Intelligent protein detection system

Abstract

The invention provides an intelligent protein detection system, and belongs to the technical field of protein detection. An intelligent protein detection system comprises a color picking device for providing a photographing environment for a coomassie brilliant blue reaction system of a to-be-detected sample under fixed light intensity, and a processor for extracting a red light value, a green light value and a blue light value from a picture of the coomassie brilliant blue reaction system of the to-be-detected sample shot in the color picking device, and calculating and outputting the concentration of the sample to be detected by using the protein concentration prediction model. According to the intelligent protein detection system disclosed by the invention, shooting is carried out in the color picking device through intelligent equipment, a red light value, a green light value and a blue light value in a picture are automatically extracted, and a protein prediction result is automatically read. The intelligent protein detection system is rapid and efficient in field detection, gets rid of detection dependence of large-scale equipment, and is suitable for large-scale popularization and reference in the market.

Description

Technical Field

[0001] The present invention belongs to the technical field of protein detection, and in particular relates to an intelligent protein detection system. Background Art

[0002] The pharmaceutical industry is experiencing a significant increase in demand for rapid testing of drugs such as polysaccharides and proteins. For example, Coomassie Brilliant Blue staining, a commonly used method for protein detection, relies on a spectrophotometer for quantitative detection. This method relies on large instruments, is complex to operate, and struggles to meet the demands of on-site, point-of-care testing. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide an intelligent protein detection system, which takes a picture of the Coomassie Brilliant Blue staining result of the protein in a specific device, intelligently extracts the red light value (R value), green light value (G value) and blue light value (B value) through the picture, and then automatically reads the protein detection results based on the protein concentration prediction model.

[0004] The present invention provides a method for constructing a protein concentration prediction model, comprising the following steps:

[0005] The protein solution to be tested with gradient concentrations is mixed with Coomassie brilliant blue dye solution to react, thereby obtaining a series of concentrations of Coomassie brilliant blue reaction system;

[0006] Detecting the absorbance values ​​of the Coomassie Brilliant Blue reaction system at the series of concentrations;

[0007] The Coomassie Brilliant Blue reaction system with the series of concentrations is placed under a fixed light intensity and photographed to obtain photographs of the Coomassie Brilliant Blue reaction system with the series of concentrations;

[0008] Extract red light value, green light value and blue light value from the photos of the Coomassie brilliant blue reaction system at each concentration series;

[0009] A nonlinear model is constructed based on the relationship between the red light value, the green light value, the blue light value and the protein concentration to obtain a protein concentration prediction model.

[0010] Preferably, the method of extracting red light values, green light values ​​and blue light values ​​from the photographs of the Coomassie Brilliant Blue reaction system of each series of concentrations is to extract the RBG values ​​in the photographs using an image processing tool.

[0011] Preferably, the protein to be tested includes BSA.

[0012] Preferably, the fixed light intensity is provided by placing a cuvette containing the Coomassie Brilliant Blue reaction system of the series of concentrations between two light sources under conditions where no light escapes;

[0013] Both light sources are 5500K LED strip lights;

[0014] The cuvette is a transparent plastic cuvette.

[0015] Preferably, the protein concentration prediction model is shown in Formula IV:

[0016] Protein concentration = 4.7980 × R + 8.3708 × G + 11.0819 × B + -0.0154 × R 2 +-0.0771×G 2 +-0.0366×B 2 +0.0546×R×G+-0.0587×R×B+0.0617×G×B+-2043.7660Formula IV;

[0017] Where R represents the red value; G represents the green value; B represents the blue value; r 2 is 0.9996.

[0018] The present invention provides a protein intelligent detection system, comprising a color picking device and an intelligent device;

[0019] The color extraction device is used to provide a photographic environment for the Coomassie Brilliant Blue reaction system of the sample to be tested under a fixed light intensity;

[0020] The smart device is used to extract red light values, green light values ​​and blue light values ​​from a picture of the Coomassie Brilliant Blue reaction system of the sample to be tested taken in the color extraction device, and calculate and output the concentration of the sample to be tested using the protein concentration prediction model constructed using the construction method described in the above technical solution.

