TW202011014A - Method, program and device for using fluorescence fingerprint information to quantify TSNAs contained in tobacco raw material - Google Patents
Method, program and device for using fluorescence fingerprint information to quantify TSNAs contained in tobacco raw material Download PDFInfo
- Publication number
- TW202011014A TW202011014A TW107132274A TW107132274A TW202011014A TW 202011014 A TW202011014 A TW 202011014A TW 107132274 A TW107132274 A TW 107132274A TW 107132274 A TW107132274 A TW 107132274A TW 202011014 A TW202011014 A TW 202011014A
- Authority
- TW
- Taiwan
- Prior art keywords
- tsnas
- aforementioned
- tobacco raw
- fingerprint information
- fluorescent fingerprint
- Prior art date
Links
Images
Landscapes
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
Description
本發明係有關於利用螢光指紋分析之處理原料中之TSNAs定量方法、程式及裝置,特別是有關於能合適地利用於TSNAs中的微量成分之定量的定量方法、程式及裝置。 The present invention relates to a method, program and device for quantifying TSNAs in processing materials using fluorescent fingerprint analysis, and in particular to a method, program and device for quantifying trace components that can be suitably used in TSNAs.
另外,TSNAs係指煙草特異亞硝基胺(Tobacco-specific nitrosamines),作為典型的TSNAs,可舉NNN(N'-亞硝基降菸鹼,N'-nitroso nornicotine)、NNK(4-甲基亞硝基胺基-1-3-吡啶基-1-丁酮,4-(methyl nitrosoamino)-1-(3-pyridyl)-1-butanone)、NAT(N'-亞硝基新煙草鹼,N'-nitroso anatabine)、NAB(N'-亞硝基毒藜鹼,N'-nitroso anabasine)之4種。 In addition, TSNAs refer to Tobacco-specific nitrosamines. As typical TSNAs, NNN (N'-nitroso nornicotine, N'-nitroso nornicotine), NNK (4-methyl Nitrosamino-1-3-pyridyl-1-butanone, 4-(methyl nitrosoamino)-1-(3-pyridyl)-1-butanone), NAT (N'-nitroso-nicotine, N'-nitroso anatabine) and NAB (N'-nitroso anabasine).
測定試料中成分的一般方法,有將試料中之各種成分萃取作為萃取液,以高效液體氣體層析 (HPLC-MS/MS)定量的方法。但是,該方法在樣本的前處理或在測定時粍費勞力與時間。 The general method for measuring the components in the sample includes a method of extracting various components in the sample as extracts and quantifying by high-performance liquid gas chromatography (HPLC-MS/MS). However, this method requires labor and time before sample preparation or measurement.
為了分析、定量試料中的成分,作為不需要樣本的前處理且可做非破壞分析之成分分析、定量手法,已知有利用近紅外(Near-Infrared)分光分析(NIR)之分析、定量。NIR係使用經測定的NIR光譜進行校準曲線的作成及驗證,並使用所作成之校準曲線進行成分分析以及定量之手法。 In order to analyze and quantify the components in the sample, as a component analysis and quantification method that does not require sample pretreatment and can be subjected to non-destructive analysis, analysis and quantification using near-infrared spectroscopic analysis (NIR) is known. NIR uses the measured NIR spectrum to create and verify a calibration curve, and uses the calibration curve to perform component analysis and quantification.
此外,作為不需要樣本的前處理且可做非破壞分析之成分分析、定量手法,已知有利用螢光指紋資訊分析、定量之手法(例如,參考專利文獻1等)。如此般之螢光指紋分析係在使用經測定、取得之螢光指紋資訊進行校準曲線之作成及驗證,並使用所作成之校準曲線進行成分分析以及定量,就此點而言,與NIR共有一部分的特徵。以下,係針對螢光指紋以及螢光指紋分析之概觀做簡單的說明。 In addition, as a component analysis and quantification method that does not require sample pretreatment and can be subjected to non-destructive analysis, a method using fluorescent fingerprint information analysis and quantification is known (for example, refer to
如第2圖所示,對含有螢光物質的試驗試料一邊階段性地改變波長一邊照射激發光,測定從試驗試料發出之光(螢光)時,於以激發波長(λEx)、螢光波長(測定波長)(λEm)、螢光強度(IEx,Em)作為3正交軸之3維空間中,將對應之點作圖。 As shown in Figure 2, the test sample containing the fluorescent substance is irradiated with excitation light while changing the wavelength step by step. When the light (fluorescence) emitted from the test sample is measured, the excitation wavelength (λEx) and the fluorescent wavelength (Measurement wavelength) (λEm), fluorescence intensity (IEx, Em) are three orthogonal axes in a three-dimensional space, and the corresponding points are plotted.
將該等點之集合予以視覺化者稱為螢光指紋,或是激發螢光矩陣(Excitation Emission Matrix;EEM)。 Those who visualize the collection of these points are called fluorescent fingerprints, or Excitation Emission Matrix (EEM).
螢光指紋係可藉由將各點之螢光強度以等高線形狀或色分布等表示,而表示為3維圖表(參考第3圖),另外,亦可表示為2維圖表(參考第4圖)。 The fluorescent fingerprint can be represented as a 3-dimensional graph (refer to Figure 3) or a 2-dimensional graph (refer to Figure 4) by expressing the fluorescent intensity of each point as a contour shape or color distribution, etc. ).
如此般之螢光指紋係顯示著具有3維的膨大資訊之試驗試料固有的圖案,可使用於各種的鑑別或定量等。 Such a fluorescent fingerprint shows the inherent pattern of the test sample with 3-dimensional swelling information, which can be used for various identification or quantification.
如此般之螢光指紋分析相較於其他的分光分析手法,係具有下述優點:可在不需對試驗試料進行螢光染色等的前處理而進行試驗試料的特性分析(characterization),操作容易且測量亦能在短時間進行,即便資訊量多定量亦可相對容易地進行,能以非破壞、非接觸方式測量,由於使用紫外光至可見光故裝置相對較便宜等。 Compared with other spectroscopic analysis methods, such a fluorescent fingerprint analysis has the following advantages: it can perform characterization of the test sample without pretreatment of the test sample such as fluorescent dyeing, and the operation is easy And the measurement can also be performed in a short time, even if the amount of information is more quantitative, it can be relatively easily performed, and it can be measured in a non-destructive, non-contact manner. The device is relatively cheap due to the use of ultraviolet light to visible light.
為了測量螢光指紋,需要具有掃描激發波長而連續地測量各別螢光光譜的機能之螢光分光光度計,而具有如此般之機能之螢光分光光度計亦有市售者(日立先端科技(hitachi-hightech)公司製F-7000等)。 In order to measure fluorescent fingerprints, a fluorescent spectrophotometer with the function of scanning the excitation wavelength and continuously measuring the respective fluorescent spectra is required, and a fluorescent spectrophotometer with such a function is also commercially available (Hitachi Advanced Technology) (hitachi-hightech) F-7000, etc.).
[專利文獻1]日本特開2018-40779號公報 [Patent Literature 1] Japanese Patent Application Publication No. 2018-40779
將高效液體氣體層析(HPLC-MS/MS)適用 於TSNAs測定時,如前述般,在樣本的前處理或測定時耗費勞力與時間。 When high-performance liquid gas chromatography (HPLC-MS/MS) is applied to the determination of TSNAs, as described above, it takes labor and time to pre-process or measure the sample.
