JP4614124B2 - Method and apparatus for determining blend ratio of blended rice - Google Patents

Description

  The present invention relates to a method and apparatus for determining a variety of rice varieties contained in blended rice and determining the mixing ratio thereof.

  Conventionally, some rice grain sellers have misrepresented rice blended with varieties other than good-tasting rice in good-tasting rice varieties represented by Koshihikari, Hitomebore, and Akitakomachi, which are traded at high prices. It became a problem that it was sold, and now, under the Food Law and the revised JAS Law, the labeling of rice varieties, origin, and year of production has become mandatory. In addition, in the government, in order to strengthen the control of false labeling of blended rice, inspections such as DNA identification and on-site inspection are performed on rice grain sellers.

  As a method for DNA identification of the presence or absence of mixing of other varieties with high-quality rice such as the conventional “Koshihikari”, and whether or not other varieties are mixed in a certain rice, the rice grain or processed product thereof When discriminating using the multiplex PCR method using DNA extracted from DNA as a template, the target variety does not express an identification band, but uses a negative band identification pair primer group that is selectively expressed only in a mixed variety. Known (Patent Document 1). However, this technique is a qualitative inspection technique for identifying whether or not different varieties are included, and the quantitative inspection for calculating the mixing ratio of rice of the varieties lacks accuracy.

  In other words, quantitative testing involves extracting dozens of rice grains at random from a rice bag, performing DNA analysis on each of these dozen grains to determine the variety name, and counting the number of grains to determine the mixing ratio. The reason for the lack of accuracy is to calculate the mixing ratio based on the ratio of the number of grains in the mixed rice process performed by the rice grain seller at the quantitative inspection stage. To do.

  FIG. 9 is a schematic view showing an example of a mixed rice metering device. This mixed rice metering device 100 includes a plurality of tanks 101A to F in which different varieties of rice are stored, and discharge valves 102A provided at the lower ends of the tanks. F, a traveling rail 103 laid on a floor (not shown), a weighing carriage 105 having a weighing means 104 and moving on the traveling rail 103, and the plurality of tanks placed on the weighing carriage 105 A container 106 that receives rice discharged from the container 101, a mixing device 107 that mixes the rice accumulated in the container 106, and a discharge basket 108 that discharges the rice mixed by the mixing device 107 as a refined product. is there.

  In this mixed rice weighing apparatus 100, for example, a case where a salesperson tries to fill a 10 kg-filled rice bag with 50% “Koshihikari” and 50% “Hitomebore” blended rice will be described. First, the weighing cart 105 movably mounted on the rail 103 is caused to travel and stopped at a position below the tank 101B in which “Koshihikari” is stored. Then, the valve 102B is operated by a measurement control unit or the like, and “Koshihikari” is put into the container 106 from the tank 101B. Then, when “Koshihikari” having a preset weight (in this case, 10 [kg] × 0.5 = 5.0 [kg]) is input to the measuring unit 104, control is performed to stop the valve 102B. Next, when measuring “pupil”, the weighing cart 105 is moved to the position where the tank 101A is provided, and “weight” is set in advance in the weighing means 104 (in this case, 10 [kg] × The valve 102A is controlled to measure only 0.5 = 5.0 [kg]). When the weighing is finished, the weighing cart 105 is moved to the vicinity of the mixing device 107, the rice in the container 106 is put into the mixing device 107, and the obtained blended rice is taken out from the refined product discharge basket 108 and filled into a rice bag.

  As described above, in the general mixed rice process, the mixed rice ratio is managed by the weight ratio, and when the mixing and mixing by the mixing device 107 is not sufficiently performed, the raw material is unevenly distributed in the refined rice bag. There is a risk. Therefore, if 48 varieties of rice grains are randomly selected and the varieties are determined to be 25 “Koshihikari” and 23 “Hitomebore” by DNA analysis, the mixing ratio of “Koshihikari” to “Hitomebore” is It is calculated as 52% vs. 48%, and an error occurs when compared with an actual mixing ratio of 50% vs. 50%.

