Screening method for engineering algal strain chromosome multi-copy integration

Abstract

The invention provides a screening method for engineering algal strain chromosome multi-copy integration, and belongs to the technical field of bioinformatics and synthetic biologication.The screening method for engineering algal strain chromosome multi-copy integration comprises the steps that firstly, morphological characteristics are obtained through microscopic image analysis, growth parameters are measured, and growth characteristic data are constructed; the method comprises the following steps: analyzing photosynthesis characteristics to construct photosynthetic data, acquiring integration site data by adopting a fluorescence in-situ hybridization technology, measuring a gene transcription level, integrating the data by utilizing a deep neural network model for evaluation, calculating a robustness evaluation value, and constructing a screening judgment equation based on a transcription intensity variable coefficient. According to the method, the genetic stability is evaluated through continuous culture, and finally, the score is calculated by adopting the comprehensive scoring equation for screening, so that the precise screening of the engineering algal strains is realized, and the technical problem that the stability and the uniformity of the engineering algal strain chromosome multi-copy integration are difficult to judge and screen in the prior art is solved.

Description

Technical Field

[0001] This invention belongs to the fields of bioinformatics and synthetic biology, and specifically relates to a screening method for multicopy integration of chromosomes in engineered algal strains. Background Technology

[0002] In the field of biotechnology, algae are widely used as important bioreactors for the expression of genetically engineered proteins and the production of bioactive substances. Traditional algal genetic engineering mainly achieves the expression of exogenous genes through ribosomal DNA integration, chloroplast DNA integration, or plasmid vectors. These methods typically employ screening marker genes for initial screening, followed by verification of integration using molecular biology methods such as PCR and Southern hybridization. Finally, the engineered strain is evaluated by detecting target protein expression or measuring the content of target products. In practical applications, chromosome integration in engineered algal strains often exhibits characteristics of multi-copy, random integration. This integration mode directly affects the stability and uniformity of exogenous gene expression.

[0003] Currently, methods for evaluating chromosome integration in engineered algal strains mainly include: fluorescence in situ hybridization (FISH) for detecting integration sites, real-time quantitative PCR for detecting gene copy numbers, Western blotting for detecting protein expression, and metabolomics for analyzing metabolites. However, these methods are independent and lack a systematic integrated evaluation framework. Furthermore, the unique characteristics of algal cells, such as rigid cell walls, autofluorescence interference, and high cell density, limit the application of traditional evaluation methods. In addition, existing methods primarily focus on static indicators, making it difficult to reflect the dynamic changes of engineered algal strains during continuous culture and accurately predict their long-term stability.

[0004] Most critically, existing technologies lack a standardized screening method that can comprehensively assess the stability and uniformity of chromosome integration in engineered algal strains. This makes it difficult to quickly and accurately screen for engineered strains with good stability and uniform expression in practical applications. Especially in large-scale production, strain stability directly affects product quality and production efficiency, and existing technologies cannot effectively predict and screen strains that meet industrial requirements, which severely restricts the industrial application of engineered algal strains. In other words, existing technologies suffer from the technical problem of difficulty in determining and screening for the stability and uniformity of multi-copy chromosome integration in engineered algal strains. Summary of the Invention

[0005] In view of this, the present invention provides a screening method for multi-copy integration of chromosomes in engineered algal strains, which can solve the technical problem in the prior art that it is difficult to determine and screen the stability and uniformity of multi-copy integration of chromosomes in engineered algal strains.

[0006] This invention is implemented as follows: A screening method for multi-copy chromosome integration in engineered algal strains includes: constructing a data matrix of morphology, growth, photosynthesis, and integration sites; calculating robustness assessment values ​​using a deep neural network model; constructing a screening decision equation to calculate screening decision values; calculating transcriptional level fluctuation rates; inputting the robustness assessment values, screening decision values, and transcriptional level fluctuation rates into a comprehensive scoring equation to calculate the engineered strain score; and selecting the strain with the highest score based on the engineered strain score ranking. The deep neural network model incorporates a biological characteristic prediction module between the convolutional layer and the fully connected layer. This module includes growth kinetics prediction equations and photosynthesis prediction equations to improve screening accuracy.

