JP2021052696A - Method for estimating concentration of target substance and concentration estimation device - Google Patents

Abstract

To provide a concentration estimation method and device which can estimate efficiently and accurately the concentration of a target substance since a volume dispersion correction can be performed when the volume dispersion correction is effective.SOLUTION: A method comprises: an acquisition step of acquiring a number N of a plurality of subsections formed by dividing a sample containing a target substance and a number Nq of subsections not containing the target substance; and an estimation step of estimating from the N and the Nq the concentration of the target substance contained in the sample. The method further comprises a determination step of determining whether it is performed in a normal mode in which the estimation is performed without correcting the volume dispersion of the plurality of subsections or a volume dispersion correction mode in which the estimation is performed by correcting the volume dispersion of the plurality of subsections.SELECTED DRAWING: Figure 9

Description

The present invention relates to a method for estimating the concentration of a target substance such as nucleic acid, and a concentration estimation device.

In recent years, the development of a technique using the concentration of a nucleic acid (target nucleic acid) having a specific base sequence has been promoted when identifying a pathogen and determining a treatment policy for cancer. As a method for quantitative analysis of a target nucleic acid, a method combining a PCR (Polymerase Chain Reaction) method for amplifying the number of nucleic acids and a fluorescent probe which emits fluorescence when bound to the target nucleic acid has been energetically developed. Since there is a correlation between the concentration of the target nucleic acid and the concentration of the dye that emits fluorescence, the concentration of the target nucleic acid can be measured from the fluorescence intensity. However, in general, it is difficult to accurately measure the gradation because the fluorescence intensity is weak.

As a solution to this problem, a digital PCR (hereinafter referred to as dPCR) method, which does not require measurement of the gradation of fluorescence intensity, is known. In the dPCR method, the solution containing the target nucleic acid is divided into a large number of small compartments, and PCR is performed for each compartment to check for the presence or absence of fluorescence emission. The target nucleic acid concentration is estimated from the ratio of the number of compartments that do not emit fluorescence to the total number of compartments.

Non-Patent Document 1 discloses a nucleic acid concentration measuring device using the dPCR method. Among them, the method of using a large number of small droplets generated in the oil as small compartments is advantageous in reducing running cost and miniaturization. In the method using droplets as subsections, the target nucleic acid concentration is estimated using the Poisson distribution, but this estimation method assumes that the droplet sizes are the same.

Non-Patent Document 2 discloses a method for estimating the target nucleic acid concentration when the volume of the subsection can be regarded as a Gaussian distribution. Further, Patent Document 1 sets a correction factor that depends on the volume fluctuation value (coefficient of variation, CV, feedback of variation) of the volume of the small compartments as compared with the average volume, to the ratio of the number of compartments that do not emit fluorescence to the total number of compartments. Disclose a method of correcting by multiplying. Specifically, a method of adapting the volume distribution of a subsection to a model curve and estimating a correction factor using the model curve is disclosed.

When the shape of the subsection is a sphere, the contour image of the subsection may be acquired, the diameter of each subsection may be extracted from the contour image, and the volume of the subsection may be obtained. Image processing methods such as binarization processing and template matching using a circular pattern as a template can be used for diameter extraction.

Japanese Unexamined Patent Publication No. 2018-46819

L. Cao et al. , Biosensors and Bioelectronics, vol. 90, pages459-474 (2017). N. Majumdar et al. , Scientific Reports, vol. 7. Article number: 9617 (2017).

In the dPCR method, the higher the number of measured sub-compartments, the higher the reliability of the measured value. Generally, the treatment is performed using a few thousand to 10,000 sub-compartments. In order to correct the volume dispersion of the subsections of spheres and estimate the concentration of the target nucleic acid, it is desirable to know the volume of all spheres, i.e. the diameter. The diameter of the sphere is about several tens of μm to 200 μm, and in order to obtain the diameter of the sphere accurately, it is necessary to acquire a contour image of the sphere at a high magnification. It takes time to acquire the contour image. Further, in order to perform image processing efficiently, it is desirable to acquire a contour image and a fluorescence image with the same optical system, but if the fluorescence image is acquired at a high magnification, it is necessary to lengthen the exposure time, and the measurement time is further increased. It becomes a factor to increase.

Therefore, the present invention provides a method (device) for estimating the concentration of a target substance that can efficiently and accurately estimate the concentration of a target by performing the volume dispersion correction when the effect of the volume dispersion correction is effective. The purpose.

The method for estimating the concentration of a target substance according to the present invention includes an acquisition step of acquiring the number N of a plurality of subdivisions formed by dividing a sample containing the target substance, the number Nq of the subdivisions not containing the target substance, and the above-mentioned. An estimation method including an estimation process for estimating the concentration of the target substance contained in the sample from N and the Nq, and a normal mode in which the estimation is performed without correcting the volume dispersion of the plurality of subsections. It is characterized by having a determination step of determining which of the volume dispersion correction modes in which the estimation is performed by correcting the volume dispersion of the plurality of subsections.

The target substance concentration estimation device according to the present invention comprises an acquisition means for acquiring the number N of a plurality of subdivisions formed by dividing a sample containing a target substance, the number Nq of the subdivisions not containing the target substance, and the above. An estimation device having an estimation means for estimating the concentration of the target substance contained in the sample from N and the Nq, and a normal mode in which the estimation is performed without correcting the volume dispersion of the plurality of subsections. It is characterized by having a determining means for determining which of the volume dispersion correction modes in which the estimation is performed by correcting the volume dispersion of the plurality of subsections.

According to the target substance concentration estimation method and the concentration estimation device according to the present invention, the volume dispersion correction can be performed when the volume dispersion correction is effective, so that the concentration of the target substance can be estimated efficiently and accurately. be able to.

Monte Carlo simulation results verifying the method for estimating the concentration of the target substance according to the embodiment of the present invention. Monte Carlo simulation results verifying the method for estimating the concentration of the target substance according to the embodiment of the present invention. Monte Carlo simulation results verifying the method for estimating the concentration of the target substance according to the embodiment of the present invention. Monte Carlo simulation results verifying the method for estimating the concentration of the target substance according to the embodiment of the present invention. Monte Carlo simulation results verifying the method for estimating the concentration of the target substance according to the embodiment of the present invention. Monte Carlo simulation results verifying the method for estimating the concentration of the target substance according to the embodiment of the present invention. Conceptual diagram of the target concentration measuring apparatus according to the first embodiment of the present invention. The figure explaining the structure of the inspection plate in 1st Example of this invention. The figure explaining the flow of image processing in 1st Example of this invention The figure explaining the fluorescence intensity distribution of the particle in 1st Example of this invention. Conceptual diagram of the target concentration measuring device in the second embodiment of the present invention Outline image in the second embodiment of the present invention The figure explaining the flow of image processing in the 2nd Example of this invention The figure explaining the volume distribution of the particle in the 2nd Example of this invention. The figure explaining the flow of image processing in the 3rd Example of this invention

(Concentration estimation method)
The method for estimating the concentration of a target substance according to the embodiment of the present invention has the following steps.
(1) An acquisition step of acquiring the number N of a plurality of subsections formed by dividing a sample containing a target substance and the number Nq of the number of subsections not containing the target substance.
(2) An estimation process for estimating the concentration of the target substance contained in the sample from N and Nq.
(3) It is determined whether the estimation is performed in the normal mode in which the estimation is performed without correcting the volume dispersion of the plurality of subsections or the volume dispersion correction mode in which the estimation is performed by correcting the volume dispersion of the plurality of subsections. Decision process.