此外,雖有將NIR適用於TSNAs測定之考量,但NIR係有例如下述之問題故難以採用NIR。 In addition, although there are considerations for applying NIR to the measurement of TSNAs, NIR has the following problems, for example, so it is difficult to adopt NIR.
(1)沒有足以用於測定煙草原料中的TSNAs濃度帶的靈敏度。 (1) There is not enough sensitivity to measure the concentration band of TSNAs in tobacco raw materials.
(2)試料中的TSNAs與菸鹼的峰值重複,而且菸鹼的存在量遠更為較多,故使得TSNAs的峰值會被埋沒在菸鹼的峰值中,難以辨識TSNAs的峰值。 (2) The peaks of TSNAs and nicotine in the sample are repeated, and the amount of nicotine is much more, so the peaks of TSNAs will be buried in the peaks of nicotine, it is difficult to identify the peaks of TSNAs.
在此,考量將可設定為比NIR更高靈敏度之螢光指紋分析適用於TSNAs測定。如此之情況,與NIR相比,由於可測定考慮到精密的化學構造之差異,故可辨識菸鹼與TSNAs之峰值位置。然而,上述的4種TSNAs之中,特別是NNK與NAB在煙草原料中的存在量非常微量,故即便單純使用螢光指紋分析,亦難以單獨定量該等成分。 Here, it is considered that fluorescent fingerprint analysis, which can be set to a higher sensitivity than NIR, is suitable for TSNAs determination. In this case, compared with NIR, the difference between the nicotine and TSNAs can be identified because the difference in the precise chemical structure can be measured. However, among the above-mentioned four types of TSNAs, especially the amount of NNK and NAB present in tobacco raw materials is very small, so even using fluorescent fingerprint analysis alone, it is difficult to quantify these components separately.
為了解決如前述般之課題,本案之發明人等想出了利用螢光指紋分析用以定量TSNAs之新穎手法。以下說明該該手法的概要。 In order to solve the aforementioned problems, the inventors of the present case have devised a novel method for quantifying TSNAs using fluorescent fingerprint analysis. The outline of this technique will be described below.
本案之發明人等對處理原料(煙草原料)中所存在之4種TSNAs進行檢討的結果,獲得下述知識見解:即使4種TSNAs之含量不同,該等的含量之比(各TSNA之存在比率)在煙草種相同的煙草原料中亦無太大的差異。 The inventors of this case reviewed the four types of TSNAs present in the processed raw materials (tobacco raw materials), and obtained the following knowledge: Even if the contents of the four types of TSNAs are different, the ratio of these contents (the ratio of the existence of each TSNA ) There is not much difference in the same tobacco raw material.
換言之,吾人獲得下述知識見解:將煙草種類相同的煙草原料中所存在之4種TSNAs的含量(例如,重量%)分別設為NNN:a、NAT:b、NNK:c、NAB:d時,該等的間之比(各TSNA之存在比率)即a:b:c:d於相同煙草種中略為一定。就煙草種而言,若以白肋煙種(Burley)為例,則得到a:b:c:d=55:35:8:2左右之數值。 In other words, we obtained the following knowledge: when the contents (eg,% by weight) of four TSNAs present in tobacco materials of the same tobacco type are set to NNN:a, NAT:b, NNK:c, NAB:d The ratio between these (the existence ratio of each TSNA), namely a: b: c: d, is slightly constant in the same tobacco species. As for the tobacco species, taking Burley as an example, a value of about a:b:c:d=55:35:8:2 is obtained.
從而,基於上述般的知識見解進一步檢討,發現當將4種TSNAs的4個峰值整體當作一個1峰值時,該1個峰值雖然會因為成分量而使峰值的高度有所變化,但於煙草種相同的煙草原料中,大致上呈現類似的形狀(詳如後述)。 Therefore, based on the above-mentioned knowledge and insights, we further found that when the four peaks of the four TSNAs are regarded as a single peak, the height of the peak may change due to the amount of ingredients. Among the same kind of tobacco raw materials, they have roughly similar shapes (details will be described later).
藉此,得到了以下結論:將4種TSNAs之總和認定作為螢光指紋之峰值,並抽出此峰值,藉此將追加如NNK與NAB般難以單獨定量的成分之定量值後的4種TSNAs的總和量予以定量,基於該總和量與各TSNA之已知的存在比率,可將4種TSNAs予以定量。 From this, the following conclusion was obtained: the sum of the four TSNAs was identified as the peak of the fluorescent fingerprint, and this peak was extracted, thereby adding the quantitative values of the four TSNAs after the quantitative values of components that are difficult to quantify separately like NNK and NAB The total amount is quantified, and based on the known ratio of the total amount to each TSNA, four TSNAs can be quantified.
本發明係基於上述般的檢討結果所思考出者,如果例示本發明的實施態樣,則如下列所述。 The present invention has been conceived based on the above-mentioned review results. If the embodiment of the present invention is exemplified, it will be as follows.
(態樣1) (Aspect 1)
一種利用螢光指紋分析之煙草原料之定量方法,係包括下列步驟:螢光指紋資訊取得步驟,係針對含有NNN、NNK、NAT以及NAB之4種煙草特異亞硝基胺(TSNAs)的煙草原料,取得由激發波長、螢光波長、螢光強度之數 據所構成之螢光指紋資訊;推測模型作成步驟,係將前述螢光指紋資訊作為解釋變數(explanatory variable),將從前述煙草原料中之4種TSNAs之已知的定量值所算出來的前述4種TSNAs之含量的總和量作為目標變數(objective variable),而取得校準曲線;總和量推測步驟,係基於前述校準曲線及前述4種TSNAs之含量為未知的煙草原料之螢光指紋資訊,推測前述煙草原料所含有之前述4種TSNAs之含量的總和量;以及含量推測步驟,係基於前述所推測之前述4種TSNAs之含量的總和量及前述4種TSNAs之已知的存在比率,推測前述4種TSNAs之含量。 A method for quantifying tobacco raw materials using fluorescent fingerprint analysis includes the following steps: fluorescent fingerprint information acquisition step for four tobacco-specific nitrosamines (TSNAs) containing NNN, NNK, NAT and NAB To obtain fluorescence fingerprint information composed of data on excitation wavelength, fluorescence wavelength, and fluorescence intensity; the speculative model creation step is to use the aforementioned fluorescence fingerprint information as an explanatory variable (explanatory variable). The total amount of the four TSNAs calculated from the known quantitative values of the four TSNAs is used as the objective variable to obtain the calibration curve; the total amount estimation step is based on the calibration curve and the four TSNAs. The content is the fluorescent fingerprint information of the unknown tobacco raw material, and the total amount of the aforementioned four TSNAs contained in the aforementioned tobacco raw material is estimated; and the content estimation step is based on the total amount of the aforementioned estimated four TSNAs In addition to the known existing ratios of the four TSNAs, the contents of the four TSNAs are speculated.