[0007] Specifically, the step of constructing the morphological data matrix involves acquiring microscopic images of the target algal strain, measuring cell diameter, cell edge integrity, nucleus-to-cytoplasm ratio, and chloroplast distribution density, and then constructing a morphological feature data matrix.

[0008] Specifically, the step of constructing the growth characteristic data matrix involves measuring the specific growth rate, cell density, and biomass of the target algal strain during the exponential growth phase to construct the growth characteristic data matrix.

[0009] Specifically, the step of constructing the photosynthesis data matrix involves measuring the photosynthetic characteristic parameters of the target algae, including light intensity, chlorophyll content, photosystem II quantum yield, and non-photochemical quenching coefficient, to construct the photosynthesis data matrix.

[0010] Specifically, the step of constructing the integration site data matrix involves using fluorescence in situ hybridization (FISH) to determine the number and distribution of chromosome integration sites, and then constructing the integration site data matrix.

[0011] Specifically, the step of constructing the screening and determination equation involves using flow cytometry to determine the transcriptional level of the target gene, calculating the transcriptional intensity variation coefficient, constructing the screening and determination equation based on the transcriptional intensity variation coefficient, and calculating the screening and determination value. The threshold range of the screening and determination value is 0.1 to 0.3.

[0012] Specifically, the step of calculating the transcriptional level fluctuation rate involves continuously culturing for 5 generations, measuring the change in the transcriptional level of the target gene, and calculating the transcriptional level fluctuation rate to assess the genetic stability of the target chromosome integration.

[0013] The deep neural network model adopts an improved residual network structure, including an input layer, three convolutional layers, a biological characteristic prediction module, two fully connected layers, and an output layer.

[0014] The growth kinetics prediction equation in the biological characteristic prediction module includes specific growth rate, cell density, and biomass as input parameters, and the output parameter is the growth prediction value.

[0015] The input parameters of the photosynthesis prediction equation in the biological characteristic prediction module include light intensity, chlorophyll content, photosystem II quantum yield, and non-photochemical quenching coefficient, and the output parameter is the predicted photosynthetic efficiency.

[0016] The robustness assessment value is used to characterize the growth and metabolic stability of the engineered strain, and the screening judgment value is used to assess the uniformity of target chromosome integration.

[0017] In the comprehensive scoring equation, the robustness assessment has a weight of 0.4, the screening judgment has a weight of 0.4, and the transcription level fluctuation rate has a weight of 0.2. The deep neural network model has a training sample size of no less than 10,000 sets.

[0018] The method of the present invention is applicable to any one of the following algae that have been domesticated in the laboratory and for which ribosome entry site sequences, chloroplast targeting sequences, and endoplasmic reticulum localization sequences have been established: Chlamydomonas reinhardtii, Chlorella vulgaris, Dunaliella salina, Prorocentrum dinoflagellates, Alexandrium, Platychophyton floccosum, Scenedesmus stenoptera, Euglena, Rhodophyta rubrum, and Crescentella spp.

[0019] Compared with existing technologies, this invention provides a screening method for multi-copy chromosome integration in engineered algal strains. This method, based on multi-dimensional data analysis and deep learning, constructs multiple data matrices including morphological features, growth characteristics, photosynthetic properties, and integration sites, and combines them with an improved deep neural network model to achieve comprehensive evaluation and precise screening of engineered algal strains. By incorporating a biological characteristic prediction module, this method can accurately predict the growth trend and photosynthetic efficiency of algal strains, thereby assessing their long-term stability.

[0020] The method of this invention overcomes the limitations of traditional techniques. First, by integrating phenotypic data from multiple dimensions, a systematic evaluation framework is established, avoiding the one-sidedness of evaluation based on a single indicator. Second, by introducing a deep learning model, intelligent analysis of large amounts of data is achieved, improving the accuracy and efficiency of screening. Third, through continuous cultivation and dynamic monitoring, the genetic stability of engineered algal strains is assessed, solving the problem that traditional methods struggle to predict long-term stability. Finally, by establishing a comprehensive scoring equation, various indicators are quantified into a unified scoring standard, achieving standardization and objectivity in the screening process.