In other words, it is dangerous to determine the mixing ratio of several tens of randomly extracted grains as the mixing ratio of the whole rice bag. In addition to the degree of stirring and mixing in the mixed rice process, The variation in weight is also a major factor in the inaccuracy of the mixing ratio.
JP 2003-135082 A

  In view of the above problems, the present invention is to provide a blended rice mixing ratio determination method and apparatus capable of reflecting a quantitative inspection performed by only a few dozen grain punching inspections in the mixing ratio of the population. Let it be an issue.

  In order to solve the above-mentioned problems, the present invention regards blended rice obtained by mixing a plurality of types of rice as a population, and extracts a plurality of sets from the population using a certain amount of rice as a sample sample. After measuring the weight of each rice, the DNA of each rice grain is extracted, and the rice is cultivated using the extracted DNA as a template to determine the variety of each rice grain. The weight of each rice grain corresponds to the weight. Calculate the mixing ratio for each of the plurality of sets of sample specimens from the variety and the number of grains counted based on the variety, and determine whether the plurality of sets of mixing ratios are data to be reflected in the population. Calculate using the test and the test of the population mean, and then use the binomial distribution test to calculate the probability of occurrence of the mixture ratios to determine whether the mixture ratio is valid. Mix proportions to population To re-display as a percentage, it took the technical means that.

  The appearance probability of the plurality of sets of mixing ratios may be calculated using a multinomial test.

  According to the present invention, by using only a few tens of grains extracted from a blended rice obtained by mixing a plurality of types of rice, it is estimated whether or not the sample is equally distributed, and the characteristics of the population distribution are estimated by a population average. Thus, the reliability of the extracted sample can be verified. And since the mixture ratio obtained by the conventional counting of the number of grains is recalculated using the binomial distribution, it can be determined how much the estimated mixture ratio can occur. That is, when a mixing ratio that cannot be normally obtained is obtained by counting the number of conventional grains, the mixing ratio can be redisplayed with higher accuracy by eliminating this by the significance level.

  In addition, the appearance probability of multiple sets of blending ratios can be tested even when blended rice with three or more varieties of rice is mixed using a multinomial test. Can be redisplayed well.

  Hereinafter, the best mode for carrying out the present invention will be described. As the target varieties of rice in the present invention, it is possible to analyze, for example, glutinous rice such as “Koshihikari”, “Hitomebore”, “Akitakomachi”, “Sasanishiki”, glutinous rice, brewery suitable rice, and glutinous rice. .

  The process of performing a quantitative inspection of these rice will be described with reference to FIG. First, 300 g to 10 kg, for example, are extracted as samples from a raw material mixed at a ratio of 50% “Koshihikari” and 50% “Hitomebore” by an appropriate sample extraction means (step 1). Next, the extracted sample is divided and reduced (step 2). For the division, for example, a sample disclosed in Japanese Utility Model Publication No. 63-23377 may be used to divide the sample 1: 1, and the 1: 1 division is repeated to reduce the sample to 48 particles. When the first sample put into the divider is 1 kg, it can be reduced to 48 grains by several tens of division operations. Then, the sample reduced to 48 grains is used as a sample reflecting the mixing ratio of the population, and the weight is measured one by one with an analytical balance (AW320, manufactured by Shimadzu Corporation) (step 3).

Next, DNA variety discrimination is performed for each grain by a known method. First, 1 grain of rice having been weighed and hard beads are put into a microtube, set in a sample crusher (manufactured by Kurabo Industries, SH-100) and pulverized (step 4). After grinding, DNA extraction is performed by the CTAB method (step 5). For example, a CTAB solution (0.1M
Tris-hydrochloric acid, 2 mM ethylenediaminetetraacetic acid (EDTA) disodium, 1.4 M NaCl (pH 8.0)) are added and stirred, and then placed in a thermostatic bath. Then, the CTAB solution is added again, and the mixture is allowed to stand for a predetermined time for extraction. Thus, genomic DNA can be obtained.