The volume dispersion correction mode is a useful method when the number of subdivisions is large and the concentration of the target substance can be estimated accurately by correcting the volume dispersion. However, the volume dispersion correction mode has a feature that it takes a relatively long time to estimate.

On the other hand, the normal mode is a useful method when the concentration of the target substance can be estimated accurately without correcting the volume variance, such as when the number of subsections is small. The normal mode is characterized in that it takes less time to estimate than the volume dispersion correction mode.

The method for estimating the concentration of the target substance according to the present embodiment can switch between the above two modes depending on how many subdivisions the sample to be estimated is divided into. As a result, it is possible to achieve both high estimation accuracy and short estimation time.

The determination step may include a step of automatically determining whether the estimation is performed in the normal mode or the volume dispersion correction mode.

In addition, the determination step may determine that the estimation is performed in the volume dispersion correction mode when the number of subsections containing the target substance is equal to or greater than a predetermined value. Here, the predetermined value can be arbitrarily set, but for example, the predetermined value can be set to 400 and 1600.

The determination step can determine that the estimation is performed in the volume dispersion correction mode when the ratio of the number of compartments to the number of compartments not containing the target substance is less than or equal to the reference value.
The determination step may also include determining whether the estimation is to be performed in normal mode or volume dispersion correction mode, based on information about the correction coefficients used to correct the volume dispersion of the plurality of compartments. .. Here, as information on the correction coefficient, examples thereof include a correction coefficient described later, a correction amount (deviation amount) obtained by subtracting the correction coefficient from 1.

Further, the volume dispersion correction mode can be estimated by any of the following modes.
(I) A normal distribution correction mode in which the volume probability distributions of a plurality of subsections are estimated as being normal distributions.
(Ii) An arbitrary distribution correction mode in which an estimation is performed using the number of subsections containing no target substance obtained based on the information on the correction coefficient.

Although the case where the determination process is automatically performed has been described, the process of determining whether to perform the estimation in the normal mode or the volume dispersion correction mode based on the instruction received from the user (worker) may be included. Good. Instructions from the user can be accepted using a switch or the like.

(Concentration estimation device)
The target substance concentration estimation device according to the embodiment of the present invention has the following configurations.

(1) An acquisition means for acquiring the number N of a plurality of subsections formed by dividing a sample containing a target substance and the number Nq of the number of subsections not containing the target substance.

(2) An estimation means for estimating the concentration of the target substance contained in the sample from N and Nq.

(3) It is determined whether the estimation is performed in the normal mode in which the estimation is performed without correcting the volume dispersion of the plurality of subsections or the volume dispersion correction mode in which the estimation is performed by correcting the volume dispersion of the plurality of subsections. Determining means.

The concentration estimation device for the target substance according to the embodiment of the present invention can achieve both high estimation accuracy and short estimation time, as in the above concentration estimation method.

In addition, the determination means can perform a step of automatically determining whether the estimation is performed in the normal mode or the volume dispersion correction mode.

In addition, the determination means can determine that the estimation is performed in the volume dispersion correction mode when the number of subsections containing the target substance is equal to or greater than a predetermined value. Here, the predetermined value can be arbitrarily set.

Further, the determination means can determine that the estimation is performed in the volume dispersion correction mode when the ratio of the number of subsections to the number of subsections not containing the target substance is equal to or less than the reference value.

Further, the determining means can determine whether the estimation is performed in the normal mode or the volume dispersion correction mode based on the information about the correction coefficient used to correct the volume dispersion of the plurality of subsections. Here, as information on the correction coefficient, examples thereof include a correction coefficient described later, a correction amount (deviation amount) obtained by subtracting the correction coefficient from 1.

The volume dispersion correction mode can be estimated by any of the following modes.
(I) A normal distribution correction mode in which the volume probability distributions of a plurality of subsections are estimated as being normal distributions.
(Ii) An arbitrary distribution correction mode in which an estimation is performed using the number of subsections containing no target substance obtained based on the information on the correction coefficient.

It should be noted that the determination means can determine whether to perform the estimation in the normal mode or the volume dispersion correction mode based on the instruction received from the user.

Hereinafter, the concentration estimation method (measurement method) and the estimation device (measurement device) according to the embodiment of the present invention will be described in detail by taking a case where a nucleic acid such as DNA is targeted as an example.

In the dPCR method, a solution (sample) containing a target nucleic acid is divided into a large number of subsections, and the PCR method is performed for each subsection to check for the presence or absence of fluorescence emission. Then, the target nucleic acid concentration is estimated from the ratio of the number of small compartments that do not emit fluorescence to the total number of small compartments.

<When the volume of the subsection is uniform: equivalent to normal mode>
Consider the case where the size of the subsections is uniform. Assuming that the total number of compartments (total number of compartments) is N, the number of compartments that do not emit fluorescence is N _q , the negative compartment ratio q = N _q / N, and the volume of each compartment is v ₀ , the target nucleic acid concentration f is It can be expressed by the following equation 1.

The above equation 1 can be derived as follows.

The probability that one nucleic acid does not enter the space of volume v ₀ in the volume (total volume) V (= Nv _{0) of all subsections is}

である。核酸の存在確率が互いに独立である時、Ｍ個の核酸が、総体積Ｖ（＝Ｎｖ_０）の中の体積ｖ_０の空間に全く入らない確率は、

Is. When the existence probabilities of nucleic acids are independent of each other, the probability that M nucleic acids do not enter the space of volume v _{0 in the total volume V (= Nv 0) is}

である。ここで、ｖ_０≪Ｖであり、Ｖ＝Ｎｖ_０であるので、式３は下記式４のように書き直すことができる。

Is. Here, since v ₀ << V and V = Nv ₀ , Equation 3 can be rewritten as Equation 4 below.

All N compartments have this probability that the target nucleic acid is absent. Therefore, this value is equal to the number of compartments in which the target nucleic acid does not exist (does not fluoresce) divided by the total number of compartments, that is, the negative compartment ratio (Equation 5).