[0021] This invention successfully solves the technical problem of determining and screening the stability and uniformity of multi-copy chromosome integration in engineered algal strains. Its core lies in establishing a complete evaluation system that organically combines indicators from multiple levels, including morphological characteristics, physiological properties, and gene expression. Through intelligent analysis and prediction using a deep learning model, it ultimately achieves precise screening of engineered algal strains. This method can not only assess the current integration state but also predict long-term stability, providing a reliable screening basis for industrial applications. Attached Figure Description

[0022] Figure 1 This is a flowchart of the method of the present invention. Figure 2 This is a growth curve of the engineered algal strain during a 72-hour culture period. Figure 3 This is a graph showing the relationship between photosynthetic parameters and light intensity. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0024] like Figure 1 The diagram shown is a flowchart of a screening method for multi-copy integration of chromosomes in engineered algal strains provided by this invention. This method includes the following steps:

[0025] S01. Collect microscopic images of the target algal strain, measure cell diameter, cell edge integrity, nucleocytoplasmic ratio, and chloroplast distribution density, and construct a morphological feature data matrix.

[0026] S02. Measure the specific growth rate, cell density, and biomass of the target algal strain during the exponential growth phase, and construct a growth characteristic data matrix.

[0027] S03. Measure the photosynthetic characteristics parameters of the target algal strain, including light intensity, chlorophyll content, photosystem II quantum yield, and non-photochemical quenching coefficient, and construct a photosynthetic data matrix.

[0028] S04. Use fluorescence in situ hybridization to determine the number and distribution of chromosome integration sites, and construct an integration site data matrix.

[0029] S05. Use flow cytometry to determine the transcriptional level of the target gene and calculate the coefficient of variation of transcriptional intensity.

[0030] S06. Construct a deep neural network model, using the morphological feature data matrix, growth feature data matrix, photosynthesis data matrix, and integration site data matrix as input parameters, with a training sample size of no less than 10,000 sets.

[0031] S07. Input the morphological feature data matrix, the growth feature data matrix, the photosynthesis data matrix, and the integration site data matrix into the deep neural network model, and calculate the robustness assessment value;

[0032] S08. Construct a screening equation based on the transcription intensity variation coefficient value, and calculate the screening value. The threshold range of the screening value is 0.1 to 0.3.

[0033] S09. After culturing for 5 generations, the changes in the transcriptional level of the target gene were measured, and the transcriptional level fluctuation rate was calculated.

[0034] S10. Input the robustness assessment value, the screening judgment value, and the transcription level fluctuation rate value into the comprehensive scoring equation to calculate the score of the engineered strain.

[0035] S11. Sort the engineered strains according to their scores and select the strain with the highest score.

[0036] The deep neural network model adopts an improved residual network structure and sets a biological characteristic prediction module between the convolutional layer and the fully connected layer. The biological characteristic prediction module includes a growth kinetics prediction equation and a photosynthesis prediction equation.

[0037] The growth kinetics prediction equation has the following input parameters: the specific growth rate value, the cell density value, and the biomass value. The output parameter is the growth prediction value.

[0038] The input parameters of the photosynthesis prediction equation include the light intensity value, the chlorophyll content value, the photosystem II quantum yield value, and the non-photochemical quenching coefficient value, and the output parameter is the predicted photosynthetic efficiency value.

[0039] The deep neural network model includes an input layer, three convolutional layers, the biological characteristic prediction module, two fully connected layers, and an output layer.

[0040] The robustness assessment values ​​are used to characterize the growth stability and metabolic stability of the engineered strains;

[0041] The screening criteria are used to assess the uniformity of target chromosome integration;

[0042] The transcriptional level fluctuation rate is used to assess the genetic stability of target chromosome integration;

[0043] In the comprehensive scoring equation, the robustness assessment has a numerical weight of 0.4, the screening judgment has a numerical weight of 0.4, and the transcription level fluctuation rate has a numerical weight of 0.2.