  Next, a PCR method using the genomic DNA as a template is performed in the presence of primers (step 6). This is performed in order to amplify a base sequence having discriminability between varieties, which is a base for designing the next PCR primer pair. Here, the PCR method in the present invention refers to a chain reaction of DNA replication by DNA polymerase. That is, about 90-96 ° C. high-temperature treatment process (denaturation), about 30-70 ° C. primer / DNA binding process (annealing), about 70-75 ° C. in the presence of heat-resistant DNA polymerase and primers. The DNA replication process (extension) is repeated for 20 to 60 cycles to increase (amplification) the reaction by about 1 million times to 1 billion times.

  Subsequently, the amplified DNA obtained by PCR is electrophoresed (step 7). This is for detecting a band indicating a base sequence in which discrimination between cultivars appears as a group for designing a target pairing primer in the amplification product. Here, electrophoresis means that DNA amplified by the PCR method is separated by the difference in molecular weight of DNA when it is moved by a direct current charge in gels of agarose or polyacrylamide, and is banded by ethidium bromide. An operation in which a difference in amplified DNA is detected as a band after staining.

  By analyzing the band pattern of the electrophoresis, the cultivar identification of one grain of rice put in the microtube is performed (step 8).

The results of the weight measurement of one rice grain in Step 1 to Step 3 and the variety identification in Step 4 to Step 8 are displayed by the apparatus shown in FIG.

  FIG. 2 shows a processing apparatus 1 for determining the mixing ratio. An information input unit 3 of a computer 2 in the processing apparatus 1 includes a weight measuring apparatus 4 for measuring the weight of each rice grain and an amplification obtained by the PCR method. An electrophoresis reading device 5 for reading DNA electrophoresis is connected. Further, the computer 2 discriminates the variety of each rice grain based on the information from the information input unit 3, and the variety discriminating unit 6 for recording the weight measured in advance for the one grain variety, The statistical processing unit 7 that performs statistical processing by summing up each cultivar and weight determined by the unit 6, the verification processing unit 8 that performs verification of the results calculated by the statistical processing unit 7, and the operator performs the verification The processing unit 8 includes a condition input unit 9 for setting conditions. The computer 2 is connected to a storage device 10 composed of a hard disk, an operation device 11 composed of a keyboard and a mouse, and a display device 12 composed of a CRT monitor or a liquid crystal monitor.

The operation of the processing apparatus 1 will be described. The processing apparatus 1 reduces the sample into 48 grains based on Step 1 to Step 2 in FIG. 1, and then weight information for each grain from the weight measuring apparatus 4 and product type information for each grain from the electrophoresis reader 5. Is input to the information input unit 3, and the product type discrimination unit 6 performs product type discrimination based on this information. The grains for which the type has been discriminated are subjected to aggregation processing in the statistical processing unit 7. This process is executed three times in different lots, and the result is displayed on the display device 12 (see Table 1).

  Next, the test processing unit 8 performs a test on the reliability of the data in Table 1. That is, the test processing unit 8 performs a binomial distribution test or a multinomial distribution test after performing an equal variance test and a population mean test.

  FIG. 3 is a flowchart showing equivariance test execution processing. First, in step 10, a null hypothesis that “there is no difference in variance” for the two groups to be tested and an alternative hypothesis that “there is a difference in variance” are set for the two groups.

Next, the sum of squares S of the samples to be compared is calculated (step 11), the average square V of the samples to be compared is calculated (step 12), and the ratio F ₀ value of the average square V to be compared is calculated (step 13). .

In the case of “no difference in variance” for the two groups to be tested, the F value follows an F distribution with degrees of freedom (n _x −1, n _y −1). The value in the F distribution table is compared with F ₀ calculated in step 13 (step 14). The value of F distribution table is a critical region of the null hypothesis, larger if the null hypothesis than the values of F ₀ is F distribution table is not rejected, the null hypothesis will not be denied.

