fig. 7b shows the baseline as a linear expression with 0 and 12 minutes as base points, subtracted from the unprocessed chromatogram, and only the negative output value is multiplied by -1 to match the sign of the output value. And a chromatogram obtained by smoothing by a simple moving average method (61 points). FIG. 7c is a diagram showing a first derivative curve, a peak start point, and an end point of the processed chromatogram. FIG. 7d is a diagram in which peak detection is performed by applying the peak start point (time) and end point (time) in the processed chromatogram to an unprocessed chromatogram.

  As can be seen from FIG. 7b, all the chromatograms are converted into curves having positive values, and therefore peak detection can be easily performed by changing the first derivative curve. The point (time) at which the first derivative of the chromatogram exceeds 0.3 mABU / min is the peak start (white circle), and the point (time) at which -0.3 mABU / min is exceeded after the peak starts is the peak end. It is recognized as (black circle) (FIG. 7c).

  Table 1 shows the results.

  The peak start time (1) and peak end time (4) were applied to an untreated chromatogram, and output values (3) and (6) in the untreated chromatogram were obtained. As a result, the peak start in the untreated chromatogram was time (1) and output (3), and the peak end was time (4) and output (6) (FIG. 7d). Based on this, quantitative calculations such as peak area and height were performed.

  Table 2 shows the peak detection / quantification results obtained by the peak detection method of the present invention.

(Comparative Example 1)
The same untreated chromatogram as used in Example 1 was subjected to peak detection by a conventional detection method in which first derivative was taken, peak start and peak end were found by the threshold value, and quantitative calculation was performed. . The peak detection sensitivity was specified as 0.3 mABU / min, and peak detection was performed. In other words, the point (time) at which the first derivative value of the chromatogram exceeded 0.3 mABU / min was used as the peak start, and the point (time) after -0.3 mABU / min after the peak started was used as the peak end. Is recognized (FIG. 8b).

  As a result, in the conventional peak detection method, as shown in FIG. 8c, the peak detection becomes abnormal between the sharp positive peak derived from the monomer or additive and the negative peak derived from the solvent. This is because the lowest point of the negative peak is recognized as the peak end of the peak derived from the monomer or additive and the peak start of the solvent peak.

  Table 3 shows the peak detection / quantification results obtained by the conventional peak detection method.

  As is clear by comparing Table 2 and Table 3, there is no change in the quantitative values such as area and height for the polymer component peak eluting at 7.26 minutes, but the low molecule eluting at 10.18 minutes. In the component peak, the conventional peak detection method calculates the peak height three times and the area seven times larger than the method of the present invention. Also, the elution time is not a correct value because the point at which the signal turns negative is determined as the peak top. The solvent peak eluting at 10.64 minutes is identified as a positive peak by the conventional peak detection method, so that the area and height are both positive values, and the peak top is also in the vicinity where the signal goes from negative to zero. Since it is identified as the peak top, the value is not correct. In the method of the present invention, both the low molecular component peak eluting at 10.18 minutes and the solvent peak eluting at 10.64 minutes can be accurately quantified, indicating the superiority of the present invention.

(Example 2)
In Example 2, verification was performed using an ion chromatography system. The actual measurement was performed using the liquid chromatogram system shown in FIG. The system consists of a solvent degassing device (SD-8020) 2, a liquid feed pump (DP-8020) 3, a sample injection device (AS-8020) 4, a column oven (CO-8020) 6, an electric conductivity meter (CM- 8000) 7 and a data processing device (LC-8020II) 9 (both manufactured by Tosoh Corporation). As the analytical column 5, TSKgel SuperIC-AnionHS (4.6 mmID × 10 cm + guard column) manufactured by Tosoh Corporation was used, and a standard anion mixture (diluted 20 times, Kanto Chemical Co., Inc.) was separated. Other conditions are as follows.
Injection volume: 30 uL, column temperature: 40 ° C., flow rate: 1.50 mL / min, eluent: 3.8 mmol / L NaHCO ₃ +3.0 mmol / L Na ₂ CO ₃ (pH = 10.1)

  FIG. 10 shows a peak detection process and result according to the peak detection method of the present invention. FIG. 10a shows an unprocessed chromatogram and a base point. Fig. 10b shows a baseline (linear line) with 0 and 5 minutes as the base points, subtracted from the unprocessed chromatogram, and only negative output values are multiplied by -1 to sign the output values. It is the figure which showed the chromatogram obtained by combining and smoothing by the simple moving average method (21 points). FIG. 10c shows the first derivative curve, peak start point and end point of the processed chromatogram. FIG. 10d is a diagram in which the peak start point (time) and end point (time) in the processed chromatogram are applied to an unprocessed chromatogram to perform peak detection.