[0044] This method is applicable to the following chromosomal integrated genes: hematopoietic cell differentiation antigen CD45 gene, T lymphocyte common antigen CD3 gene, helper T lymphocyte marker CD4 gene, cytotoxic T lymphocyte marker CD8 gene, natural killer cell marker CD16 gene, neural cell adhesion molecule CD56 gene, B lymphocyte marker CD19 gene, complement regulatory protein CD55 gene, complement regulatory protein CD59 gene, interleukin-2 receptor α chain CD25 gene, and white blood cell... Intercytokinin 7 receptor α chain CD127 gene, programmed death receptor 1 gene, T cell co-stimulatory molecule CD28 gene, immunoglobulin Fc receptor CD64 gene, complement receptor 3 gene, complement receptor 1 gene, perforin gene; This method is applicable to the following algae that have been domesticated in the laboratory and for which ribosome entry site sequences, chloroplast targeting sequences, and endoplasmic reticulum localization sequences have been established: Chlamydomonas reinhardtii, Chlorella vulgaris, Dunaliella salina, Prorocentrum dinoflagellates, Alexandrium, Platychophyton floccosum, Scenedesmus stenoptera, Euglena spp., Rhodophyta rubrum, and Microceratops spp.

[0045] The specific implementation methods of the above steps are described in detail below.

[0046] The specific implementation of step S01 involves using a microscopic image acquisition system to analyze the morphological characteristics of the target algal strain. First, a high-resolution image of the algal strain is acquired under a microscope, with an image resolution of no less than 1920×1080 pixels, an acquisition magnification of 400x, and a sampling quantity of no less than 100 cells. The contour information of a single cell is extracted using an image segmentation algorithm, and the cell diameter is calculated using the Hough circle transform algorithm. The measurement accuracy of the cell diameter should reach 0.1 micrometers. An edge detection algorithm is used to calculate the cell edge integrity value. The edge integrity is quantified by the continuity and smoothness of the edge pixels, with a value range of 0 to 1 and an edge integrity threshold of no less than 0.85. The cell nucleus and cytoplasm regions are segmented using a region growing algorithm, and the nucleus-cytoplasm ratio is calculated. The normal range of the nucleus-cytoplasm ratio is 0.15 to 0.25. Exceeding this range indicates abnormal cell structure. The spatial distribution characteristics of chloroplasts are analyzed using a density clustering algorithm, and the chloroplast distribution density value is calculated. The density value reflects the uniformity and aggregation degree of chloroplasts. Finally, these feature parameters are organized into a morphological feature data matrix, with the matrix dimension being the number of samples × the number of features.

[0047] The specific implementation of step S02 involves determining the growth characteristic parameters of the target algal strain under standard culture conditions. First, the algal strain is cultured under conditions of light intensity of 200 micromoles per square meter per second, temperature of 25 degrees Celsius, and pH of 7.0. The optical density value is sampled and measured every 4 hours. The specific growth rate is calculated by fitting the optical density versus time curve. The specific growth rate is calculated using an exponential growth equation. The culture conditions should be kept stable during the measurement. At the same time, the cell density value is measured using a hemocytometer. The cell density measurement is repeated 3 times and the average value is taken. The dry weight biomass value is measured by centrifugation at a temperature of 60 degrees Celsius for 24 hours. Finally, these growth parameters are organized into a growth characteristic data matrix.