As a result of the determination in step 14, when it is determined that the data in Table 1 is “difference in equivariance”, the verification operation is completed, and the process returns to step 1 in FIG. Step 15). If it is determined that “there is no difference in equal variance”, it is determined whether or not all test items have been completed for the two groups to be tested (step 16). For the two groups we want to test, see Table 1.
(1). Lot 1 “Koshihikari” and Lot 2 “Koshihikari”
(2). Lot 1 “Koshihikari” and Lot 3 “Koshihikari”
(3). Lot 2 “Koshihikari” and Lot 3 “Koshihikari”
(4). Lot 1 “Hitomebore” and Lot2 “Hitomebore”
(5). Lot 1 “Hitomebore” and Lot3 “Hitomebore”
(6). Lot 2 “Hitomebore” and Lot3 “Hitomebore”
(7). Lot 1 + 2 + 3 “Koshihikari” and Lot 1 + 2 + 3 “Hitomebore”,
7 items can be selected. When the equal variance test is completed for the above seven items, the process proceeds to the test of the population average of the next process.

  FIG. 4 is a flowchart showing the population average test execution process. First, in step 20, a null hypothesis that “the population mean is not different” for the two groups to be tested and an alternative hypothesis that “the population mean is different” are set for the two groups.

Next, an estimated value of the mother standard deviation is calculated from the mass average, the degree of freedom, and the mother standard deviation of the sample to be compared (step 21), and the t ₀ value of the sample to be compared is calculated (step 22).

In the case of “no difference in population mean” for the two groups to be tested, the t ₀ value follows a t distribution with degrees of freedom φ = n _x + _ny −2). From the t distribution table, the value at t (φ, 0.05) is compared with the absolute value of t ₀ calculated in step 22 (step 14). The value of the t distribution table is a rejection area of the null hypothesis. If the value of t ₀ is larger than the value of the t distribution table, the null hypothesis is not rejected and the null hypothesis is not denied.

As a result of the determination in step 23, when it is determined that the data in Table 1 is “difference in population mean”, the verification operation is completed, and the process returns to step 1 in FIG. Step 24). If it is determined that “the population mean is not different”, it is determined whether or not all the test items have been completed for the two groups to be tested (step 25). For the two groups you want to test,
(1). Lot 1 “Koshihikari” and Lot 2 “Koshihikari”
(2). Lot 1 “Koshihikari” and Lot 3 “Koshihikari”
(3). Lot 2 “Koshihikari” and Lot 3 “Koshihikari”
(4). Lot 1 “Hitomebore” and Lot2 “Hitomebore”
(5). Lot 1 “Hitomebore” and Lot3 “Hitomebore”
(6). Lot 2 “Hitomebore” and Lot3 “Hitomebore”
(7). Lot 1 + 2 + 3 “Koshihikari” and Lot 1 + 2 + 3 “Hitomebore”,
7 items can be selected. When the test of the population average is completed for the above seven items, the process proceeds to the binomial distribution test in the next step.

The binomial distribution means that when a probability that an event A is realized in one trial is p (0 <p <1), the trial is repeated n times independently, and the probability p that this event A is realized x times (X) refers to a distribution that follows the probability distribution determined by equation (1), and this distribution is determined by n and p.

  In the case of a finite population, this binomial distribution differs depending on whether it is a restoration extraction that returns to the population after sample extraction or a non-restoration extraction that does not return to the population. The distribution is consistent even with non-restored extraction. The population of the present embodiment (that is, the number of rice grains filled in the rice bag) is inherently finite and is non-restored extraction, so a hypergeometric distribution must be assumed, but compared with the number n of extracted grains. Since the number of grains in the population is very large, it is assumed that the population is regarded as an infinite population and follows a binomial distribution.

When n grains of rice are extracted from an infinite population, the probability that the A variety is x grains follows the binomial distribution of equation (2).