From Equation 5, the target nucleic acid concentration f is

と表すことができ、ｆは式１のように表すことができる。

Can be expressed as, and f can be expressed as in Equation 1.

The "normal mode" in the present embodiment means that the target nucleic acid concentration is calculated by the above formula 6.

<When the volume variance of the subsection is Gaussian distribution: Corresponds to the normal distribution correction mode>
In a subdivision group having an average volume v ₀ , the total number of subdivisions N, and a total volume V (= Nv ₀ ), the probability that one nucleic acid does not fit in a subdivision having a volume v is

である。Ｍ個の核酸が、この小区画に全く入らない確率は、

Is. The probability that M nucleic acids will not fit in this compartment at all is

である。

Is.

The volume probability distribution p (v) of a subdivision group having a Gaussian distribution (normal distribution) with a mean v _{0 and a standard deviation σ is}

で表すことができる。

Can be represented by.

The negative subdivision ratio can be expressed as in Equation 10.

Equation 10 can be transformed as follows.

This is a quadratic equation for M / N, the solution of which is

で与えられる。ここで、体積変動係数ｃｖ（ただしｃｖ＝σ／ｖ_０）で置き換えている。

Given in. Here, it is replaced by the volume coefficient of variation cv (where cv = σ / v _0).

From this, the target nucleic acid concentration f is

と表すことができる。これは、式６と比較して、体積分散の補正が含まれている、と見なすことができる。

It can be expressed as. This can be considered to include the correction of volume variance as compared to Equation 6.

The "normal distribution correction mode" in the present embodiment means that the target nucleic acid concentration is calculated by the above formula 13.

The above equation 10 can be rewritten as x = (Mcv) / N as follows.

For comparison, consider the case where the volume variance is _{a uniform distribution with mean v 0} and standard deviation σ. Although the details are omitted, the negative subdivision ratio for the uniform distribution can be expressed as follows.

Similarly, for comparison, _{consider the case where the volume variance is a binary distribution with mean v 0} and standard deviation σ. Although the derivation of details is omitted, the negative subdivision ratio for the binary distribution can be expressed as follows.

Comparing the above equations 14, 15 and 16, the second term in parentheses (the squared term of x) after the exp term matches. This means that if x is small to some extent and the fourth or more terms of x can be ignored, even if the volume variance is a uniform distribution or a binary distribution, it can be corrected in the "normal distribution correction mode". Further, when x is smaller and the squared term or more of x can be ignored, it means that the volume dispersion correction is meaningless and the "normal mode" may be used. On the contrary, when x is large and x 4 or more terms cannot be ignored, if the volume variance deviates from the Gaussian distribution (normal distribution), it may not be possible to correct in the "normal distribution correction mode". In this case, the "arbitrary distribution correction mode" described below is effective.

<When the volume variance of the small section is irregular: Corresponds to the arbitrary distribution correction mode>
Consider the case where the volume of each compartment is known but its variance is amorphous. The negative subdivision ratio can be expressed as follows, where the volume probability distribution of the subdivision group is p (v).

However, Δv = v−v ₀ .

Here, consider the volume dispersion correction coefficient w (v) given by the following equation, which is determined according to the volume of each subsection.

Then, the integrand of Equation 20 is multiplied by this volume dispersion correction coefficient to integrate. Let the result be q'.

q'represents the true value. However, the volume dispersion correction coefficient includes M / N, which is an unknown value.

Therefore, the following processing is performed.

_{(1) First, the first negative subdivision ratio q 1} is calculated from the number of negative subdivisions / the total number of subdivisions.

(2) Substituting into Equation 18 as M / N = -log (q ₁ ), the volume dispersion correction coefficient w (v) determined according to the volume of the subsection is determined.

(3) Multiply the volume dispersion correction coefficient to perform integration, and calculate the _{second negative subdivision ratio q 2.}

(4) If necessary, multiply by a new volume dispersion correction coefficient and perform integration to calculate the _{mth negative subdivision ratio q m.}

By performing such an operation, the negative compartment ratio closer to the true value, that is, the target nucleic acid concentration can be estimated.

In fact, in the dPCR method, each compartment shows a binary value of negative (without target nucleic acid) or positive (with target nucleic acid). Based on the above ideas (1) to (4), the inventor has discovered that a result close to the true value can be obtained by performing the processing by the following method.

(A) The number of negative subdivisions (N _q1 ) not including the target of interest is counted to determine the _first negative subdivision ratio q1.

(B) Substituting into Equation 18 as M / N = -log (q ₁ ), the first volume dispersion correction coefficient w ₁ (n) for each subsection is determined. Here, n is the number of the subsection n, Δv = v (n) −v ₀ , and v (n) is the volume of the subsection n.

(C) When counting the small sections of the negative, the volume dispersion correction coefficient is multiplied as a weight to count. That is, the sum of the volume dispersion correction coefficients for all the compartments not containing the target nucleic acid is taken. Let this be the number of second negative subdivisions N _q2 . When the subsection numbers 1 to N _q1 are renumbered as negative subsections and the subsection numbers N _q1 +1 to N are positive subsections, it can be expressed as the following equation.

(D) The second negative subdivision ratio q ₂ is determined.

The following (e), (f), and (g) are repeatedly executed as necessary.

(E) Substituting into Equation 18 as M / N = -log (q _m-1 ), the volume dispersion correction coefficient w _m-1 (n) for each subsection is determined.

(F) Sum the volume dispersion correction coefficients for all compartments that do not contain the target nucleic acid. Let this be the number of mth negative subsections N _qm .

(G) The mth negative subdivision ratio q _m is determined.

(H) if negative cubicle number N _qm, it is negative cubicle rate q _m converges, and outputs a value obtained by dividing -log a (q _m) in average volume of all the small compartment as the target nucleic acid concentration.

By performing such an operation, even if the volume dispersion of the subdivisions is indefinite, the negative subdivision ratio closer to the true value, that is, the target nucleic acid concentration can be estimated.

<Judgment of necessity of volume dispersion correction by confidence interval>
It is assumed that a sample of volume v is taken out from a liquid (population) in which X target nucleic acids are dispersed in volume V. For example, if a sample is extracted multiple times and the number of target nucleic acids contained therein is measured, which is the case when minute blood is taken out from the human body, in the case of v << V, the number of target nucleic acids in the sample is measured. The distribution can be approximated to a normal distribution with an expected value of Xv / V and a variance of Xv / V. That is, the number of target nucleic acids in a sample from a trial has a 95.4% probability of being between Xv / V ± 2√ (Xv / V).