(態樣2) (Aspect 2)
如態樣1所述之利用螢光指紋分析之煙草原料之定量方法,其中,於前述推測模型作成步驟中,係藉由多變量分析作成上述校準曲線。 The method for quantifying tobacco raw materials by fluorescent fingerprint analysis as described in
(態樣3) (Aspect 3)
如態樣2所述之利用螢光指紋分析之煙草原料之定量方法,其中,前述多變量分析係PLS回歸分析。 The quantitative method of tobacco raw material using fluorescent fingerprint analysis as described in
(態樣4) (Aspect 4)
如態樣1至3所述之利用螢光指紋分析之煙草原料之定量方法,更包含前處理步驟,該前處理步驟包括針對前述螢光指紋資訊之刪除處理。 The method for quantifying tobacco raw materials by fluorescent fingerprint analysis as described in
(態樣5) (Aspect 5)
如態樣1至4所述之利用螢光指紋分析之煙草原料之定量方法,其中,前述試驗試料係在照射激發光之前被粉 碎、混合成粉末狀。 The quantitative method of tobacco raw material by fluorescent fingerprint analysis as described in
(態樣6) (Aspect 6)
如態樣5所述之利用螢光指紋分析之煙草原料之定量方法,係藉由前述試驗粉碎使試料之粒徑成為1mm以下。 As described in aspect 5, the quantitative method of the tobacco raw material by fluorescent fingerprint analysis is to make the particle size of the sample less than 1 mm by the aforementioned test pulverization.
(態樣7) (Aspect 7)
如態樣1至6所述之利用螢光指紋分析之煙草原料之定量方法,其中,為了在事前使水分量固定化,前述煙草原料係以既定的調和條件儲藏放置既定時間。 The method for quantifying tobacco raw materials using fluorescent fingerprint analysis as described in
(態樣8) (Aspect 8)
如態樣7所述之利用螢光指紋分析之煙草原料之定量方法,其中,前述調和條件為溫度22℃、濕度60%之室內之條件,前述既定時間為24小時以上。 The method for quantifying tobacco raw materials by fluorescent fingerprint analysis as described in aspect 7, wherein the conditioning conditions are indoor conditions at a temperature of 22°C and a humidity of 60%, and the predetermined time is more than 24 hours.
(態樣9) (Aspect 9)
一種程式,係用以使電腦執行如態樣1至8所述之煙草原料之定量方法。 A program used to make the computer execute the quantitative method of tobacco raw materials as described in
(態樣10) (Aspect 10)
一種裝置,係具備:推測模型作成手段,係輸入有關含有NNN、NNK、NAT以及NAB之4種煙草特異亞硝基胺(TSNAs)的煙草原料之由激發波長、螢光波長、螢光強度之數據所構成之螢光指紋資訊,將前述螢光指紋資訊作為解釋變數,並將從前述煙草原料中之4種TSNAs之已知的定量值所算出來的前述4種TSNAs之含量的總和量作為目標變數,而取得校準曲線;總和量推測手段,係基於前述校準曲線及前述4種TSNAs之含量為未知的煙草原料之 螢光指紋資訊,推測前述煙草原料所含有之前述4種TSNAs之含量的總和量;以及含量推測手段,係基於前述所推測之前述4種TSNAs之含量之總和量及前述4種TSNAs之已知的存在比率,推測前述4種TSNAs之含量。 A device equipped with speculative model creation means to input the excitation wavelength, fluorescence wavelength, and fluorescence intensity of four tobacco-specific nitrosamines (TSNAs) containing NNN, NNK, NAT, and NAB The fluorescent fingerprint information constituted by the data, using the aforementioned fluorescent fingerprint information as an explanatory variable, and taking the sum of the contents of the aforementioned four TSNAs calculated from the known quantitative values of the four TSNAs in the tobacco raw material as The target variable is used to obtain the calibration curve; the total amount estimation method is based on the aforementioned calibration curve and the fluorescent fingerprint information of the tobacco raw material whose content is unknown for the aforementioned four TSNAs, to speculate the content of the aforementioned four TSNAs contained in the tobacco raw material The total amount; and the content estimation means are based on the total amount of the above-mentioned four kinds of TSNAs and the known existence ratio of the four kinds of TSNAs, so as to estimate the contents of the four kinds of TSNAs.
(態樣11) (Aspect 11)
態樣10所述之裝置,更具備前處理手段,該前處理手段係對前述螢光指紋資訊進行前處理。 The device described in aspect 10 further includes pre-processing means for pre-processing the fluorescent fingerprint information.
(態樣12) (Aspect 12)
如態樣11所述之裝置,其中,前述前處理包括針對前述螢光指紋資訊之刪除處理。 The device according to aspect 11, wherein the pre-processing includes deleting processing of the fluorescent fingerprint information.
另外,「程式」係指基於任意的言語或記述方法所記述之數據處理方法,不論為原始碼或二元碼等形式均在所不問。此外,「程式」可以單一形態構成,亦可分散構成複數個模組或函式庫(Library),此外,亦可以與其他既存的程式聯動而達成其功能之方式構成。 In addition, "program" refers to the data processing method described based on any language or description method, regardless of the form of the original code or binary code. In addition, the "program" can be composed in a single form, or can be distributed to form a plurality of modules or libraries (Library), in addition, it can also be linked with other existing programs to achieve its function.
此外,「裝置」可為由硬體所構成者,亦可為藉由電腦的軟體實現各種機能之機能實現手段的組合所構成者。機能實現手段例如可包括程式模組。 In addition, the "device" may be constituted by hardware, or may be constituted by a combination of functional realization means for realizing various functions by software of a computer. The function realization means may include a program module, for example.
依據本發明,可迅速且確實地進行處理原料中所含之4種TSNAs之定量。 According to the present invention, the quantification of the four TSNAs contained in the raw materials for processing can be performed quickly and reliably.
S01、S02、S03、S04、S05、S06、S07‧‧‧步驟 S01, S02, S03, S04, S05, S06, S07
100‧‧‧TSNAs定量裝置 100‧‧‧TSNAs quantitative device
110‧‧‧前處理手段 110‧‧‧Pre-processing means
120‧‧‧推測模型作成手段 120‧‧‧ Speculation model creation method
130‧‧‧總和量推測手段 130‧‧‧ Total guessing method
140‧‧‧含量推測手段 140‧‧‧ content speculation means
第1圖係用以說明本發明之實施的一態樣的概要之流 程圖。 Fig. 1 is a flowchart for explaining an outline of an embodiment of the present invention.
第2圖係顯示朝測量對象物照射激發光時,從該測量對象物發出之螢光的光譜之概要之說明圖。 FIG. 2 is an explanatory diagram showing an overview of the spectrum of the fluorescent light emitted from the measurement object when the excitation light is irradiated to the measurement object.
第3圖係將螢光指紋之一例以3維顯示之等高線形狀之圖表。 Figure 3 is a graph showing the contour of a three-dimensional display of an example of a fluorescent fingerprint.
第4圖係將螢光指紋之一例以2維顯示之等高線形狀之圖表。 Figure 4 is a graph showing the contour of a two-dimensional display of an example of a fluorescent fingerprint.
第5圖係表示樣本之螢光指紋之一例之圖。 Fig. 5 is a diagram showing an example of a fluorescent fingerprint of a sample.