[0021] Preferably, the color extraction device comprises two vertically distributed hollow chambers;

[0022] The upper chamber is provided with an openable upper cover; a symmetrical light source is fixed on the inner side wall of the upper chamber; a card slot is provided on the inner bottom surface of the upper chamber, and the card slot is used to fix the cuvette; a shooting window for smart device photography is opened on the front side wall of the upper chamber;

[0023] A smart device fixing element is provided inside the lower chamber; the smart device fixing element includes an elastic structure and a fixing part connected to one end of the elastic structure; the other end of the elastic structure is vertically fixed to the side wall of the lower chamber; the smart device fixing element is retracted into the lower chamber in the restored state and is pulled out to fix the smart device when in use.

[0024] Preferably, the material of the color picking device includes a light-shielding material;

[0025] The inner side of the upper chamber of the color extraction device also includes a layer of white paint.

[0026] Preferably, the external dimensions of the color picking device are 15cm×15cm×15cm;

[0027] The length, width and height specifications of the upper chamber in the color picking device are 12cm×12cm×10cm; the length, width and height specifications of the lower chamber in the color picking device are 12cm×12cm×4cm.

[0028] Preferably, the upper cover is opened in one of the following ways: slide rail push-pull connection, hinged connection and complete detachment;

[0029] The upper cover and the upper edge of the upper chamber are respectively provided with magnetic components with adapted positions.

[0030] The present invention provides a method for constructing a protein concentration prediction model, comprising the following steps: mixing gradient concentrations of a protein solution to be tested with a Coomassie Brilliant Blue dye solution to react, respectively, to obtain a series of concentrations of Coomassie Brilliant Blue reaction systems; detecting the absorbance values ​​of the series of concentrations of the Coomassie Brilliant Blue reaction systems; placing the series of concentrations of the Coomassie Brilliant Blue reaction systems under a fixed light intensity and photographing them, to obtain series of concentrations of Coomassie Brilliant Blue reaction system photos; extracting red light values, green light values, and blue light values ​​from the photos of the Coomassie Brilliant Blue reaction systems of each series of concentrations; constructing a nonlinear model based on the red light values, green light values, blue light values, and protein concentrations to obtain a protein concentration prediction model. In order to avoid the use of large instruments to detect protein concentration and meet the requirements of rapid and efficient on-site detection, the present invention has developed a protein concentration prediction model based on the Coomassie Brilliant Blue staining method, combined with image processing and software analysis, and specifically using the red light values, green light values, and blue light values ​​in the staining result image as quantitative parameters, analyzing the nonlinear relationship between them and the absorbance of different protein concentrations, and the constructed nonlinear model can accurately predict the concentration of protein, R 2 The protein concentration error obtained by the protein concentration prediction model is 0.9996; compared with the traditional Coomassie Brilliant Blue method, the protein concentration error value obtained by the protein concentration prediction model is within 3.29%. This shows that the present invention provides a new protein concentration detection method that not only overcomes the limitations of traditional large-scale instrumentation but also allows for rapid and accurate detection results.

[0031] The present invention provides a protein intelligent detection system, including a color picking device and an intelligent device; the color picking device is used to provide a photographic environment for the Coomassie Brilliant Blue reaction system of the sample to be tested under a fixed light intensity; the intelligent device is used to extract the red light value, green light value and blue light value from the image of the Coomassie Brilliant Blue reaction system of the sample to be tested taken in the color picking device, and calculate and output the concentration of the sample to be tested using the protein concentration prediction model constructed by the construction method described in the above technical solution. In order to ensure the stability of the photographic environment provided by the Coomassie Brilliant Blue reaction system of the sample to be tested, the present invention provides a color picking device, which can provide a set of standard photos with stable light intensity and avoid color difference in shooting, thereby ensuring stable and uniform light intensity, so that the extracted red light value, green light value and blue light value are not affected by the shooting environment, and ensuring the accuracy of the prediction results based on the protein concentration prediction model. The protein intelligent detection system is directly operated by the intelligent device, which is simple and convenient, and can quickly read the prediction results, greatly shortening the protein concentration detection steps, meeting the requirements of on-site detection, and is suitable for large-scale promotion and application in the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a flowchart for the use of the protein intelligent detection system;

[0033] Figure 2 This is a side cross-sectional view of the color picker; 1. Top cover; 2. Magnetic component; 3. LED light strip; 4. Card slot; 5. Pull-out mounting bracket; 6. Shooting window;