Generating a plurality of small sections from a portion of a liquid, as a result of fluorescence measurement by dPCR method, the number of small sections containing the target nucleic acid (positive small sections number) N _p, the number of target nucleic acid in the total measured Is M. This number N _p or M is a numerical value including statistical variation, and when considered in 2σ, the true value of the number of target nucleic acids exists between M ± 2√M with a probability of 95.4%. ..

_{N p} ≤ M because a plurality of target nucleic acids may be contained in one subsection, but for convenience, it is considered as _{N p = M.}

If positive small section number _{N p} is 400, statistically, a number that includes a dispersion of 40 at 2 [sigma] (10% in the ratio), 440 pieces of 360 becomes 95.4% confidence interval. Also, if positive cubicle number _{N p} of 1600, statistically, a number that includes a dispersion of 80 at 2 [sigma] (5% in the ratio), 1680 from 1520 pieces is a 95.4% confidence interval Become.

Therefore, the accuracy error of the target nucleic acid concentration, if 10% or less ± 95.4% confidence interval, it is desirable that 400 or more as a positive cubicle number N _p. Further, if the ± 5% or less 95.4% confidence interval of accuracy errors of a target nucleic acid concentration is desirably 1600 or more as a positive cubicle number N _p.

That is, when the number of positive compartments is equal to or more than a certain reference value, the accuracy of the target nucleic acid concentration can be expected to be high. Therefore, the volume dispersion correction processing is performed. When the number is less than a certain reference value, the accuracy of the target nucleic acid concentration is low. It is possible to determine that the volume dispersion correction process is not performed.

Further, the configuration may be such that the operator can arbitrarily set the number of positive subdivisions that determine the necessity of the volume dispersion correction processing.

<Judgment of necessity of volume dispersion correction based on negative subdivision ratio>
The volume of the parcel cannot be negative. As an extreme example, if half the volume of a plurality of compartments is 0 and the volume of the other half is 2v ₀ , the volume variance of the compartments is mean v ₀ , standard deviation v ₀ , cv = 1. Become. Actually, the volume does not become 0, and there is no problem even if it is considered that the cv value becomes 1 or less as an actual model.

The fraction at the end of Equation 13 is the correction coefficient for volume variance. When this correction coefficient can be regarded as 1, the processing in the "normal mode" is appropriate, and when it is significantly larger than 1, it is appropriate to perform the volume dispersion correction processing. Let a be the amount of deviation from 1, that is, the amount of correction, and derive the relationship between a and q.

Equation 32 is expanded and organized as follows.

Table 1 shows the values of q when a and cv are changed.

That is, if the accuracy error of the measurement result is aimed at 1% or less, it is better to perform the volume dispersion correction process when the correction amount exceeds 1%. Therefore, if the volume dispersion of the measurement target can be predicted empirically in advance, the necessity of the volume dispersion correction processing may be determined based on the equation 33.

For example, empirically, when the volume variance of the measurement target can be regarded as cv = 0.2, if the accuracy error of the measurement result is aimed at 1% or less, the volume when the negative subdivision ratio q is 0.613 or less. It is better to perform the dispersion correction process. Similarly, when the accuracy error is 5% or less, it is possible to determine that the volume dispersion correction process should be performed with q ≦ 0.104, and when the accuracy error is 10% or less, the volume dispersion correction process should be performed with q ≦ 0.016.

In addition, the configuration may be such that the operator can arbitrarily set the negative subdivision ratio that determines the necessity of the volume dispersion correction processing.

<Verification by Monte Carlo simulation: Gaussian distribution>
Prepare a group of 500,000 subdivisions distributed in a Gaussian distribution with an average volume of v ₀ = 113 pl, a standard deviation of σ = _{0.2 v 0, and a cv value of 0.2.} Then, a population in which a predetermined number of target nucleic acids are randomly distributed is prepared.

A sample obtained by randomly extracting a predetermined number of subdivisions from the population is prepared, and the target nucleic acid concentration is calculated according to the following method. This is tried 100 times and its mean, standard deviation, maximum and minimum values are plotted.

(Non) No volume dispersion correction According to Equation 6, the target nucleic acid concentration is calculated from the ratio of the number of negative compartments to the total number of compartments. (A) Normal distribution correction The ratio of the number of negative compartments to the total number of compartments, and the correction when the distribution is Gaussian. Calculate the target nucleic acid concentration according to the formula (Equation 12) (B) Arbitrary distribution correction According to formulas 26, 27 or 29, 30, the number of negative subdivisions is counted in consideration of the correction coefficient of the negative subdivisions, and the target nucleic acid concentration is calculated. Calculation In this example, since the volume variance of the subsection is Gaussian, it can be corrected to the same extent by either (A) normal distribution correction or (B).

(1) When 5000 target nucleic acids are randomly distributed The results of the calculated target nucleic acid concentrations are shown in FIG. (A) is a case where the number of subsections of the sample is 1000, and similarly, (b) is 2000, (c) is 5000, (d) is 10000, and (e) is 20000. f) is the case of 50,000 pieces.

Further, in the figure, (Non) is the case without volume dispersion correction, (A) is the case where normal distribution correction is performed, and (B1) to (B4) are cases where correction is performed according to the present embodiment. Means the number of repetitions. In addition, the vertical axis represents the target nucleic acid density standardized by a true value.

Further, in the figure, ◯ represents the average value of the results of 100 trials, and the error bar represents ± standard deviation × 2 (2σ). Further, x at the bottom of the mean value is the minimum value, and x at the top of the mean value is the maximum value.

In this example, from (a) to (f), the variation due to the increase in the number of positive subsections is reduced, that is, the confidence interval range is reduced. However, the negative subdivision ratio is about 0.9, which is larger than the threshold value of 0.613 of 1% in Table 1, and the correction amount is less than 1% even if corrected, and there is no need to correct. In fact, no significant difference can be obtained when comparing with and without correction.

(2) When 25,000 target nucleic acids are randomly distributed The results of the calculated target nucleic acid concentrations are shown in FIG. The notation rules are the same as in FIG.

In this example, the negative parcel ratio is 0.608, which is smaller than the 1% threshold of 0.613 in Table 2. Therefore, when aiming for a value accuracy error of 1% or less, it is better to perform correction. Further, from (a) to (f), a decrease in variation due to an increase in the number of positive subdivisions, that is, a decrease in the confidence interval range is observed, and in particular, from (c) where the number of positive subdivisions exceeds 1600 (c). In f), the effect of correction is remarkable.

(3) When 50,000 target nucleic acids are randomly distributed The results of the calculated target nucleic acid concentrations are shown in FIG. The notation rules are the same as in FIG.

In this example, the negative subdivision ratio is 0.375, which is a correction amount of about 2% from Table 2. The effect of the correction is remarkable in any of (a) to (f).