FIG. 5 is a flowchart showing a binomial distribution test execution process. First, assuming that the weight ratio per rice grain is the same, in step 30, the null hypothesis H ₀ that “the probability of being an A variety is half” and “the probability of being an A variety” It is set and the alternative hypothesis H ₁ to is not 1 of 2 minutes ".

Next, a test is performed for each lot shown in Table 1 (step 31). In the case of lot 1, the probability that 0 to 20 A variety (Koshihikari) is included in the extracted 48 grains is calculated by equation (3).

In the case of p = 1/2, the binomial distribution is symmetric with respect to the average. Therefore, the probability that 28 to 48 A varieties (Koshihikari) are included in the extracted 48 grains is the same as the expression (3). Become. Therefore, the probability that the probability that 0 to 20 A variety (Koshihikari) is included in the extracted 48 grains and the probability that 28 to 48 A variety (Koshihikari) is included is added to Equation (4). .

Next, a determination is made at a reference significance level of 5% for determining whether to reject the null hypothesis (step 32). Here, a significance level of 5% means that if the same test is performed, the conclusion obtained is incorrect once every 20 times. Fig. 6 is a graph showing the binomial distribution for 48 grains, but the rejection area with a significance level of 5% or less shown by diagonal lines "deny that A variety is mixed at 50%", and the null hypothesis is established Indicates not to. In the case of lot 2, the probability of adding 0 to 17 grains of A variety (Koshihikari) and 31 to 48 grains of A variety (Koshihikari) in the extracted 48 grains is 5 0.94%, which is barely above the 5% significance level. Table 2 below shows the number of grains from which A varieties to be rejected with a significance level of 5% or less are extracted.

  If the result of determination in step 32 is that the probability is below the significance level, the verification operation is terminated, and the process returns to step 1 in FIG. 1 to collect a sample again (step 33). If the probability exceeds the significance level, it is determined whether or not the test has been completed for all lots (step 35).

  As described above, when the test of equal variance, population mean, and binomial distribution is completed, the display device 12 shown in FIG. 2 corrects the result of Table 1 from the test result of the binomial distribution. “Koshihikari” 50%, “Hitomi” 50% can occur with a very high probability. ”

Next, consider the mixing of three varieties. Varieties A, B, because a mixture of 3 varieties and C, the proportion of each _breed, P _A, P _B, and were mixed so that the _{P C.} However, P _A + P _B + P _C = 1. When n grains are examined, it is n _A grains, n _B grains, and n _C grains, respectively, and a total of n grains are measured. A total of n grains are examined, and the probability that the A variety is nA grains, the B variety is nB grains, and the C variety is nC grains is expressed by equation (5).

FIG. 7 is a flowchart for performing a test execution process by mixing three types. First, in step 40, as initial data, the null hypothesis H ₀ become A variety blending ratio, B varieties blending ratio, and sets the C varieties blend ratio, A cultivar is the actual determination particle number, B varieties, C Enter the number of grains of each variety.

  Next, it is determined whether or not the sum of the blend ratios of the A, B, and C varieties is 100%, and the sum of the determined number of grains of the A, B, and C varieties is 48 (step). 41). If the sum of the blend ratios is not 100% or the sum of the determined number of grains is not 48, an error is displayed (step 42). When the sum of blend ratios is 100% and the sum of the number of judgment grains is 48, probabilities are calculated for all possible combinations (step 43), rearrangement in the order of probability, and data output of the probability list Is performed (step 44). Thereafter, whether or not to reject the null hypothesis is determined at a reference significance level of 5% (step 45). A graph of the probability distribution at this time is shown in FIG.

  When in the 5% region, the blend ratio assumed at the 5% risk rate (step 40) can be negated, so the test operation is completed and the process returns to step 1 in FIG. 46). If it is not in the 5% area, the blend ratio assumed at the 5% risk rate cannot be denied, so “the blend ratios of the A, B, and C types assumed in step 40 can occur with a very high probability”. It is displayed (step 47).