  As is clear from FIG. 10b, all the chromatograms are converted into curves having positive values, so that peak detection can be easily performed by changing the first derivative curve. The point (time) at which the first derivative of the chromatogram exceeds 2.0 mABU / min is set as the peak start (white circle), and the point (time) at which -2.0 mABU / min is exceeded after the peak starts is the peak end. It is recognized as (black circle) (FIG. 10c).

  Table 4 shows the results.

The peak start time (1) and peak end time (4) are applied to an unprocessed chromatogram, and output values (3) and (6) in the unprocessed chromatogram are obtained. Thus, the peak start in the unprocessed chromatogram is time (1) and output (3), and the peak end is time (4) and output (6) (FIG. 10d). Based on this, quantitative calculation such as peak area and height is performed.

  Table 5 shows the peak detection / quantification results obtained by the peak detection method of the present invention.

(Comparative Example 2)
The same untreated chromatogram as used in Example 2 was subjected to peak detection by a conventional detection method in which a first derivative was taken, peak start and peak end were found by the threshold value, and quantitative calculation was performed. . The peak detection sensitivity was specified as 2.0 mABU / min, and peak detection was performed. That is, the point (time) at which the first derivative value of the chromatogram exceeded 2.0 mABU / min was set as the peak start, and the point (time) after -2.0 mABU / min after the peak started was set as the peak end. Is recognized (FIG. 11b).

  As a result, in the conventional peak detection method, as shown in FIG. 11c, the peak detection becomes abnormal between the water dip that is a negative peak and the F ion peak that elutes first. This is because the peak start of the negative peak is the peak start, and the actual peak start of F ions cannot be detected. In such a state, the quantitative value of F ions becomes abnormal.

  Table 6 shows the peak detection / quantification results obtained by the conventional peak detection method.

  As is clear from comparison between Tables 5 and 6, there is no change in the quantitative values such as area and height for the Cl ion peak eluted after 1.84 minutes, but F eluted at 1.44 minutes. In the case of the ion peak, the conventional peak detection method calculates the peak height 7% and the area 17% larger than the present method. In the method of the present invention, the negative water dip peak can be detected and quantified as a negative peak, and the peak of the F ion eluted immediately after the water dip can be correctly detected and quantified. Is shown.

1. 1. Eluent 2. Deaeration device Liquid feed pump (sample side)
4). 4. Sample injection valve Analysis column 6. 6. Column constant temperature bath Differential refractometer8. Waste liquid9. System control and data processing device 10. Liquid feed pump (reference side)
11. Resistance tube 12. Electric conductivity meter

Claims (2)

As a first step, with respect to an unprocessed chromatogram, two or more data points serving as a base point of a baseline are selected, and a baseline is established by mathematical processing based on approximate calculation from the data points;
As a second step, subtracting the baseline from the untreated chromatogram to obtain a chromatogram without baseline fluctuations;
As a third step, the sign of the output value is obtained by multiplying the chromatogram without the baseline fluctuation by −1 only when the output value is negative or by −1 only when the output value is positive. And the process of combining
As a fourth step, for the chromatogram combined with the sign of the output value, obtaining a primary differential value, obtaining time information of the peak start point and peak end point,
As a fifth step, the time information of the peak start point and peak end point is adapted to the unprocessed chromatogram and the peak is determined,
A method for detecting a peak of a chromatogram, comprising:   The chromatogram peak detection method according to claim 1, further comprising a step of performing a smoothing process on the chromatogram having no baseline fluctuation between the second step and the third step.

Images