第6圖係表示包含TSNAs試藥之標準品的螢光指紋之一例之圖。 Fig. 6 is a diagram showing an example of fluorescent fingerprints containing standard products of TSNAs reagents.
第7圖係對複數個樣本,將4種TSNAs之含量之總和量及藉由來自螢光指紋資訊之推測值所規定之點予以作圖之圖表。 Figure 7 is a graph that plots the sum of the contents of the four TSNAs and the points specified by the estimated value from the fluorescent fingerprint information for a plurality of samples.
第8圖係用以說明本發明之實施的另一態樣的概要之方塊圖。 FIG. 8 is a block diagram for explaining the outline of another aspect of the implementation of the present invention.
以下,說明本發明之實施的態樣,但請留意,本發明不限於以下說明之實施的態樣。 The following describes the implementation of the present invention, but please note that the present invention is not limited to the implementation described below.
<本發明之實施的一態樣之概要> <Summary of an aspect of implementation of the present invention>
第1圖係用以說明實施本發明之一態樣的概要之流程圖。 FIG. 1 is a flowchart for explaining the outline of an aspect of implementing the present invention.
最先,準備4種TSNAs之含量為已知的試驗試料(煙草原料)(S01),對如此般之已知的試驗試料測定螢光指紋,取得螢光指紋資訊(S02)。 First, prepare four types of TSNAs with known test samples (tobacco raw materials) (S01), measure fluorescent fingerprints on such known test samples, and obtain fluorescent fingerprint information (S02).
其次,對已取得之螢光指紋資訊,視需要進行前處理(S03)。另外,此前處理亦可省略。前處理的細節係如後述。 Secondly, pre-processing the acquired fluorescent fingerprint information (S03) if necessary. In addition, the previous processing may be omitted. The details of the pre-processing will be described later.
其次,將已取得之螢光指紋資訊及試驗試料中所含之4種TSNAs之總和量之關係進行模型化,作成推測模型(校準曲線)(S04)。具體而言,該模型化係藉由使用各式各樣多變量分析手法或數據挖掘(data mining)手法而構築推測式來進行,該推測式係將所取得之螢光指紋資訊作為解釋變數,並將4種TSNAs之總和量作為目的變數者,以作成從螢光指紋資訊試驗試料中推測4種TSNAs之總和量之校準曲線(回歸式)。另外,推測式之構築所使用之演算法可以是更汎用且可對應非線形現象之機械學習演算法,例如支援向量機(support vector machine,SVM)、隨機森林(random forest,RF)、神經網路(neural network)等。模型化之際所使用之多變量分析手法之例係如後述。 Next, the relationship between the obtained fluorescent fingerprint information and the sum of the four TSNAs contained in the test sample is modeled to make a speculative model (calibration curve) (S04). Specifically, the modeling is performed by constructing a speculative formula using various multivariate analysis techniques or data mining techniques, which uses the obtained fluorescent fingerprint information as an explanatory variable, Using the sum of the four TSNAs as the target variable, a calibration curve (regression formula) was made to estimate the sum of the four TSNAs from the fluorescent fingerprint information test sample. In addition, the algorithm used in the speculative construction can be a more general machine learning algorithm that can correspond to the non-linear phenomenon, such as support vector machine (SVM), random forest (RF), neural network (neural network) etc. Examples of multivariate analysis techniques used in modeling are described below.
驗證以如此方式構築之推測模型(校準曲線),確認其有效性(S05)。 The speculative model (calibration curve) constructed in this way is verified to confirm its validity (S05).
使用有效性經確認之推測模型(校準曲線),基於未知的處理原料(煙草原料)之螢光指紋資訊,推測前述未知的處理原料(煙草原料)中作為成分含有之4種TSNAs之總和量(S06)。 Using the validated speculative model (calibration curve), based on the fluorescent fingerprint information of the unknown processing raw material (tobacco raw material), the total amount of the four TSNAs contained as an ingredient in the aforementioned unknown processing raw material (tobacco raw material) is estimated ( S06).
基於所推測之4種TSNAs之總和量及各TSNA之已知的存在比率,推測4種TSNAs之含量(S07)。 Based on the estimated sum of the four TSNAs and the known existence ratio of each TSNA, the content of the four TSNAs is estimated (S07).
<本發明之實施的另一態樣之概要> <Summary of another aspect of implementation of the present invention>
第8圖係用以說明本發明之實施的另一態樣之概要之方塊圖。 FIG. 8 is a block diagram for explaining the outline of another aspect of the implementation of the present invention.
利用螢光指紋分析以定量處理原料中之TSNAs之TSNAs定量裝置100係具備下列手段:前處理手段110,係輸入有關試料(煙草原料)之由激發波長、螢光波長、螢光強度之數據所構成之螢光指紋資訊,對所輸入之螢光指紋資訊進行前處理;推測模型作成手段120,係輸入前述前處理手段110之輸出,將已實施前處理的螢光指紋資訊作為解釋變數,並將前述試驗試料所含之4種TSNAs的總和量作為目標變數,而取得校準曲線;總和量推測手段130,係基於藉由前述推測模型作成手段120取得之前述校準曲線與未知的處理原料(煙草原料)的螢光指紋資訊,推測前述未知的處理原料(煙草原料)中作為成分含有之4種TSNAs之總和量;以及含量推測手段140,係基於所推測之4種TSNAs之總和量及各TSNA之已知的存在比率,推測4種TSNAs之含量。另外,前處理手段110亦可省略。 The
首先,使用已知的螢光分光光度計等,自成分的種類與含量為已知的試驗試料取得螢光指紋資訊。 First, using a known fluorescent spectrophotometer, etc., the fluorescent fingerprint information is obtained from a test sample with known types and contents of ingredients.
其次,將已取得之螢光指紋資訊輸入前述前處理手段110,對所輸入之螢光指紋資訊進行前處理。另外,此前處理可予以省略。此外,進行前處理時,前處理之細節詳如後述。 Secondly, the obtained fluorescent fingerprint information is input into the aforementioned pre-processing means 110 to perform pre-processing on the input fluorescent fingerprint information. In addition, the previous processing can be omitted. In addition, the details of the pre-processing will be described later when performing the pre-processing.
其次,藉由推測模型作成手段120,將完成 前處理之螢光指紋資訊與4種TSNAs之總和量之關係予以模型化,作成推測模型(校準曲線)。該模型化係與前述本發明之實施的一態樣相同。然後,驗證以如此方式構築成之推測模型(校準曲線),確認其有效性,將有效性經確認之推測模型(校準曲線)預先記憶在未圖示之記憶體等。 Next, by means of the speculation model creation means 120, the relationship between the fluorescent fingerprint information that has been pre-processed and the sum of the four TSNAs is modeled to make a speculation model (calibration curve). The modelling is the same as the aforementioned embodiment of the present invention. Then, the speculative model (calibration curve) constructed in this way is verified, and its validity is confirmed, and the speculative model (calibration curve) whose validity is confirmed is stored in a memory not shown in advance.