[0048] The specific implementation of step S03 involves determining the photosynthetic characteristics of the target algal strain. First, a quantum optical sensor is used to measure the light intensity of the culture environment. The measurement location is the surface of the culture bottle, and three different locations are measured for each culture bottle, with the average value taken. A spectrophotometer is used to measure the chlorophyll content. Chlorophyll extraction is performed using 90% acetone, and absorbance is measured at wavelengths of 663 nm and 645 nm. The contents of chlorophyll a and chlorophyll b are calculated using the Arnnon formula. A pulse-modulated chlorophyll fluorescence spectrometer is used to measure the photosystem II quantum yield. During the measurement, dark adaptation is performed for 15 minutes, the saturation pulse light intensity is 3000 μmol / m² / s, and the pulse duration is 0.8 seconds. The non-photochemical quenching coefficient is also measured simultaneously. Finally, these photosynthetic characteristic parameters are organized into a photosynthetic data matrix.

[0049] The specific implementation of step S04 involves using fluorescence in situ hybridization (FISH) to analyze chromosome integration. First, chromosome slides are prepared. Cells are digested with trypsin, treated with hypotonicity, and fixed with methanol and acetic acid. When preparing probes, DNA fragments complementary to the target gene sequence are selected. Probes are labeled with biotin or digoxigenin. The hybridization reaction is carried out at 42 degrees Celsius for 16 hours. Detection is performed using streptavidin-fluorescein labeling or anti-digoxigenin antibody. Observation and photography are conducted under a fluorescence microscope. Image analysis software is used to count the number and distribution of chromosome integration sites, calculating the numerical values ​​of the number and distribution of chromosome integration sites. Finally, these integration characteristic parameters are organized into an integration site data matrix.

[0050] The specific implementation of step S05 involves using flow cytometry to analyze the transcriptional level of the target gene. First, algal cells in the logarithmic growth phase are collected, total RNA is extracted using an RNA extraction kit, cDNA is synthesized using a reverse transcription kit, and specific primers are designed to perform real-time quantitative PCR amplification of the target gene. The housekeeping gene is used as an internal control, and the relative expression level of the target gene is calculated using the negative ΔΔCt method of 2. Each sample is repeated three times, and the mean and standard deviation of the transcriptional level are calculated. The coefficient of variation of transcription intensity is calculated by dividing the standard deviation by the mean. The coefficient of variation reflects the stability of gene expression; the smaller the coefficient of variation, the more stable the expression.

[0051] The specific implementation of step S06 involves constructing a deep neural network model to evaluate engineered algal strains. First, an improved residual network structure is designed, adding a biological characteristic prediction module to the standard residual network. This module, located between the convolutional layers and fully connected layers, is used to extract biological features. The input layer receives morphological feature data matrices, growth feature data matrices, photosynthetic data matrices, and integration site data matrices. The first convolutional layer uses 64 3×3 convolutional kernels with a stride of 1, the second uses 128 3×3 convolutional kernels with a stride of 2, and the third uses 256 3×3 convolutional kernels with a stride of 2. Batch normalization and ReLU activation functions are applied after each convolutional layer. The biological characteristic prediction module includes growth kinetics prediction... The growth kinetics prediction equation uses a modified Monod equation, taking specific growth rate, cell density, and biomass as input parameters to predict the growth trend of the algal strain. The photosynthesis prediction equation uses a modified light response curve model, taking light intensity, chlorophyll content, photosystem II quantum yield, and non-photochemical quenching coefficient as input parameters to predict the photosynthetic efficiency of the algal strain. The number of neurons in the two fully connected layers are 512 and 256, respectively. Dropout is used to prevent overfitting, and the output layer uses the Softmax activation function. The number of training samples is no less than 10,000 sets. A stochastic gradient descent optimizer is used with a learning rate of 0.001, a batch size of 64, and 100 training rounds.

[0052] The specific implementation of step S07 involves using a trained deep neural network model to evaluate the robustness of the engineered algal strain. First, the morphological feature data matrix, growth feature data matrix, photosynthetic data matrix, and integration site data matrix are standardized. The data are transformed to the same scale using the Z-score standardization method. The standardized data is then input into the deep neural network model. The model extracts features through convolutional layers and calculates growth prediction values ​​and photosynthetic efficiency prediction values ​​through the biological characteristic prediction module. The consistency between these prediction values ​​and actual observation values ​​reflects the robustness of the engineered algal strain. All features are integrated through fully connected layers, and finally, a robustness evaluation value is output. The value ranges from 0 to 1, with a larger value indicating better robustness of the engineered algal strain. The recommended threshold is 0.8.