[0034] Figure 3 Graphs of standard curves predicted by linear models and nonlinear models;

[0035] Figure 4 Image of the Coomassie Brilliant Blue reaction system of the sample to be tested, taken by a smart device. DETAILED DESCRIPTION

[0036] The present invention provides a method for constructing a protein concentration prediction model, comprising the following steps:

[0037] The protein solution to be tested with gradient concentrations is mixed with Coomassie brilliant blue dye solution to react, thereby obtaining a series of concentrations of Coomassie brilliant blue reaction system;

[0038] Detecting the absorbance values ​​of the Coomassie Brilliant Blue reaction system at the series of concentrations;

[0039] The Coomassie Brilliant Blue reaction system with the series of concentrations is placed under a fixed light intensity and photographed to obtain photographs of the Coomassie Brilliant Blue reaction system with the series of concentrations;

[0040] Extract red light value, green light value and blue light value from the photos of the Coomassie brilliant blue reaction system at each concentration series;

[0041] A nonlinear model is constructed based on the relationship between the red light value, the green light value, the blue light value and the protein concentration to obtain a protein concentration prediction model.

[0042] The invention mixes the protein solution to be tested with the Coomassie brilliant blue dye solution with gradient concentrations to react respectively, so as to obtain the Coomassie brilliant blue reaction system with a series of concentrations.

[0043] The present invention has no particular limitation on the type of protein to be tested, and any type of protein to be tested that is of interest in the art can be used. In the examples of the present invention, BSA is used as an example to illustrate the specific method for constructing a protein concentration prediction model, but this does not limit the scope of protection of the present invention.

[0044] The present invention has no special restrictions on the specific concentration range and the number of gradient concentrations in the gradient concentration of the protein solution to be tested. The gradient concentration range and number of the standard curve of the protein solution to be tested, which are well known in the art, can be used. In an embodiment of the present invention, the gradient concentration range of the BSA solution is 10 to 100 μg / mL, distributed according to an arithmetic progression. The present invention also has no special restrictions on the mixing reaction. The processing method for determining protein concentration by Coomassie Brilliant Blue staining, which is well known in the art, can be used. For example, in an embodiment of the present invention, the volume ratio of the protein solution to be tested and the Coomassie Brilliant Blue dye solution is 1:5; the mass percentage of the Coomassie Brilliant Blue dye solution is 10%. The temperature of the mixing reaction is preferably room temperature, such as 18 to 25°C. The time of the mixing reaction is 3 to 4 minutes.

[0045] After the mixing reaction is completed, the present invention detects the absorbance values ​​of the Coomassie Brilliant Blue reaction system with the series of concentrations.

[0046] In the present invention, the absorbance value is preferably detected by a spectrophotometer. The detection wavelength of the absorbance value is preferably 595 nm.

[0047] After the mixing reaction is completed, the present invention further comprises placing the Coomassie Brilliant Blue reaction system with a series of concentrations under a fixed light intensity and photographing the system to obtain photographs of the Coomassie Brilliant Blue reaction system with a series of concentrations.

[0048] In the present invention, the fixed light intensity helps ensure uniform lighting for photographs taken at different reaction batches, different protein types, and at different times, conforming to standardization and avoiding the effects of large color differences in photographs due to unstable lighting. The fixed light intensity is preferably provided by placing a cuvette containing the series of Coomassie Brilliant Blue reaction systems between two light sources under closed conditions, where no light escapes; both light sources are preferably 5500K LED light strips; and the cuvette is preferably a transparent plastic cuvette. The photographs are preferably taken under closed conditions, with a consistent light source and a fixed shooting distance.

[0049] After obtaining the photos of the Coomassie Brilliant Blue reaction system with a series of concentrations, the present invention extracts the red light value, the green light value and the blue light value from the photos of the Coomassie Brilliant Blue reaction system with each series of concentrations.

[0050] In the present invention, the method of extracting red light values, green light values, and blue light values ​​from photographs of Coomassie Brilliant Blue reaction systems of various concentration series is preferably to use an image processing tool to extract red light values, green light values, and blue light values ​​in the photographs. The present invention has no particular restrictions on the image processing tool, and a tool well known in the art that can extract RBG values ​​in a picture can be used, such as Adobe Photoshop or GIMP. In order to improve the prediction accuracy, the red light value is the average value of the values ​​obtained from the red channel of the selected area in the photograph. The number of the selected areas is preferably 3 to 5, and can be 4. The green light value and the blue light value are both average values ​​of multiple selected areas.