<Verification by Monte Carlo simulation: uniform distribution>
It assumes 500,000 small parcels which is uniformly distributed from 0.1 v ₀ to 1.9V ₀ in average volume _v 0 = _113pl. At this time, the standard deviation σ is 0.52v ₀ and the cv value is 0.52. Then, a population in which a predetermined number of target nucleic acids are randomly distributed is prepared.

A sample obtained by randomly extracting a predetermined number of subdivisions from the population is prepared, and the target nucleic acid concentration is calculated according to the following method. This is tried 100 times and its mean, standard deviation, maximum and minimum values are plotted.

(1) When 2500 target nucleic acids are randomly distributed, the results are shown in FIG. The notation rules are the same as in FIG.

In this example, from (a) to (f), the variation due to the increase in the number of positive subsections is reduced, that is, the confidence interval range is reduced. However, the negative subdivision ratio is 0.952, which is larger than the threshold value of 0.930 of 1% in Table 1, and the correction amount is less than 1% even if corrected, and there is no need to correct. In fact, no significant difference can be obtained when comparing with and without correction. Moreover, there is no significant difference between the normal distribution correction and the arbitrary distribution correction.

(2) When 5000 target nucleic acids are randomly distributed, the results are shown in FIG. The notation rules are the same as in FIG.

In this example, the negative parcel ratio is 0.906, which is smaller than the 1% threshold of 0.930 in Table 2. Therefore, when aiming for a value accuracy error of 1% or less, it is better to perform correction. Further, from (a) to (f), a decrease in variation due to an increase in the number of positive subdivisions, that is, a decrease in the confidence interval range is observed, and in particular, from (e) where the number of positive subdivisions exceeds 1600 (e). In f), the effect of the correction can be confirmed. Moreover, there is no significant difference between the normal distribution correction and the arbitrary distribution correction.

(3) When 50,000 target nucleic acids are randomly distributed, the results are shown in FIG. The notation rules are the same as in FIG.

In this example, since the effect of correction of 10% or more can be expected, it is better to perform volume dispersion correction. Further, in (A) normal distribution correction, the average value shows a value slightly larger than the true value, but in (B) arbitrary distribution correction, the result is closer to the true value. In this example, the volume dispersion has a uniform distribution, and when the target nucleic acid concentration is high, it means that the deviation from the Gaussian distribution cannot be ignored. Further, in (B) arbitrary distribution correction, it can be seen that the convergence occurs when the process is repeated three times.

From the above, it is possible to determine whether to perform the processing in the normal mode or the volume dispersion correction processing based on the negative subdivision ratio, the number of positive subdivisions, or both. Of course, the present invention is not limited to this, and a mode changeover switch may be provided so that the operator can measure the mode. There may be at least two modes in which the changeover switch can be selected: a normal mode, an automatic changeover mode, a normal distribution correction mode, and an arbitrary distribution correction mode. The target concentration measuring device according to the present embodiment is configured so that these modes can be selected.

<Comparison between normal distribution correction mode and arbitrary distribution correction mode>
If the volume information of all the subsections is known, the arbitrary distribution correction mode gives a result closer to the true value. Therefore, it is better to use the arbitrary distribution correction mode when the volume of all the subsections is measured together with the fluorescence measurement.

However, the arbitrary distribution correction mode requires volume information for all subsections. For example, when the volume information of the positive subsection is obtained from the light intensity distribution at the time of fluorescence measurement, when there is only a part of the volume information, the arbitrary distribution correction mode cannot be applied, and it is appropriate to use the normal distribution correction mode. is there. Inferring the volume information (mean value and standard deviation) of a sample from the volume information (mean value and standard deviation) of a part of the set in the sample may use a statistically valid method.

In addition, if the average volume and volume fluctuation amount (standard deviation / average value) of the measurement target are known empirically, correction by the normal distribution correction mode may be performed based on the values.

[First Example]
Hereinafter, the target concentration measuring apparatus according to the first embodiment of the present invention will be described with reference to the drawings.

<Measuring device>
FIG. 7 is a conceptual diagram of the target concentration measuring device according to this embodiment. In the figure, reference numeral 101 denotes an inspection plate to be measured, the details of which will be described later. 103 is an excitation light source composed of an LED, 105 is a collimating lens, 107 is a wavelength filter corresponding to the absorption spectrum of the fluorescent probe included in the inspection plate 101, and 109 is a dichroic mirror. 111 is an objective lens, 113 is a wavelength filter that selectively transmits the wavelength of fluorescence emitted from the fluorescent probe, 115 is an imaging lens, and 117 is a CCD that is an image sensor.

The wavelength characteristics of the dichroic mirror 109 are selected so as to reflect the excitation light and transmit the fluorescence.

121 is an image acquisition system, and 123 is an image processing system. Reference numeral 125 denotes a measurement mode selection switch.

<Inspection plate>
The structure of the inspection plate 101 will be described with reference to FIG. FIG. 3 is a diagram showing an example of the structure of the inspection plate 101 that holds the small section according to the present embodiment. FIG. 8A is a cross-sectional view of the inspection plate 101. FIG. 8B is a plan view of the inspection plate 101. The test plate 101 is an instrument that holds a solution of a mixture of a target nucleic acid and a reagent (a drug containing a chemical substance necessary for performing a PCR method such as a primer and a fluorescent probe) as a plurality of particles 222 in oil 221. .. The particles 222 correspond to subsections.

The inspection plate 101 has a structure in which a spacer 228 is sandwiched between an upper plate 226 and a lower plate 227, and oil 221 is filled between the two plates. The solution 220 is pressurized by a syringe or the like and injected into the solution injection unit 223, and a large number of particles 222 are generated by the porous membrane provided in the particle generation unit 224. In the present embodiment, the porous membrane is used for particle generation, but a microchannel chip or the like may be used. Since the particles 222 have a higher specific gravity than the oil 221, the particles 222 are arranged on the lower plate 227 so that the particles do not overlap with each other.

The target nucleic acids present in the plurality of particles 222 in the test plate 101 are amplified by a commonly known PCR method.

<Imaging>
The image acquisition system 121 has a function of controlling ON / OFF of the excitation light source 103 and imaging by the CCD 117 to acquire an image and transfer the acquired image to the image processing system 123.

Imaging of a fluorescence image will be described with reference to FIG. 7.

First, the inspection plate 101 is installed at a desired position. This position has an imaging relationship with CCD117.

Next, the excitation light source 103 is turned on. The emitted excitation light is converted into substantially parallel light by the collimating lens 105 and reaches the wavelength filter 107. The wavelength filter 107 has a property of transmitting only a desired wavelength band according to the absorption spectrum of the fluorescent probe. The excitation light transmitted through the wavelength filter 107 is reflected by the dichroic mirror 109 and is irradiated to the inspection plate 101 via the objective lens 111.