[0053] The specific implementation of step S08 is to construct a screening determination equation based on the transcription intensity variation coefficient. First, the baseline value of the transcription intensity variation coefficient is calculated. The baseline range is determined by analyzing the gene expression variation of wild-type algae. When establishing the screening determination equation, the deviation between the variation coefficient and the baseline value is considered. At the same time, a correction factor is introduced. The value of the correction factor is related to the cell growth state and environmental conditions. The output of the screening determination equation is the screening determination value, which ranges from 0.1 to 0.3. This range reflects the uniformity of target gene expression. The closer the value is to 0.2, the better the expression uniformity.

[0054] The specific implementation of step S09 is to evaluate the genetic stability of the engineered algal strain. First, the strain is continuously passaged for 5 generations under standard culture conditions, with an interval of 7 days between each generation. Samples are taken from each generation to measure the transcription level of the target gene. Real-time quantitative PCR is used, and each measurement is repeated 3 times. The average transcription level of each generation is calculated. The transcription level fluctuation rate is calculated by comparing the changes in transcription level between different generations. The fluctuation rate is calculated using the coefficient of variation method, which reflects the genetic stability of gene expression.

[0055] The specific implementation of step S10 is to calculate the comprehensive score of the engineered algal strain. First, a comprehensive scoring equation is established, which includes three main parameters: robustness assessment value, screening judgment value, and transcriptional level fluctuation rate value. Each parameter is assigned a different weight. The weight of the robustness assessment value is 0.4, reflecting the importance of the growth and metabolic stability of the engineered algal strain to the final score. The weight of the screening judgment value is also 0.4, reflecting the importance of gene expression uniformity. The weight of the transcriptional level fluctuation rate value is 0.2, representing the proportion of genetic stability in the score. Substituting these parameters into the comprehensive scoring equation, the score value of the engineered algal strain is calculated, with a score range of 0 to 100.

[0056] The specific implementation of step S11 is to conduct a final screening of the engineered algal strains. First, all the algal strains to be screened are sorted from high to low according to their comprehensive scores. The score distribution characteristics are analyzed to determine the screening threshold. The recommended threshold is 80 points. Algal strains with scores higher than this threshold have good growth stability, gene expression uniformity and genetic stability. The algal strain with the highest score is selected as the final screening result. The selected algal strains are subjected to morphological observation and functional verification to confirm their phenotypic characteristics and the functional expression of the target gene.

[0057] Among them, the negative ΔΔCt method of 2 (2 -ΔΔCt The ΔΔCt method is a commonly used computational method for relative quantitative PCR (qPCR) analysis, used to compare the relative differences in gene expression. Developed by Kenneth Livak and Thomas Schmittgen, it is also known as the Livak method or the ΔΔCt method.

[0058] In the existing technology, the calculation steps of this method are as follows:

[0059] 1. Calculate the ΔCt value of the target gene relative to the internal reference gene: ΔCt = Ct 目标基因 -Ct 内参基因 ;

[0060] 2. Calculate the ΔΔCt value of the treatment group relative to the control group: ΔΔCt = ΔCt 处理组 -ΔCt 对照组 ;

[0061] 3. Calculate relative expression level: Relative expression level = 2 -ΔΔCt ;

[0062] This formula is based on the principle of PCR amplification, assuming that in an ideal PCR reaction, the amount of DNA doubles with each cycle (efficiency of 100%). Therefore, powers of 2 can be used to represent changes in the amount of DNA. The negative sign is used because a smaller Ct value indicates a larger initial template amount, so this relationship needs to be reversed.

[0063] The advantage of this method is its simplicity and ease of use, as it does not require a standard curve and is suitable for high-throughput analysis. However, it assumes that the PCR efficiency is close to 100% and that the amplification efficiencies of the target gene and the internal reference gene are similar. If these conditions are not met, other methods that take into account actual amplification efficiency may need to be used.