[0051] The present invention uses red light value (R), green light value (G) and blue light value (B) to construct a prediction model, which is beneficial to improve the accuracy of predicting protein concentration compared with gray value. In the early optimization research of the present invention, the gray value is calculated according to formula I based on the red light value, green light value and blue light value. By gray value modeling, r 2 is 0.99.

[0052] Grayscale value = 0.299R + 0.587G + 0.114B Formula 1.

[0053] The R, G, and B values ​​represent the red, green, and blue light values, respectively, and the grayscale value is a simplification of the RGB value, retaining only the "brightness" information.

[0054] In addition, in the early optimization research of the present invention, considering that the Coomassie brilliant blue color development system is mainly based on blue light, a single channel B value was also tried to model the r 2 It is only 0.9, which is not as accurate as the prediction result obtained by direct modeling using R, G, and B values.

[0055] After obtaining the red light value, green light value and blue light value, the present invention constructs a nonlinear model based on the relationship between the red light value, green light value and blue light value and protein concentration to obtain a protein concentration prediction model.

[0056] The present invention has no particular limitations on the method for constructing the nonlinear model; any nonlinear model known in the art may be used. In an embodiment of the present invention, a nonlinear model was constructed using MATLAB software. Experiments have shown that nonlinear models have advantages over linear models in terms of detection accuracy. In an embodiment of the present invention, using BSA as the protein to be tested, the protein concentration prediction model constructed is preferably as shown in Formula IV:

[0057] Protein concentration = 4.7980 × R + 8.3708 × G + 11.0819 × B + -0.0154 × R 2 +-0.0771×G 2 +-0.0366×B 2 +0.0546×R×G+-0.0587×R×B+0.0617×G×B+-2043.7660Formula IV;

[0058] Where R represents the red value; G represents the green value; B represents the blue value, r 2 is 0.9996.

[0059] The examples of the present invention demonstrate that the prediction results of the nonlinear model are relatively accurate, and compared with the results detected by the spectrophotometer, the error rate is within 3.29%.

[0060] The present invention provides a protein intelligent detection system, comprising a color picking device and an intelligent device; the color picking device is used to provide a photographic environment for the Coomassie Brilliant Blue reaction system of a sample to be tested under a fixed light intensity;

[0061] The smart device is used to extract red light values, green light values ​​and blue light values ​​from a picture of the Coomassie Brilliant Blue reaction system of the sample to be tested taken in the color extraction device, and calculate and output the concentration of the sample to be tested using the protein concentration prediction model constructed using the construction method described in the above technical solution.

[0062] In the present invention, the color extraction device preferably comprises two vertically spaced hollow chambers; the upper chamber is provided with an openable upper cover. The upper cover can be opened in one of the following ways: a sliding push-pull connection, a hinged connection, or a fully detachable opening method. The upper edges of the upper cover and the upper chamber are preferably each provided with a positionally compatible magnetic component to seal the upper cover to the housing, preventing light from escaping and affecting color aberration in the captured image. Symmetrical light sources are fixed to the inner sidewalls of the upper chamber; a slot is provided in the center of the inner bottom surface of the upper chamber for securing a cuvette. The cuvette is preferably a transparent plastic cuvette. Compared to glass cuvettes, transparent plastic cuvettes reduce light refraction. A shooting window is provided on the front sidewall of the upper chamber for use with a smart device to take photos. The size of the shooting window can be determined based on the type of smart device, preferably to ensure that it does not block the lens and prevents light from escaping. The interior of the upper chamber preferably also includes a layer of white paint. The white paint is preferably titanium dioxide, a hydrophobic polyurethane material. The thickness of the white paint is preferably 0.05mm. The white paint can neutralize the light absorption problem of the black background, achieve uniform diffuse reflection of light, and has the advantages of being waterproof, stain-resistant, and easy to clean. The light source is preferably a 5500K LED light strip; the length of the LED light strip is preferably consistent with the width of the inner wall. The external length, width, and height of the color extraction device are preferably 15cm×15cm×15cm; the length, width, and height of the upper chamber of the color extraction device are preferably 12cm×12cm×10cm.