The fluorescent probe contained in the particles 222 in the inspection plate 101 absorbs the excitation light and emits fluorescence having a longer wavelength than that. The fluorescence emitted by the plurality of particles 222 is converted into substantially parallel light by the objective lens 111, passes through the dichroic mirror 109, and reaches the wavelength filter 113. The wavelength filter 113 has a property of transmitting only a desired wavelength band according to the fluorescence spectrum of the fluorescent probe. The fluorescence transmitted through the wavelength filter 13 is imaged on the CCD 117 by the imaging lens 115.

An example of a fluorescence image is shown in FIG. 8 (b). In the particles containing the target nucleic acid, the fluorescent probe is activated by binding to the target nucleic acid and exhibits bright fluorescence. Also, in particles that do not contain the target nucleic acid, weak fluorescence from unactivated fluorescent probes is observed.

Further, although not shown in the figure, the inspection plate 101 may be arranged on the XY movable stage. In this case, the number of particles used for analysis can be increased by acquiring a plurality of images by changing the location of the inspection plate 101 by the XY movable stage.

<Image processing>
The image processing system 123 discriminates negative particles, estimates particle volume, discriminates negative particles, estimates the number of negative particles in consideration of volume dispersion, and targets nucleic acid concentration based on the image data transferred from the image acquisition system 121. It has a function such as output of. Hereinafter, the flow of image processing (density estimation method) will be described with reference to FIG. The "subsection" in the figure means the "particle" of this embodiment.

(A) Measurement mode setting The measurement mode is set using the measurement mode selection switch 125. The measurement mode is selected from "normal ode", "automatic selection mode", and "volume dispersion correction mode".

The volume dispersion correction mode can be selected when the total number of subsections including the target of interest in the entire measurement target region is equal to or greater than the reference value. Further, the volume dispersion mode can be selected when the ratio of the total number of the plurality of subsections to the number of subsections not including the target of interest in the entire measurement target region is equal to or less than the reference value.

(B) Acquisition of fluorescence image As described above.

(C) Determination of negative ratio (negative subdivision ratio) (acquisition process)
First, the fluorescence intensity of each particle is extracted from the fluorescence image as shown in FIG. 8B. Then, it is determined that the particles having an intensity equal to or higher than a certain threshold value are positive particles, and the particles having an intensity lower than the threshold value are negative particles.

An example of measuring the fluorescence intensity is shown in FIG. This is a histogram showing the distribution of fluorescence intensity, the horizontal axis shows the fluorescence intensity, and the vertical axis shows the frequency. If a unit of fluorescence intensity is given, it is a digit of image luminance information, but its absolute value has no meaning. In this example, particles having 50 digits or more may be determined to be positive, and particles having less than 50 digits may be determined to be negative.

The total number of particles is N, the number of positive particles is Np, and the number of negative particles is Nq, which are stored in the memory. Further, the negative ratio obtained by dividing the number of negative particles by the total number of particles N is stored in the memory as q.

It is not always necessary to store the fluorescence intensity of each particle, but it may be stored for verification of the data.

(D) Branch when the normal mode is selected. (Decision process)
_{(E) Obtaining the average volume v 0} of all particles If the average volume of particles is empirically grasped in advance and the value is stored in the memory, the value is referred to. Alternatively, the average volume of the positive particles may be estimated from the fluorescence intensity of the positive particles, and the value may be equal to the overall average volume. Further, as shown in the second embodiment, a contour image may be acquired and the average volume may be obtained from the image.

(F) Output of target nucleic acid concentration (step of estimating the concentration of target substance)
-Log (q) / v ₀ is output as the concentration of the target nucleic acid. There are various methods for outputting the output, such as displaying it on a display device such as a display, printing it with a printer, or storing it in a memory.

(G) When the volume dispersion correction mode is selected, branching occurs. (Decision process)
(H) Obtaining the average volume v ₀ and the volume fluctuation amount (cv) of all particles In advance, the average volume and the volume fluctuation amount (cv) of the particles are empirically grasped, and the values are stored in the memory. If so, refer to that value. Alternatively, the average volume and volume fluctuation amount of the positive particles may be estimated from the fluorescence intensity of the positive particles, and the values may be equal to the whole. Further, as shown in Example 2, a contour image may be acquired and the average volume and the amount of volume fluctuation may be obtained from the image.

(I) Output of target nucleic acid concentration The concentration of the target nucleic acid corrected for volume dispersion according to Equation 13 is output. There are various methods for outputting the output, such as displaying it on a display device such as a display, printing it with a printer, or storing it in a memory.

(J) Branching based on the number of positive particles This branching is reached when the automatic selection mode is selected.

When the number of positive particles is 1600 or more, the measurement result can be expected to have a confidence interval of 95.4% and an error of ± 5% or less. In that case, volume dispersion correction is performed in order to further improve the accuracy of the measured value. On the contrary, when the number of positive particles is less than 1600, the normal mode is set.

Here, the threshold value for whether or not to perform volume dispersion correction is set to 1600 positive particles, but it is not limited to this, and a case where an error of ± 10% or less is allowed in a confidence interval of 95.4%. The threshold value may be 400 positive particles. Further, the configuration may be such that the operator can set an arbitrary value.

According to this embodiment, the measurement time can be shortened by performing the volume dispersion correction when the effect of the volume dispersion correction is high.

[Second Example]
In the volume dispersion correction mode, the accuracy of the target nucleic acid concentration can be further improved by actually measuring the volume of each particle.

Hereinafter, the target concentration measuring apparatus according to the second embodiment of the present invention will be described with reference to the drawings.

FIG. 11 is a conceptual diagram of the target concentration measuring device according to the second embodiment of the present invention. The same numbers are added to the parts common to those in FIG. 7, and the description thereof will be omitted.

The <measuring device> and <inspection plate> are the same as those in the first embodiment, and the description thereof will be omitted.

<Imaging>
Unlike the first embodiment, in order to capture the contour image of the particles, an illumination light source 119 composed of LEDs for capturing a transmitted image of the inspection plate 101 is provided.

The imaging of the contour image will be described.

The emission light emitted by turning off the excitation light source 103 and turning on the illumination light source 119 is applied to the inspection plate 101. The light transmitted through the inspection plate 101 is imaged on the CCD 117 by the objective lens 111 and the imaging lens 115. The dichroic mirror 109 and the wavelength filter 113 are left inserted so as not to cause a misalignment between the fluorescence image and the contour image.

An example of the contour image is shown in FIG. Both the particles containing the target nucleic acid and the particles not containing the target nucleic acid are imaged with dark edges.

Here, the lighting method is changed between the fluorescence image and the contour image. This is to avoid an error due to dark edges when trying to extract contours using a fluorescent image.