[0064] To describe the calculation process involved in this invention in detail, the various data matrices and equations are now explained.

[0065] The morphological feature data matrix is ​​specifically represented as follows:

[0066]

[0067] In the formula, d ij e represents the cell diameter of the i-th sample, in micrometers; ij r represents the cell edge integrity value for the i-th sample, which is dimensionless; ij c is the nucleus-to-cytoplasm ratio of the i-th sample, dimensionless; ij denoted as chloroplast distribution density for the i-th sample, in units of chloroplasts per square micrometer; n is the number of samples.

[0068] The specific equation for calculating edge integrity is expressed as follows:

[0069]

[0070] In the formula, E is the edge integrity value; N is the number of edge pixels; and I is the image grayscale value. and These are the grayscale gradients in the x and y directions, respectively; L cont L represents the length of the continuous edge. total α is the total edge length; α is the weighting coefficient, ranging from 0.3 to 0.7.

[0071] The equation for calculating chloroplast distribution density is expressed as follows:

[0072]

[0073] In the formula, D chloro A represents the numerical value of chloroplast distribution density. cell n is the cell area; n is the number of chloroplasts; w i The weight of the i-th chloroplast; (x i y i (μ) represents the coordinates of the i-th chloroplast; x μ y ) represents the coordinates of the center of chloroplast distribution; (σ) x , σ y ) represents the standard deviation in the x and y directions.

[0074] The growth characteristic data matrix is ​​specifically represented as follows:

[0075]

[0076] In the formula, μ ij ρ represents the specific growth rate of the i-th sample, in units of hourly. ij B represents the cell density value for the i-th sample, in cells per milliliter; ij This represents the biomass value for the i-th sample, in grams per liter.

[0077] The specific equation for calculating the specific growth rate is expressed as follows:

[0078]

[0079] In the formula, μ is the specific growth rate; OD1 and OD2 are the optical density values ​​at time points t1 and t2, respectively; β is the growth regulation coefficient; γ is the time constant; and δ is the environmental impact correction term.

[0080] The photosynthesis data matrix is ​​represented as follows:

[0081]

[0082] In the formula, I ij Chl represents the light intensity value for the i-th sample, expressed in micromoles per square meter per second. ij Φ represents the chlorophyll content of the i-th sample, in milligrams per liter; ij Let NPQ be the dimensionless quantum yield value of the optical system II for the i-th sample; ij Let be the non-photochemical quenching coefficient value of the i-th sample, which is dimensionless.

[0083] The equation for calculating chlorophyll content is expressed as follows:

[0084] C chl =20.2D 645 +8.02D 663 +λ·T+η·pH;

[0085] In the formula, C chl This refers to the chlorophyll content value; D 645 and D 663 The absorbance values ​​are at wavelengths of 645 nm and 663 nm, respectively; T is the temperature in degrees Celsius; pH is the acidity / alkalinity; λ and η are the correction coefficients for temperature and pH, respectively.

[0086] The equation for calculating the quantum yield of the optical system II is expressed as follows:

[0087]

[0088] In the formula, Φ PSII F represents the quantum yield value of the optical system II; m F' represents the maximum fluorescence value under photoadapted conditions; s This represents the steady-state fluorescence value; F o ′ represents the minimum fluorescence value under photoadaptive conditions; I represents the light intensity; k represents the photoinhibition coefficient; and ξ represents the quantum efficiency correction term.

[0089] The specific equation for calculating the non-photochemical quenching coefficient is as follows:

[0090]

[0091] In the formula, NPQ is the non-photochemical quenching coefficient; F m The maximum fluorescence value under dark-adapted conditions; F m ′ represents the maximum fluorescence value under light-adapted conditions; ΔpH represents the proton gradient; t represents the illumination time; θ represents the pH sensitivity coefficient; and ω represents the time decay constant.

[0092] The integration site data matrix is ​​specifically represented as follows:

[0093]

[0094] In the formula, N ij D represents the number of chromosome integration sites in the i-th sample; ij denoted as the numerical value of the chromosome integration site distribution for the i-th sample.