[0063] In the present invention, a smart device securing element is provided within the lower chamber. The smart device securing element comprises an elastic structure and a securing member connected to one end of the elastic structure. The other end of the elastic structure is vertically secured to the side wall of the lower chamber. When in its restored state, the smart device securing element retracts into the lower chamber and is pulled out to secure the smart device during use. The elastic structure includes a spring. The securing member comprises a pull-out securing bracket. The pull-out securing bracket is preferably 4 cm in height.

[0064] In the present invention, the material of the color picking device preferably includes a light-shielding material to prevent light loss caused by the light source. The light-shielding material is preferably a black resin material, specifically a carbon fiber reinforced polyamide composite material. The present invention has no special restrictions on the source of the carbon fiber reinforced polyamide composite material, and any carbon fiber reinforced polyamide composite material well known in the art can be used. The external length, width and height specifications of the color picking device are preferably 15cm×15cm×15cm. The length, width and height specifications of the upper chamber in the color picking device are preferably 12cm×12cm×10cm; the length, width and height specifications of the lower chamber in the color picking device are preferably 12cm×12cm×4cm.

[0065] In the present invention, the smart device preferably includes a smart phone or a smart tablet. The smart device is installed with a program for taking pictures, inputting the extracted red light value, green light value, and blue light value into a protein concentration prediction model to calculate the protein concentration, and outputting and displaying the protein concentration.

[0066] The protein intelligent detection system provided by the present invention is described in detail below with reference to the following examples, but they should not be construed as limiting the scope of protection of the present invention.

[0067] Example 1

[0068] A design method for a color picking device in a protein intelligent detection system

[0069] The cross-sectional structure diagram of the color picking device is shown in FIG. Figure 2 1. Top cover; 2. Magnetic component; 3. LED light strip; 4. Card slot; 5. Pull-out bracket; 6. Photographing window. The color extraction device provides a photographic environment for the cuvette containing the Coomassie Brilliant Blue reaction system, ensuring that the photographic environment for different test samples is identical. The design features and preparation method of the color extraction device are as follows:

[0070] (1) The color picking device is made of black resin material through 3D printing, which can effectively block external light interference, provide a stable background environment for smartphone photography, and ensure the consistency of shooting conditions. The color picking device has a cube shape and is equipped with an openable upper cover to facilitate the removal and placement of the colorimetric dish. The external dimensions of the color picking device are 15 cm long × 15 cm wide × 15 cm high; the internal dimensions are 12 cm long × 12 cm wide × 10 cm high.

[0071] (2) Lighting system design: A 5500K LED light strip is embedded horizontally in the middle of the inner wall of the color picking device. The length of the LED light strip is consistent with the width of the inner wall. This lighting system can achieve a uniform and soft lighting effect, avoiding deviations in sample color reading due to uneven lighting.

[0072] (3) Cuvette Fixing and Supporting Components: A slot compatible with the UV-visible spectrophotometer cuvette is designed at the center of the bottom of the colorimeter. The slot is 1 cm deep and 1.3 cm long and wide. The slot is used to accurately fix the cuvette position to prevent the cuvette position from drifting and affecting the experimental results. At the same time, when testing the Coomassie Brilliant Blue reaction, a disposable plastic cuvette with four colorless and translucent surfaces is used to effectively eliminate the adverse effects of strong glass light reflection.

[0073] (4) Internal coating: The walls, floor, and top cover of the color extraction device are sprayed with matte white paint (titanium dioxide, polyurethane hydrophobic material) with a thickness of 0.05mm. The paint helps to neutralize the light absorption problem of the black background and achieve uniform diffuse reflection of light. It is also waterproof, anti-fouling, and easy to clean.

[0074] (5) Sealing and placement design: The inner edge of the lid of the color picking device is equipped with a magnet, which contacts the magnet on the upper edge of the inner wall of the color picking device box, which can better keep the cavity sealed and effectively isolate the interference of ambient light.