<Image processing>
The image processing system 123 discriminates negative particles, estimates particle volume, discriminates negative particles, estimates the number of negative particles in consideration of volume dispersion, and targets nucleic acid concentration based on the image data transferred from the image acquisition system 121. It has a function such as output of. Hereinafter, the flow of image processing will be described with reference to FIG. The volume dispersion correction according to this embodiment is referred to as "arbitrary distribution correction".

(A) Acquisition of Fluorescent Image The same as in the first embodiment.

(B) Determination of Negative Ratio (Negative Subsection Ratio) First, the fluorescence intensity of each particle is extracted from the fluorescence image as shown in FIG. 8 (b) while numbering each particle. Then, it is determined that the particles exhibiting an intensity equal to or higher than a certain threshold value are positive particles, and the particles exhibiting an intensity lower than the threshold value are positive particles. The determination result is stored in the memory as, for example, g (n), where n is an integer from 1 to N, where N is the assumed number of particles. Here, g (n) = 0 and g (n) = 1 with respect to the positive particles.

The total number of particles is N, the number of positive particles is N _p1 , and the number of negative particles is N _q1 . Further, the negative ratio obtained by dividing the number of negative particles by the total number of particles N is stored in the memory as _{q 1.}

It is not always necessary to store the fluorescence intensity of each particle, but it may be stored for verification of the data.

(C) Acquisition of particle volume An outline image is acquired by the above method. In order to reduce the influence of time-dependent changes such as the movement of particles with time, the acquisition of the contour image may be performed continuously with the acquisition of the fluorescence image. From the obtained contour image as shown in FIG. 12, the particles are numbered according to the fluorescence data, and the diameter of each particle is extracted. This uses commonly known methods such as binarization and template matching using a circular pattern as a template. Further, the value of the diameter is converted into the value of the sphere, and the particle number and the volume thereof are associated with each other and stored in the memory of the image acquisition system 121. For example, v (n), where n is an integer from 1 to N, is stored in memory.

An example of measuring the volume of particles is shown in FIG. This is a histogram showing the volume distribution, the horizontal axis shows the volume of particles, and the vertical axis shows the number of particles. The average volume was 113 pl. This average volume is stored in memory as _{v 0.}

(D) Determination of correction coefficient According to Equation 25, the volume dispersion correction coefficient w _m-1 (n) of each particle is stored in the memory. Here, m is an integer of 2 or more.

Since the volume dispersion correction coefficient of the negative particles is required for the calculation, renumber the particle numbers 1 to N _q1 as negative particles and the particle numbers N _q1 +1 to N as positive particles, and renumber them from 1 to N _q1 . The volume dispersion correction coefficient of N _{q1 may be stored.}

(D) Determination of correction coefficient _{The sum of the volume dispersion correction coefficients w m-1} (n) of the negative particles is stored in the memory as the _{number of mth negative particles N qm.}

Without renumbering, the number of mth negative particles N _qm can also be calculated as the product of the row vector g (n) and the column vector w _{m-1 (n).}

(F) Determination of the mth negative ratio The negative ratio (the mth negative ratio) obtained by dividing the number of _{mth negative particles N qm} by the total number of particles is calculated and stored in the memory as _{q m.}

(G) Judgment of convergence _{Compare the number of negative particles N qm} of the mth and the number of negative particles N _{q (m-1) of the m-1} to check whether this is smaller than a certain set threshold value. .. If no, the process from (d) is repeated with m = m + 1. This threshold value is, for example, 0.5.

(F) Output of target nucleic acid concentration −log (q _m ) / v ₀ is output as the target nucleic acid concentration. There are various methods for outputting the output, such as displaying it on a display device such as a display, printing it with a printer, or storing it in a memory.

According to this embodiment, the volume of each particle can be actually measured, and the volume dispersion correction can be performed based on the value, so that the accuracy of the target nucleic acid concentration can be improved.

[Third Example]
In the automatic selection mode, when the correction amount is large, the accuracy of the target nucleic acid concentration can be improved more effectively by configuring the mode to shift to the volume dispersion correction mode.

Hereinafter, the target concentration measuring apparatus according to the third embodiment of the present invention will be described with reference to the drawings. The device configuration is the same as in FIG. 7, and the description thereof will be omitted.

The <measuring device>, <inspection plate>, and <imaging> are the same as those in the first embodiment, and the description thereof will be omitted.

<Image processing>
The image processing system 123 discriminates negative particles, estimates particle volume, discriminates negative particles, estimates the number of negative particles in consideration of volume dispersion, and targets nucleic acid concentration based on the image data transferred from the image acquisition system 121. It has a function such as output of. Hereinafter, the flow of image processing will be described with reference to FIG. The "subsection" in the figure means the "particle" of this embodiment. It is assumed that the "automatic selection mode" is selected as the measurement mode.

(A) Acquisition of Fluorescent Image The same as in the first embodiment.

(B) Setting the initial value of the volume fluctuation amount In order to estimate the correction amount, it is necessary to grasp the volume fluctuation amount. For example, if the volume fluctuation amount of the measurement target can be grasped empirically, the value is stored in the memory as _{cv 0.} Alternatively, the configuration may be such that the operator can input an arbitrary numerical value.

(C) Determining the negative ratio (negative subdivision ratio) The same as in the first embodiment.

(D) Branching by correction amount When the correction amount exceeds a certain threshold value, the mode shifts to the volume dispersion correction mode. Assuming that the threshold value of the correction amount is a ₀ , the following conditional expression can be obtained with reference to the equation 32.

Equation 34 is expanded and organized as follows.

For example, when cv ₀ = 0.52 and the correction amount threshold value a ₀ = 1%, the negative ratio q ≦ 0.930 is obtained from Table 1. That is, when the negative ratio is equal to or less than a certain reference value, it can be determined to shift to the volume dispersion correction mode.

Regarding the judgment of branching, if the negative ratio is below a certain reference value, it may be set to shift to the volume dispersion correction mode, and if the correction amount exceeds a certain threshold value, it may be set to the volume dispersion correction mode. It may be set to migrate. Further, these threshold values may be configured to be arbitrarily set by the operator.

_{(E) Obtaining the average volume v 0} of all particles If the average volume of particles is empirically grasped in advance and the value is stored in the memory, the value is referred to. Alternatively, the average volume of the positive particles may be estimated from the fluorescence intensity of the positive particles, and the value may be equal to the overall average volume. Further, as shown in the second embodiment, a contour image may be acquired and the average volume may be obtained from the image.

(F) Output of target nucleic acid concentration −log (q) / v ₀ is output as the target nucleic acid concentration. There are various methods for outputting the output, such as displaying it on a display device such as a display, printing it with a printer, or storing it in a memory.