總和量推測手段130係使用有效性經確認之推測模型(校準曲線),基於未知的處理材料(煙草原料)之螢光指紋資訊,推測前述未知的處理材料(煙草原料)中作為成分含有之4種TSNAs之總和量。另外,較佳係對於未知的試料的螢光指紋資訊亦藉由前述前處理手段110進行前處理(第8圖之TSNAs定量裝置100係採用如此之構成),惟視需要亦可省略如此般之前處理。 The total amount estimation means 130 uses a speculative model (calibration curve) whose validity is confirmed, based on the fluorescent fingerprint information of the unknown processing material (tobacco raw material), to estimate the content of the aforementioned unknown processing material (tobacco raw material) as an ingredient 4 The total amount of TSNAs. In addition, it is preferable that the fluorescent fingerprint information of the unknown sample is also pre-processed by the aforementioned pre-processing means 110 (the
含量推測手段140係基於所推測之4種TSNAs之總和量及各TSNA之已知的存在比率,推測4種TSNAs之含量。另外,各TSNA之已知的存在比率之資訊可預先儲納在含量推測手段140或TSNAs定量裝置100中,亦可採用從外部提供至TSNAs定量裝置100之方式。 The content estimation means 140 estimates the content of the four TSNAs based on the estimated total sum of the four TSNAs and the known existence ratio of each TSNA. In addition, information on the known existence ratio of each TSNA may be stored in the content estimation means 140 or the
<模型化之際所使用之多變量分析手法> <Multivariate analysis technique used in modeling>
作為模型化之際所使用之多變量分析手法,可使用PLS(偏最小平方法,Partial Least Squares)回歸分析、複回歸分析、主成分回歸分析、最小平方法等各種解析手法。 As a multivariate analysis method used in modeling, various analytical methods such as PLS (Partial Least Squares) regression analysis, complex regression analysis, principal component regression analysis, and least square method can be used.
PLS回歸分析係以使主成分與目的變數之共變異數變成最大的方式抽出主成分之手法,在解釋變數之間有強的相關時(產生多重共線性時)為有效的手法。 PLS regression analysis is a method of extracting principal components in such a way that the covariance between the principal component and the target variable becomes the largest. It is an effective method when there is a strong correlation between the explanatory variables (when multicollinearity is generated).
主成分回歸分析係以使主成分之分散為最大之方式抽出主成分之手法,係僅使用解釋變數進行主成分分析,在所得的主成分與目的變數間藉由最小平方法進行複回歸分析。 Principal component regression analysis is a method of extracting principal components in such a way as to maximize the dispersion of principal components. It uses only explanatory variables to perform principal component analysis, and performs complex regression analysis by the least squares method between the obtained principal components and the target variables.
複回歸分析係在解釋變數與目的變數之間適用最小平方法者,具有與主成分回歸分析為不同的特徵。 The compound regression analysis system applies the least squares method between the explanatory variable and the target variable, and has different characteristics from the principal component regression analysis.
另外,上述之各解析手法本身已為習知者,並非本發明在模型化之際特殊的處理所必要者,故將處理內容之細節予以省略,但關於PLS與校準曲線作成的關聯將如後述。 In addition, the above analysis methods themselves are already known, and are not necessary for the special processing of the present invention at the time of modeling, so the details of the processing content are omitted, but the relationship between the PLS and the calibration curve will be described later .
<關於將4種TSNAs之總和認定為螢光指紋之峰值,並取出該峰值之手法> <About identifying the sum of the four types of TSNAs as the peak of fluorescent fingerprints and removing the peak>
本發明係如前述,係可藉由採用將4種TSNAs之總和認定為螢光指紋之峰值,並取出該峰值之手法,將追加如NNK與NAB般難以單獨定量的成分之定量值後的4種TSNAs的總和量予以定量,基於該總和量與各TSNA之已知的存在比率,將4種TSNAs予以定量。 The present invention is as described above, by adopting the method of identifying the sum of four types of TSNAs as the peak of fluorescent fingerprints and taking out the peak, adding the quantitative values of components that are difficult to quantify separately like NNK and NAB. The total amount of TSNAs is quantified, and based on the known existence ratio of the total amount to each TSNA, four TSNAs are quantified.
參照第5圖以及第6圖,其係說明將4種TSNAs之4個峰值整體當作1個峰值時,不論峰值之高低,峰值的形狀會於煙草種相同的煙草原料中,大致上呈現類似的形狀。 Referring to Figure 5 and Figure 6, it shows that when the four peaks of the four TSNAs are regarded as one peak in general, the shape of the peak will be similar to that of the same tobacco material, regardless of the peak value. shape.
第5圖係表示樣本之螢光指紋之一例之圖。使用白肋煙種作為樣本,該樣本中所含之4種TSNAs之總 和為2.856μg/g。另外,對於樣本係基於後述般的調整條件實施處理。 Fig. 5 is a diagram showing an example of a fluorescent fingerprint of a sample. Burley tobacco was used as a sample, and the total of the four TSNAs contained in the sample was 2.856 μg/g. In addition, the sample system is processed based on adjustment conditions as described later.
第6圖係表示包含TSNAs試藥之標準品的螢光指紋之一例之圖。標準品係將4種TSNAs之試藥以與白肋煙種大致相同之比率混合於乙醇溶液中所作成者,該標準品所含之4種TSNAs之總和為1.28μg/g。 Fig. 6 is a diagram showing an example of fluorescent fingerprints containing standard products of TSNAs reagents. The standard product was prepared by mixing 4 TSNAs in the ethanol solution at approximately the same ratio as Burley tobacco. The total of the 4 TSNAs contained in this standard was 1.28 μg/g.
另外,兩者之螢光指紋之測定條件均相同,為激發光200至600nm,螢光200至700nm,解析能力2.5nm,狹縫寬2.5nm,光電倍增管(Photomultiplier)靈敏度950V。 In addition, the measurement conditions of the two fluorescent fingerprints are the same. The excitation light is 200 to 600 nm, the fluorescence is 200 to 700 nm, the resolution is 2.5 nm, the slit width is 2.5 nm, and the sensitivity of the photomultiplier is 950V.
對比第5圖的螢光指紋資訊與第6圖的螢光指紋資訊,大致上可得到以下的知識見解。 Comparing the fluorescent fingerprint information in FIG. 5 with the fluorescent fingerprint information in FIG. 6, the following knowledge can be roughly obtained.
˙經第6圖的標準品確認之4種TSNAs的螢光指紋資訊(參考第6圖標有「A’」的橢圓形框區域(「區域A’」)),於第5圖的樣本中亦可確認得到(參考第5圖標有「A」的橢圓形框區域(「區域A」))。 ˙Fluorescent fingerprint information of the 4 TSNAs confirmed by the standard in Figure 6 (refer to the oval frame area with "A'" in the sixth icon ("Area A'")), also in the sample in Figure 5 It can be confirmed (refer to the oval frame area with "A" in the fifth icon ("Area A")).
˙將第6圖的區域A’與第5圖的區域A相比較時,在螢光指紋資訊之峰值的高度雖然有所不同,但兩者具有略為類似的形狀。 ˙When comparing the area A'in Figure 6 with the area A in Figure 5, although the height of the peak of the fluorescent fingerprint information is different, the two have a slightly similar shape.