[0075] (6) Mobile phone holder design: The color picking device is also equipped with a pull-out mobile phone holder at the bottom of the front box, which is used to fix the smartphone for taking photos and provide stable support for shooting. The pull-out mobile phone holder is connected to the elastic structure and can be freely pulled out and pushed in to facilitate the removal and placement of the smartphone. When not in use, the pull-out mobile phone holder is retracted by the elastic structure to return to its original position; when in use, the pull-out mobile phone holder can be pulled out and fixed to the smartphone. The fixed height of the pull-out mobile phone holder is 4 cm.

[0076] (7) Shooting window setting: a shooting window is set on the front of the color picking device to adapt to the position of the mobile phone camera, with a length and width of 4.0 cm respectively, to ensure that all types of smart phones can take pictures of the cuvette through the shooting window.

[0077] Example 2

[0078] Construction method of mathematical model M in protein intelligent detection system

[0079] (1) Coomassie Brilliant Blue protein detection method

[0080] S1: Drawing of standard curve

[0081] Weigh 100 mg of Coomassie Brilliant Blue G250, dissolve it in 50 mL of 95% ethanol, add 85 mL of phosphoric acid, and dilute to 1000 mL with distilled water to obtain Coomassie Brilliant Blue staining solution.

[0082] To create a standard curve: Accurately weigh 50 mg of bovine serum albumin (BSA) standard and add 0.9% sodium chloride solution to prepare a 0.1 mg / ml standard protein solution. Dispense 0 ml, 0.10 ml, 0.20 ml, 0.40 ml, 0.60 ml, 0.80 ml, and 1.00 ml of the BSA standard solution into seven test tubes, then add 0.9% sodium chloride solution to a total volume of 1.00 ml and mix thoroughly. The concentrations of these standard solutions are 0 μg / ml, 10 μg / ml, 20 μg / ml, 40 μg / ml, 60 μg / ml, 80 μg / ml, and 100 μg / ml, respectively. Set up eight replicates for each concentration. Add 5 ml of Coomassie Brilliant Blue staining solution to each test tube, shake well, and let stand for 3 minutes for color development. Measure the absorbance of the resulting solution at 595 nm and average the absorbance of the three replicates. A standard curve was established with the mass concentration of bovine serum albumin in the standard series of solutions as the horizontal axis and the absorbance as the vertical axis.

[0083] The remaining five samples of each concentration were placed in parallel in the color extraction device prepared in Example 1, and photos were taken using a smartphone (Huawei nova11) using the photo function in the protein concentration prediction applet. The red light value (R), green light value (G), and blue light value (B) of the obtained cuvette images of the reaction system with different concentrations were extracted. The test results are shown in Table 1.

[0084] Table 1 Absorbance and RGB values ​​of protein solutions with different concentrations

[0085]

[0086]

[0087] MATLAB software was used to model the average values ​​of R, G, and B in Table 1. The modeling methods included linear regression model and nonlinear regression model.

[0088] The linear regression model is shown in Formula II.

[0089] y=1+x1+x2+x3 Formula II

[0090] Among them, X1 represents the R value, X2 represents the G value, and X3 represents the B value.

[0091] Linear regression model see Figure 3 and Table 2.

[0092] Table 2 Linear model calculation results

[0093] Estimate SE tStat P-value intercept 6.8024 119.11 0.057109 0.95605 x1 -0.064183 0.41701 -0.15391 0.88202 x2 -1.1116 0.307 -3.6208 0.0085016 x3 1.3302 0.46855 2.8389 0.025085

[0094] Note: Estimated coefficient: Estimate : Estimated value of the regression coefficient; ·SE: Standard error of the estimated value; ·tStat: t=Estimate / SE, used to test whether the coefficient is 0; ·p-value: Two-tailed significance level corresponding to the t-test.

[0095] Number of observations: 11, degrees of freedom (the number of independent information reflecting model error): 7, root mean square error (an indicator that measures the average error between the predicted value and the actual value of the regression model, reflecting the accuracy of the model fitting): 6.24, r 2 (indicates the proportion of the dependent variable's variation that can be explained by the independent variable): 0.975, adjusted r 2 is 0.965, F statistic (constant model): 91.9, p value = 5.5e-06; linear regression r 2 Value: 0.9752.

[0096] The linear regression equation is shown in Formula III:

[0097] Protein concentration = -0.0642×R + -1.1116×G + 1.3302×B + 6.8024 (Formula III).

[0098] The nonlinear regression model formula is shown in Formula IV.