  Although FIG. 7 shows the test execution process by mixing three types, the test process can be executed only by adding the parameters of the type even if mixing four or more types, five types,..., N types. It becomes possible.

  As described above, the varieties, the number of grains and the weight of 48 samples extracted from blended rice are counted (see Table 1), and it is estimated whether or not the 48 samples are equally distributed for each lot. It is possible to estimate whether each lot is a population average or not, and to test the reliability of the extracted sample. Since the calculation of the mixing ratio is performed again using the binomial distribution or the multinomial distribution for each lot, it can be determined how much the estimated mixing ratio can occur. In other words, when a mixing ratio that is not normal can be obtained by counting the number of grains as shown in Table 1, this can be eliminated depending on the significance level, and the mixing ratio can be displayed again with higher accuracy. .

It is a flowchart which shows the process of the method of determining the mixing ratio of blended rice. It is a block diagram of the processing apparatus which determines the mixing ratio of blended rice. It is a flowchart which shows equivalence test execution processing. It is a flowchart which shows the test execution process of a population average. It is a flowchart which shows the test execution process of binomial distribution. It is a graph which shows binomial distribution about 48 grains. It is a flowchart which performs the test | inspection execution process by 3 types mixing. It is a graph which shows probability distribution of 3 kinds blend rice. It is the schematic which shows an example of a mixed rice measuring device.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Processing apparatus 2 Computer 3 Information input part 4 Weight measuring apparatus 5 Electrophoresis reader 6 Kind discrimination | determination part 7 Statistical processing part 8 Test | inspection processing part 9 Condition input part 10 Memory | storage device 11 Operation apparatus 12 Display apparatus

Claims (4)

Considering blended rice that is a mixture of multiple types of rice as a population, a plurality of sets of samples are extracted from the population using a certain amount of rice as a sample,
After measuring the weight of each grain of rice for the sample sample, DNA for each grain of rice is extracted, and using the extracted DNA as a template, the variety of rice grains is determined using the PCR method,
Calculate the mixing ratio for each of the plurality of sets of sample samples from the weight per rice grain, the variety corresponding to the weight and the number of grains counted based on the variety,
It is calculated using the test of equal variance and the test of population mean whether or not the data is reflected in the population with respect to the mixture ratio of the plurality of sets,
Then, using a binomial distribution test, calculate the probability of appearance of the multiple sets of mixing ratios to determine whether the mixing ratio is valid,
A blending ratio determination method for blended rice, wherein an effective blending ratio is redisplayed as a blending ratio of a population.   The blended rice mixture ratio determination method according to claim 1, wherein the appearance probability of the plurality of sets of mixture ratios is calculated using a multinomial test. A sample extraction means that regards blended rice obtained by mixing a plurality of types of rice as a population, and extracts a plurality of sets from the population using a certain amount of rice as a sample sample,
A weight measuring means for measuring the weight of each sample of rice for each grain;
An electrophoresis reading means for extracting DNA for each grain of rice, and reading electrophoresis of amplified DNA obtained by PCR using the extracted DNA as a template;
Variety discriminating means for discriminating the variety of one grain of rice by analyzing the electrophoresis band pattern obtained from the variety information;
A statistical processing means for calculating a mixing ratio for each of the plurality of sets of sample samples from the weight per rice grain, the variety corresponding to the weight, and the number of grains counted based on the variety;
After calculating using the test of equal variance and the test of population mean whether or not the data reflected in the population with respect to the mixture ratio of the plurality of sets, the mixture ratio of the plurality of sets using a binomial test A test processing means for calculating the occurrence probability of and determining whether or not the mixing ratio is valid;
Display means for redisplaying the effective mixture ratio determined by the test processing means as the mixture ratio of the population;
An apparatus for determining a blend ratio of blended rice, comprising: 4. The blended rice mixture ratio determining apparatus according to claim 3, wherein the verification processing means calculates the appearance probability of the plurality of sets of mixing ratios using a multinomial distribution test.

Images