˙於第5圖中,受到樣本中其他材料的螢光的影響,多數的帶域發生飽合(因為靈敏度取最大值),即便於如此般之狀況下中,仍然可辨識區域A,咸認夾雜物的影響較低。 ˙In Figure 5, due to the influence of the fluorescence of other materials in the sample, most of the bands are saturated (because the sensitivity takes the maximum value), even in such a situation, the area A can still be identified, salty recognition The impact of inclusions is low.
˙在實際測定時,藉由刪除預測會產生飽和的帶域之螢光指紋資訊,咸認能有效使TSNAs的螢光指紋資訊區別化。 此外,關於前處理,咸認能有效避免例如2次微分等飽和有可能對其他波長造成影響的處理。 ˙During actual measurement, by deleting the predicted fluorescent fingerprint information that will produce saturation bands, Xianren can effectively distinguish the fluorescent fingerprint information of TSNAs. In addition, with regard to pre-processing, Xianxin can effectively avoid the processing that saturation, such as the second derivative, may affect other wavelengths.
以下,針對本發明之實施的一態樣的各步驟進行說明。 Hereinafter, each step of the embodiment of the present invention will be described.
〔試驗試料(煙草原料)的準備〕 [Preparation of test samples (tobacco raw materials)]
將4種TSNAs之含量為已知的各樣本粉碎成粒徑1mm以下並充分混合,將此準備作為試驗試料。 Each sample with known content of 4 TSNAs was crushed to a particle size of 1 mm or less and mixed thoroughly, and this preparation was used as a test sample.
由於4種TSNAs有可能局部存在於煙草原料中,故較佳係如此般在測定前將試料粉碎為一定粒徑(1mm徑)以下,並充分混合後再取得螢光指紋。另外,針對各樣本中的4種TSNAs之含量,事前以高效液體氣體層析(HPLC-MS/MS)予以定量。 Since the four types of TSNAs may be localized in the tobacco raw materials, it is better to crush the sample to a certain particle size (1 mm diameter) or less before the measurement in this way, and obtain the fluorescent fingerprint after fully mixing. In addition, the content of the four TSNAs in each sample was quantified in advance by high-performance liquid gas chromatography (HPLC-MS/MS).
此外,為了在事前使水分量固定化,試驗試料係使用經儲藏放置者。為了使試驗試料的水分量固定化,較佳係以調和條件(22度60%的室內)儲藏放置24小時以上。如此般藉由在測定前使水分量預先固定化,會使得螢光之峰值位移(shift)較不易發生。 In addition, in order to fix the amount of water beforehand, the test sample system used a storage placer. In order to fix the moisture content of the test sample, it is preferable to store it under the condition of conditioning (22 degrees 60% indoor) for more than 24 hours. In this way, by pre-fixing the water content before measurement, the peak shift of the fluorescence is less likely to occur.
〔螢光指紋資訊的取得〕 [Acquisition of Fluorescent Fingerprint Information]
為了取得試驗試料的螢光指紋資訊,使用日立先端科技公司製F-7000作為螢光指紋測定裝置,藉由反射法(FrontFace)進行測定。 In order to obtain the fluorescent fingerprint information of the test sample, F-7000 manufactured by Hitachi Advanced Technologies Co., Ltd. was used as the fluorescent fingerprint measurement device, and the measurement was performed by the reflection method (FrontFace).
測定條件係激發光200至600nm,螢光200至700nm,解析能力2.5nm,狹縫寬2.5nm,光電倍增管靈敏度950V。另外,若考量解析能力5nm,則測定波長至 少容許5nm左右的誤差。 The measurement conditions are excitation light 200 to 600 nm,
〔對螢光指紋資訊之前處理〕 [Pre-processing of fluorescent fingerprint information]
取得試料的螢光指紋資訊之際,可直接使用螢光指紋的測量值(每激發波長的螢光光譜),但視需要,亦有進行各種前處理的必要。 When obtaining the fluorescence fingerprint information of the sample, the measurement value of the fluorescence fingerprint (fluorescence spectrum per excitation wavelength) can be directly used, but various pre-processing is necessary as necessary.
作為用以從所測量之螢光指紋中除去雜訊而取得有效的螢光指紋資訊的前處理手法,例如可採用非螢光成分的除去處理,散射光的除去處理,低靈敏度區域的刪除處理中的1種或複數種的組合。此外,亦可對所取得之螢光指紋資訊採用中心化、規格化、標準化、基本線(base-line)補正、平滑化、自動調整,對數變換(Log10)、2次微分處理等演算處理中的1種或複數種的組合作為前處理。此外,在組合複數種演算處理時的處理順序,例如可採用對數變換(Log10)→二次微分→規格化(normalize)→自動調整(autoscale)。 As a pre-processing method for removing noise from the measured fluorescent fingerprints to obtain effective fluorescent fingerprint information, for example, non-fluorescent component removal processing, scattered light removal processing, and low sensitivity area deletion processing can be used One or a combination of plural. In addition, the obtained fluorescent fingerprint information can also be used in calculation processing such as centralization, normalization, standardization, base-line correction, smoothing, automatic adjustment, logarithmic transformation (Log10), and second-order differential processing. One kind or a combination of plural kinds is used as pretreatment. In addition, for the processing order when combining plural kinds of calculation processing, for example, logarithmic transformation (Log10)→secondary differentiation→normalization→autoscale can be used.
另外,作為除去對成分資訊無貢獻的波長之處理,例如可採用如下般之手法,各處理手法本身已為已知,並省略對其詳細之說明。 In addition, as the processing for removing the wavelength that does not contribute to the component information, for example, the following methods can be adopted, and each processing method itself is already known, and a detailed description thereof is omitted.
(a)Variable important projection(VIP) (a)Variable important projection(VIP)
(b)interval PLS(iPLS) (b)interval PLS(iPLS)
(c)Genetic algorithms(GA) (c) Genetic algorithms (GA)
(d)Jack-knife分析 (d) Jack-knife analysis
(e)Forward interval PLS (e)Forward interval PLS
(f)Backward interval PLS(biPLS) (f)Backward interval PLS(biPLS)
(g)Synergy interval PLS(siPLS) (g) Synergy interval PLS(siPLS)
(h)LASSO type method (h)LASSO type method
前處理的適用順序可適宜設定,但從處理效率化的觀點來看,較較係優先進行非螢光成分的除去處理、散射光的除去處理、低靈敏度區域的刪除處理等之處理。 The order of application of the pre-processing can be appropriately set, but from the viewpoint of efficiency of processing, processing such as removal processing of non-fluorescent components, removal processing of scattered light, and deletion processing of low-sensitivity regions is preferred.
如此般之前處理中係例如可使用Matlab或PLS_toolbox等專門軟體。 For such pre-processing, special software such as Matlab or PLS_toolbox can be used.
另外,於本態樣中,可僅使用自動調整作為前處理,但前處理未必只限於此手法。 In addition, in this aspect, only automatic adjustment may be used as the pre-processing, but the pre-processing is not necessarily limited to this method.
〔校準曲線的作成、驗證〕 〔Creation and verification of calibration curve〕
具體而言,校準曲線係將所取得的螢光指紋資訊作為解釋變數,並將4種TSNAs之總和量作為目的變數,使用例如PLS回歸分析(以下有時簡稱「PLS」)而作成。 Specifically, the calibration curve is obtained by using the obtained fluorescent fingerprint information as an explanatory variable and using the sum of the four TSNAs as the target variable, for example, using PLS regression analysis (hereinafter sometimes referred to as "PLS").