[0099] Y=1+x1+x2+x3+x4+x5+x6+x7+x8+x9 Formula IV.

[0100] Linear regression model see Figure 3 and Table 3.

[0101] Table 3 Nonlinear model calculation results

[0102] Estimate SE tStat pValue (Intercept -2043.8 15877 -0.12872 0.9185 x1 4.798 81.448 0.058909 0.96254 x2 8.3708 17.913 0.4673 0.7217 X3 11.082 105.92 0.10462 0.93364 X4 -0.0154 0.097503 -0.15794 0.90027 X5 -0.07709 0.025482 -3.0253 0.20324 X6 -0.036641 0.17075 -0.21459 0.86543 X7 0.054611 0.076682 0.71217 0.60603 X8 -0.058669 0.26282 -0.22323 0.86018 X9 0.061704 0.082004 0.75245 0.58934

[0103] Number of observations: 11, degrees of freedom for error: 1; root mean square error: 2.02; F statistic (constant model): 300, p-value = 0.0448; nonlinear regression r 2 Value: 0.9996.

[0104] The nonlinear regression equation is shown in Formula V:

[0105] Protein concentration = 4.7980 × R + 8.3708 × G + 11.0819 × B + -0.0154 × R 2 +-0.0771×G 2 +

[0106] -0.0366×B 2 +0.0546×R×G+-0.0587×R×B+0.0617×G×B+-2043.7660Formula V.

[0107] Comparison of the two models:

[0108] In the linear model, r 2 is 0.9752, and r 2 Compared with the linear model, the r 2 The improvement was 0.0244, indicating that the prediction results of the nonlinear model were more accurate. Therefore, the nonlinear regression equation was selected to calculate the protein concentration.

[0109] Example 3

[0110] A protein intelligent detection system and detection method

[0111] 1. Composition of a protein intelligent detection system

[0112] 1) The color extraction device prepared in Example 1 is used to provide a photographic environment for the Coomassie Brilliant Blue reaction system;

[0113] 2) A smartphone equipped with a protein concentration prediction program, used for taking photos of the Coomassie Brilliant Blue reaction system, analyzing the results, and outputting the results; the protein concentration prediction program is used to run the nonlinear regression equation obtained in Example 2 and output the predicted protein concentration to a display screen.

[0114] 3) Coomassie Brilliant Blue Dye.

[0115] 2. Detection Method

[0116] A. Preparation of Coomassie Brilliant Blue Staining Solution: Dissolve 0.1 g of Coomassie Brilliant Blue G50 in 50 mL of 95% ethanol, add 85 mL of phosphoric acid, and then add distilled water to 1000 mL.

[0117] B. Sample processing: Take 0.1mL of the BSA protein solution to be tested, add 5mL of Coomassie Brilliant Blue staining solution, mix well, and let it stand for 3 minutes. The reaction is completed in a disposable cuvette with light-transmitting surfaces. Place the cuvette of the Coomassie Brilliant Blue reaction system into the card slot of the color picking device and fix it. Use a smartphone to perform the test. The specific steps are as follows: open the protein concentration prediction program installed on the smartphone, adjust the position of the smartphone so that the cuvette appears in the middle of the picture (see Figure 4 ), click to take a photo, and then read the protein concentration.

[0118] Example 4

[0119] Validation of a protein concentration detection method based on a protein intelligent detection system

[0120] Take 1 mL of the appropriately diluted BSA protein solution to be tested in a test tube, add 5 mL of Coomassie Brilliant Blue staining solution, shake well, and let it stand for 3 minutes to develop color.

[0121] ① The obtained solution was placed in a cuvette and the absorbance was measured at a wavelength of 595 nm using a spectrophotometer. The absorbances were 0.194, 0.197, and 0.196, respectively. The standard curve constructed in Example 1 was used to obtain protein concentrations of 31.00 mg / ml, 31.59 mg / ml, and 31.40 mg / ml, and the average concentration was calculated to be 31.33 mg / ml.

[0122] ② Use a smartphone to take a photo of the above solution in a color picking device, use a protein concentration prediction program to extract the average R value, G value, and B value in the image, and read the corresponding protein concentration.

[0123] The R, G, and B values ​​of the BSA protein solution were obtained, as shown in Table 4. The program read an average protein concentration of 30.3 mg / ml. Compared with the Coomassie Brilliant Blue colorimetric method, the error in protein concentration prediction based on the mathematical model of the present invention was 3.29%.