_{(G) Obtaining the average volume v 0} and volume fluctuation amount (cv) of all particles The average volume and volume fluctuation amount (cv) of all particles are empirically grasped in advance, and the values are stored in the memory. If so, refer to that value. Alternatively, the average volume and volume fluctuation amount of the positive particles may be estimated from the fluorescence intensity of the positive particles, and the values may be equal to the whole. Further, as shown in Example 2, a contour image may be acquired and the average volume and the amount of volume fluctuation may be obtained from the image.

(H) Output of target nucleic acid concentration The concentration of the target nucleic acid corrected for volume dispersion according to Equation 13 is output. There are various methods for outputting the output, such as displaying it on a display device such as a display, printing it with a printer, or storing it in a memory.

In this embodiment, as described in the first embodiment, the determination of branching based on the number of positive particles may be used in combination.

According to this embodiment, the measurement time can be shortened by performing the volume dispersion correction when the effect of the volume dispersion correction is high.

101 Inspection plate 103 Excitation light source 105 Collimating lens 107 Wavelength filter 109 Dichroic mirror 111 Objective lens 113 Wavelength filter 115 Imaging lens 117 CCD
119 Illumination light source 121 Image acquisition system 123 Image processing system 125 Measurement mode selection switch 221 Oil 222 Particles 223 Solution injection part 224 Particle generation part 226 Upper plate 227 Lower plate 228 Spacer

Claims (27)

An acquisition step of acquiring the number N of a plurality of subsections formed by dividing a sample containing a target substance and the number Nq of the number of subsections not containing the target substance.
An estimation method including an estimation process for estimating the concentration of the target substance contained in the sample from the N and the Nq.
In a normal mode in which the estimation is performed without correcting the volume variance of the plurality of subsections,
A method for estimating the concentration of a target substance, which comprises a determination step of determining which of the volume dispersion correction modes in which the estimation is performed by correcting the volume dispersion of the plurality of subsections. The determination process
The concentration estimation method according to claim 1, further comprising a step of automatically determining whether the estimation is performed in the normal mode or the volume dispersion correction mode. The determination step is
The concentration estimation method according to claim 2, wherein when the number of subsections containing the target substance is equal to or greater than a predetermined value, it is determined that the estimation is performed in the volume dispersion correction mode. The concentration estimation method according to claim 3, wherein the predetermined value can be arbitrarily set. The determination step is
2. The concentration estimation method according to any one of 4 to 4. A step of determining whether the estimation is performed in the normal mode or the volume dispersion correction mode based on the information about the correction coefficient used for correcting the volume dispersion of the plurality of subsections. The concentration estimation method according to any one of claims 2 to 5, wherein the concentration estimation method comprises. The volume dispersion correction mode includes a normal distribution correction mode in which the volume probability distribution of the plurality of subsections is regarded as a normal distribution and the estimation is performed, and a subsection that does not include the target substance based on information on the correction coefficient. The concentration estimation method according to claim 6, wherein the estimation is performed by any of the arbitrary distribution correction modes in which the estimation is performed using the number of. The concentration according to claim 1, wherein the determination step includes a step of determining whether the estimation is performed in the normal mode or the volume dispersion correction mode based on an instruction received from the user. Estimating method. An acquisition means for acquiring the number N of a plurality of subdivisions formed by dividing a sample containing a target substance and the number Nq of the subdivisions not containing the target substance.
An estimation device including an estimation means for estimating the concentration of the target substance contained in the sample from the N and the Nq.
In a normal mode in which the estimation is performed without correcting the volume variance of the plurality of subsections,
A device for estimating the concentration of a target substance, which comprises a determining means for determining which of the volume dispersion correction modes in which the estimation is performed by correcting the volume dispersion of the plurality of subsections. The concentration estimation device according to claim 9, wherein the determination means automatically determines whether the estimation is performed in the normal mode or the volume dispersion correction mode. The concentration estimation device according to claim 10, wherein the determination means determines that the estimation is performed in the volume dispersion correction mode when the number of subsections containing the target substance is equal to or greater than a predetermined value. .. The concentration estimation device according to claim 11, wherein the predetermined value can be arbitrarily set. The determining means determines that the estimation is performed in the volume dispersion correction mode when the ratio of the number of the sub-compartments to the number of sub-compartments containing no target substance is equal to or less than a reference value. The concentration estimation device according to any one of claims 9 to 12. The determining means determines whether the estimation is performed in the normal mode or the volume dispersion correction mode, based on the information about the correction coefficients used to correct the volume variance of the plurality of compartments. The concentration estimation device according to any one of claims 9 to 13, wherein the concentration estimation device is characterized. The volume dispersion correction mode includes a normal distribution correction mode in which the volume probability distribution of the plurality of subsections is regarded as a normal distribution and the estimation is performed, and a subsection that does not include the target substance based on information on the correction coefficient. The concentration estimation device according to claim 14, wherein the estimation is performed by any of the arbitrary distribution correction modes in which the estimation is performed using the number of. The concentration estimation device according to claim 9, wherein the determination means determines whether the estimation is performed in the normal mode or the volume dispersion correction mode based on an instruction received from the user. In a target concentration measuring device that estimates the target concentration from the ratio of the total number of the plurality of compartments to the number of compartments not containing the target of interest among the plurality of compartments obtained by dividing the sample containing the target of interest.
A normal mode in which the volume variance of the plurality of subsections is not corrected and
A device for measuring a target concentration, which comprises two measurement modes of a volume dispersion correction mode for correcting the volume dispersion of the plurality of subsections. The target concentration measuring apparatus according to claim 17, further comprising an automatic selection mode for automatically switching between the normal mode and the volume dispersion correction mode. It has a mode changeover switch that can be switched by the operator.
The target concentration measuring device according to claim 17 or 18, wherein the mode changeover switch can select at least two of the normal mode, the volume dispersion correction mode, and the automatic selection mode. The target concentration measuring apparatus according to claim 18, wherein a volume dispersion correction mode is selected when the total number of subsections including the target of interest is equal to or greater than a reference value. The device for measuring a target concentration according to claim 20, wherein the reference value is 400 or 1600. The device for measuring a target concentration according to claim 20, wherein the reference value can be arbitrarily set. The target concentration measuring apparatus according to claim 18, wherein the volume dispersion correction mode is selected when the ratio of the total number of the plurality of compartments to the number of compartments not including the target of interest is equal to or less than a reference value. .. The target concentration measuring device according to claim 23, wherein the reference value is an amount determined by a volume fluctuation amount (standard deviation / average value) of the plurality of subsections. The device for measuring a target concentration according to claim 23, wherein the reference value can be arbitrarily set. The target concentration measuring device according to claim 18, wherein the volume dispersion correction mode is selected when the correction amount is equal to or larger than a certain reference value. The target concentration measuring device according to claim 17, wherein the volume dispersion correction mode has a normal distribution correction mode and an arbitrary distribution correction mode.

Images