關於作成校準曲線之際使用之PLS回歸分析,先簡潔說明其概觀。 Regarding the PLS regression analysis used when creating the calibration curve, first briefly explain its overview.
PLS的解釋變數X(矩陣)與目的變數y(向量)滿足以下的二個基本式(1)、(2)。 The interpretation variable X (matrix) and the target variable y (vector) of the PLS satisfy the following two basic formulas (1) and (2).
X=TPT+E (1) X=TP T +E (1)
y=Tq+f (2) y=Tq+f (2)
在此,T為潛在變數(矩陣),P為載荷矩陣(loading)(矩陣),E為解釋變數X之殘差(矩陣),q為係數(向量),f為目的變數的殘差(向量),PT為P的轉置矩陣。 Here, T is the latent variable (matrix), P is the loading matrix (matrix), E is the residual (matrix) explaining the variable X, q is the coefficient (vector), and f is the residual of the target variable (vector) ), P T is the transpose matrix of P.
順帶一提,PLS並非將解釋變數X的資訊 直接用於目的變數y的模型,而是將解釋變數X的資訊之一部分轉換成潛在變數t,並使用潛在變數t來模擬目的變數y者。另外,潛在變數之數值,例如可藉由來自交叉驗證之預測的說明分散值作為指標來決定。此外,潛在變數有時亦稱為主成分。 Incidentally, PLS does not directly use the information of the explanatory variable X as the model of the target variable y, but converts a part of the information of the explanatory variable X into the latent variable t, and uses the latent variable t to simulate the target variable y. In addition, the value of the potential variable can be determined, for example, by using the predicted divergence value from cross-validation as an indicator. In addition, latent variables are sometimes called principal components.
特別是,在1成分模型的情況中,上述之(1)、(2)係表示為下列的(3)、(4)。 In particular, in the case of the one-component model, the above (1) and (2) are expressed as the following (3) and (4).
X=t1p1 T+E (3) X=t 1 p 1 T +E (3)
y=t1q1+f (4) y=t 1 q 1 +f (4)
在此,t1為潛在變數(向量),p1為載荷矩陣(向量),q1為係數(純量)。 Here, t 1 is a latent variable (vector), p 1 is a load matrix (vector), and q 1 is a coefficient (scalar).
現在,假設t1係以X的線形組合表示,則成立以下的(5)。 Now, assuming that t 1 is represented by a linear combination of X, the following (5) holds.
t1=Xw1 (5) t 1 =Xw 1 (5)
在此,w1為經規格化的加權向量。 Here, w 1 is a normalized weight vector.
PLS係求取如y與t1之共變異數yTt1在w1的範數(norm)為1(|w1|=1)的條件下會最大化之t1,要算出t1時,使用所謂拉格朗日(Lagrange)的未定乘數法即可。由於使用拉格朗日的未定乘數法之計算手法已為習知,故將計算的細節予以省略,僅將關於w1、p1、q1之計算結果以下列之(6)至(8)表示。 The PLS system is obtained with co-variance t y T y t 1 the number 1 (NORM) 1 w 1 under conditions FAN (| | w 1 = 1) will maximize the t 1, t 1 is calculated to be In this case, the so-called Lagrange's undetermined multiplier method may be used. Since the calculation method using Lagrange’s undetermined multiplier method is already known, the details of the calculation are omitted, and only the calculation results of w 1 , p 1 , q 1 are as follows (6) to (8 ) Means.
w1=XTy/|XTy| (6) w 1 =X T y/|X T y| (6)
p1=XTt1/t1 Tt1 (7) p 1 =X T t 1 /t 1 T t 1 (7)
q1=yTt1/t1 Tt1 (8) q 1 = yTt1/t 1 T t 1 (8)
另外,(7)、(8)式之t1係將以(6)式求得之w1代入(5)式而算出之向量。 In addition, t 1 in equations (7) and (8) is a vector calculated by substituting w 1 obtained by equation (6) into equation (5).
於多成分模型中亦可用相同的手法計算,計算手法係周知而省略其細節。 The same technique can also be used in multi-component models. The calculation technique is well known and its details are omitted.
作成校準曲線,且為了驗證所作成之校準曲線,將試料中的4種TSNAs之含量為已知的複數個樣本,區分並準備成用於作成校準曲線之校正(calibration)用樣本群,以及驗證校準曲線以確認有效性之驗證(validation)用樣本群。另外,本例中,作為校正用樣本準備有40個樣本,作為驗證用樣本準備有19個樣本。此外,雖然因VIP將波長限定於1000波長左右來進行,但本發明不限於此。 Create a calibration curve, and in order to verify the resulting calibration curve, the content of the four TSNAs in the sample is a known plurality of samples, and distinguish and prepare a sample group for calibration to make a calibration curve, and verify The calibration curve is used to verify the validity of the validation sample group. In this example, 40 samples are prepared as calibration samples, and 19 samples are prepared as verification samples. In addition, although the VIP limits the wavelength to about 1000 wavelengths, the present invention is not limited to this.
對校正用樣本群,例如適用上述之PLS回歸分析(例如,潛在因子數6),從所取得之螢光指紋資訊推測4種TSNAs之總和量,作成校準曲線。另外,用以作成校準曲線所使用之各樣本的4種TSNAs之總和量,係藉由下述方式求得:追加在事前以高效液體氣體層析(HPLC-MS/MS)定量之各樣本中的4種TSNAs之含量。 For the calibration sample group, for example, the above-mentioned PLS regression analysis (for example, the number of latent factors 6) is applied, and the total amount of the four TSNAs is estimated from the obtained fluorescent fingerprint information to prepare a calibration curve. In addition, the total amount of the four TSNAs for each sample used to prepare the calibration curve is obtained by adding the following to each sample quantified by high performance liquid gas chromatography (HPLC-MS/MS) beforehand The content of 4 kinds of TSNAs.
其次,針對驗證用樣本群,利用前述校準曲線從所取得之螢光指紋資訊推測4種TSNAs之總和量,進行校準曲線之驗證。 Secondly, for the verification sample group, the aforementioned calibration curve is used to estimate the total amount of the four TSNAs from the obtained fluorescent fingerprint information to verify the calibration curve.
第7圖係於橫軸取來自高效液體氣體層析(HPLC-MS/MS)之實測值(化學分析值),於縱軸取來自螢光指紋之4種TSNAs之總和量的推測值,對屬於驗證用樣本群之各樣本,將對應之點予以作圖之圖表。 Figure 7 is the measured value (chemical analysis value) from high performance liquid gas chromatography (HPLC-MS/MS) on the horizontal axis, and the estimated value of the sum of the four TSNAs from the fluorescent fingerprint on the vertical axis. Each sample belonging to the verification sample group will be plotted on the corresponding point.