[0124] Table 4 Validation group results

[0125] parallel R-value G value B value Protein concentration 1 125.82 187.91 183.77 31.44 2 122.79 187.85 184.23 31.17 3 123.63 188.06 182.86 28.29

[0126] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for constructing a protein concentration prediction model, characterized in that: The following steps are involved: The protein solution to be tested with gradient concentrations is mixed with Coomassie brilliant blue dye solution to react, thereby obtaining a series of concentrations of Coomassie brilliant blue reaction system; Detecting the absorbance values ​​of the Coomassie Brilliant Blue reaction system at the series of concentrations; The Coomassie Brilliant Blue reaction system with the series of concentrations is placed under a fixed light intensity and photographed to obtain photographs of the Coomassie Brilliant Blue reaction system with the series of concentrations; Extract red light value, green light value and blue light value from the photos of the Coomassie brilliant blue reaction system at each concentration series; A nonlinear model is constructed based on the relationship between the red light value, the green light value, the blue light value and the protein concentration to obtain a protein concentration prediction model.

2. The construction method according to claim 1, characterized in that: The method for extracting red light values, green light values, and blue light values ​​from photos of Coomassie Brilliant Blue reaction systems of various concentration series is to use an image processing tool to extract RBG values ​​in the photos.

3. The construction method according to claim 1, characterized in that: The protein to be tested includes BSA.

4. The construction method according to claim 1, characterized in that: The fixed light intensity is provided by placing a cuvette containing the series of concentrations of the Coomassie Brilliant Blue reaction system between two light sources under conditions of sealing and no light escape; Both light sources are 5500K LED strip lights; The cuvette is a transparent plastic cuvette.

5. The construction method according to any one of claims 1 to 4, characterized in that: The protein concentration prediction model is shown in Formula IV: Protein concentration = 4.7980 × R + 8.3708 × G + 11.0819 × B + -0.0154 × R 2 +-0.0771×G 2 +-0.0366×B 2 +0.0546×R×G+-0.0587×R×B+0.0617×G×B+-2043.7660Formula IV; Where R represents the red value; G represents the green value; B represents the blue value; r 2 is 0.9996.

6. A protein intelligent detection system, characterized in that: Including color picking device and smart device; The color extraction device is used to provide a photographic environment for the Coomassie Brilliant Blue reaction system of the sample to be tested under a fixed light intensity; The smart device is used to extract red light values, green light values ​​and blue light values ​​from a picture of a Coomassie brilliant blue reaction system of a sample to be tested taken in a color extraction device, and calculate and output the concentration of the sample to be tested using a protein concentration prediction model constructed using the construction method described in any one of claims 1 to 5.

7. The protein intelligent detection system according to claim 6, characterized in that: The color extraction device includes two vertically distributed hollow chambers; The upper chamber is provided with an openable upper cover; a symmetrical light source is fixed on the inner side wall of the upper chamber; a card slot is provided on the inner bottom surface of the upper chamber, and the card slot is used to fix the cuvette; a shooting window for smart device photography is opened on the front side wall of the upper chamber; A smart device fixing element is provided inside the lower chamber; the smart device fixing element includes an elastic structure and a fixing part connected to one end of the elastic structure; the other end of the elastic structure is vertically fixed to the side wall of the lower chamber; the smart device fixing element is retracted into the lower chamber in the restored state and is pulled out to fix the smart device when in use.

8. The protein intelligent detection system according to claim 7, characterized in that: The material of the color picking device includes light-shielding material; The inner side of the upper chamber of the color extraction device also includes a layer of white paint.

9. The protein intelligent detection system according to claim 7, characterized in that: The external length, width and height of the color picking device are 15cm×15cm×15cm; The length, width and height specifications of the upper chamber in the color picking device are 12cm×12cm×10cm; the length, width and height specifications of the lower chamber in the color picking device are 12cm×12cm×4cm.

10. The protein intelligent detection system according to claim 7, characterized in that: The opening method of the upper cover includes one of the following: slide rail push-pull connection, hinged connection and complete detachment; The upper cover and the upper edge of the upper chamber are respectively provided with magnetic components with adapted positions.

Images