於校正用樣本群中,決定係數R2=0.99(SEC=0.08μg/g),在化學分析值與來自校準曲線之推測值之間有高度相關,確定有良好的推測精度。此外,依據第7圖,於驗證用樣本群中之推測精度,決定係數R2=0.86(SEP=0.22μg/g),該校準曲線之有效性得到確認。 In the calibration sample group, the determination coefficient R 2 =0.99 (SEC=0.08μg/g), there is a high correlation between the chemical analysis value and the estimated value from the calibration curve, and it is determined that there is a good guess accuracy. In addition, according to Fig. 7, the estimation accuracy in the verification sample group is determined by the coefficient R 2 = 0.86 (SEP=0.22 μg/g), and the validity of the calibration curve is confirmed.
〔於4種TSNAs之含量為未知的試料(煙草原料)中4種TSNAs之總和量的推測〕 [Prediction of the total amount of 4 TSNAs in a sample (tobacco raw material) in which the content of 4 TSNAs is unknown]
使用有效性經確認之校準曲線,基於4種TSNAs之含量為未知的試料(煙草原料)的螢光指紋資訊,推測前述試料所含有之4種TSNAs之總和量。 Using the calibration curve with validated validity, based on the fluorescent fingerprint information of the sample (tobacco raw material) whose contents of the four TSNAs are unknown, the total amount of the four TSNAs contained in the aforementioned sample is estimated.
另外,針對4種TSNAs之含量為未知的試料,雖然可以省略對所取得之螢光指紋的前處理,但亦可進行與在取得校準曲線之際實施之前處理為相同內容的前處理。 In addition, for the samples of which the contents of the four TSNAs are unknown, the pre-processing of the obtained fluorescent fingerprint can be omitted, but it is also possible to perform the pre-processing that is the same as the pre-processing before the calibration curve is acquired.
〔於未知試料中4種TSNAs之各別之含量的推測〕 [Prediction of the respective contents of 4 TSNAs in unknown samples]
基於所推測之4種TSNAs之總和量及各TSNA之已知的存在比率,推測4種TSNAs之含量。 Based on the estimated total sum of the four TSNAs and the known existence ratio of each TSNA, the content of the four TSNAs is estimated.
順帶一提,有關4種TSNAs之含量,因為如前所述,各TSNA之存在比率於相同的煙草種中略為一定,故在將各TSNA之已知的存在比率設為a:b:c:d之情況時,可藉由將4種TSNAs之總和量各別乘以a/(a+b+c+d)、b/(a+b+c+d)、c/(a+b+c+d)、d/(a+b+c+d)來算出。 Incidentally, regarding the content of the four TSNAs, as mentioned above, the existence ratio of each TSNA is slightly constant in the same tobacco species, so the known existence ratio of each TSNA is set to a:b:c: In the case of d, you can multiply the sum of the four TSNAs by a/(a+b+c+d), b/(a+b+c+d), c/(a+b+ c+d) and d/(a+b+c+d).
請注意,本發明除了上述之實施態樣以外,於專利申請專利範圍所記載之技術的思想的範圍內亦可採用各種不同的實施態樣。 Please note that in addition to the above-mentioned embodiments, the present invention can also adopt various embodiments within the scope of the technical idea described in the patent application.
S01、S02、S03、S04、S05、S06、S07‧‧‧步驟 S01, S02, S03, S04, S05, S06, S07
Claims (12)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW107132274A TW202011014A (en) | 2018-09-13 | 2018-09-13 | Method, program and device for using fluorescence fingerprint information to quantify TSNAs contained in tobacco raw material |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW107132274A TW202011014A (en) | 2018-09-13 | 2018-09-13 | Method, program and device for using fluorescence fingerprint information to quantify TSNAs contained in tobacco raw material |
Publications (1)
Publication Number | Publication Date |
---|---|
TW202011014A true TW202011014A (en) | 2020-03-16 |
Family
ID=70766753
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW107132274A TW202011014A (en) | 2018-09-13 | 2018-09-13 | Method, program and device for using fluorescence fingerprint information to quantify TSNAs contained in tobacco raw material |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW202011014A (en) |
-
2018
- 2018-09-13 TW TW107132274A patent/TW202011014A/en unknown
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11946807B2 (en) | Hyperspectral imaging system for monitoring agricultural products during processing and manufacturing | |
Mishra et al. | Sequential fusion of information from two portable spectrometers for improved prediction of moisture and soluble solids content in pear fruit | |
Romero et al. | Internal soil standard method for the Brazilian soil spectral library: Performance and proximate analysis | |
Zamora-Rojas et al. | Handheld NIRS analysis for routine meat quality control: Database transfer from at-line instruments | |
EP2784484B1 (en) | Optimized calibration method for determining the content of chlorophyll using pre-processing and independent component analysis | |
Mishra et al. | Parallel pre-processing through orthogonalization (PORTO) and its application to near-infrared spectroscopy | |
Bevilacqua et al. | Application of near infrared (NIR) spectroscopy coupled to chemometrics for dried egg-pasta characterization and egg content quantification | |
Tahmasbian et al. | Laboratory-based hyperspectral image analysis for predicting soil carbon, nitrogen and their isotopic compositions | |
Bricklemyer et al. | Improved intact soil-core carbon determination applying regression shrinkage and variable selection techniques to complete spectrum laser-induced breakdown spectroscopy (LIBS) | |
Shetty et al. | Selection of representative calibration sample sets for near-infrared reflectance spectroscopy to predict nitrogen concentration in grasses | |
Jiang et al. | Using an optimal CC-PLSR-RBFNN model and NIR spectroscopy for the starch content determination in corn | |
EP3290908A1 (en) | Unknown sample determining method, unknown sample determining instrument, and unknown sample determining program | |
Mishra et al. | Improved prediction of minced pork meat chemical properties with near-infrared spectroscopy by a fusion of scatter-correction techniques | |
Mishra et al. | Pre-processing ensembles with response oriented sequential alternation calibration (PROSAC): A step towards ending the pre-processing search and optimization quest for near-infrared spectral modelling | |
JP7021755B2 (en) | Sample evaluation / estimation methods, programs, and equipment by fluorescent fingerprint analysis | |
Sun et al. | Variables selection for quantitative determination of cotton content in textile blends by near infrared spectroscopy | |
Zamora-Rojas et al. | Evaluation of a new local modelling approach for large and heterogeneous NIRS data sets | |
Davis et al. | Rapid, non-destructive, textile classification using SIMCA on diffuse near-infrared reflectance spectra | |
Divyanth et al. | Non-destructive prediction of nicotine content in tobacco using hyperspectral image–derived spectra and machine learning | |
Taous et al. | Geographical classification of authentic Moroccan Argan oils and the rapid detection of soya and sunflower oil adulteration with ATR-FTIR spectroscopy and chemometrics | |
TW202011015A (en) | Method, program and device for quantifying TSNAs contained in tobacco raw material by using fluorescence fingerprint information | |
Tena et al. | Assessment of vibrational spectroscopy performance in geographical identification of virgin olive oils: A world level study | |
Ahmed et al. | Enhancing corn quality prediction: Variable selection and explainable AI in spectroscopic analysis | |
TW202011014A (en) | Method, program and device for using fluorescence fingerprint information to quantify TSNAs contained in tobacco raw material | |
Nkansah et al. | Determination of concentration of ACQ wood preservative components by UV-Visible spectroscopy coupled with multivariate data analysis |