[0095] The growth kinetics prediction equation is expressed as follows:

[0096]

[0097] In the formula, G is the predicted growth value; μ max Maximum specific growth rate; S is substrate concentration; K s X is the half-saturation constant; X is the biomass; X max Maximum biomass; t is the culture time; α is the time decay coefficient; β i For environmental factor weights; f i (E) is the environmental factor response function; ∈ is the random disturbance term.

[0098] The specific equation for predicting photosynthesis is expressed as follows:

[0099]

[0100] In the formula, P is the predicted photosynthetic efficiency; P max I represents the maximum photosynthetic rate; I represents the light intensity; I k I is the light saturation parameter; h The light-inhibition parameter is represented by Chl; the chlorophyll content is represented by K. chl Φ is the chlorophyll half-saturation constant. PSII denoted as the photosystem II quantum yield; NPQ is the non-photochemical quenching coefficient; γ is the energy dissipation regulation coefficient; and σ is the systematic error term.

[0101] The numerical calculation equation for robustness assessment is expressed as follows:

[0102]

[0103] In the formula, R is the robustness assessment value; CV g CV is the coefficient of variation of growth parameters. pS is the coefficient of variation of photosynthetic parameters; h denoted as Shannon entropy; ΔX represents biomass change; X0 represents initial biomass; w1, w2, w3, and w4 are weighting coefficients; and ψ represents the system fluctuation term.

[0104] The specific expression of the screening and judgment equation is as follows:

[0105]

[0106] In the formula, J is the screening and judgment value; CV exp The coefficient of variation for the experimental group; CV base The coefficient of variation for the control group; ΔE i For environmental factor bias; λ i φ represents the environmental factor weights; t represents the culture time; τ represents the time decay coefficient; and φ represents the random fluctuation term.

[0107] The equation for calculating the transcriptional level fluctuation rate is expressed as follows:

[0108]

[0109] In the formula, F represents the numerical value of transcriptional level fluctuation; T i The transcriptional level of the i-th generation; denoted as average transcription level; n is the generation number; ΔG is the genome instability index; κ is the genome instability weight; and ω is the systematic error term.

[0110] The comprehensive scoring equation is expressed as follows:

[0111]

[0112] In the formula, S is the engineered strain score; R is the robustness assessment score; J is the screening decision score; F is the transcriptional level fluctuation rate; Q i For quality control parameters; α i ρ represents the quality control weight; ρ is the system correction term.

[0113] The above equations are constructed based on the following considerations:

[0114] 1. The edge integrity calculation equation adopts a dual evaluation of gradient and continuity, taking into account both edge sharpness and integrity;

[0115] 2. The equation for calculating chloroplast distribution density uses a Gaussian kernel function to describe spatial distribution characteristics, reflecting the degree of chloroplast aggregation;

[0116] 3. The specific growth rate calculation equation adds a time-dependent term and an environmental correction term to the traditional calculation;

[0117] 4. The chlorophyll content calculation equation takes into account the effects of temperature and pH, thus improving the accuracy of the measurement;

[0118] 5. The quantum yield calculation equation for photosystem II introduces a light intensity dependence term, which more accurately describes the photochemical efficiency;

[0119] 6. The equation for calculating the non-photochemical quenching coefficient takes into account the influence of the proton gradient, reflecting the energy dissipation mechanism;

[0120] 7. The growth kinetics prediction equation combines the Monod equation and the environmental response function, which improves the prediction accuracy;

[0121] 8. The photosynthesis prediction equation integrates the light response curve and energy dissipation mechanism, thus improving the prediction model;

[0122] 9. The robustness assessment numerical calculation equations comprehensively consider variability, diversity, and stability;

[0123] 10. The screening and judgment equation incorporates environmental factor correction, which improves the accuracy of the judgment;

[0124] 11. The equation for calculating the transcriptional level fluctuation rate takes into account the impact of genomic instability;

[0125] 12. The comprehensive scoring equation adopts a weighted summation method and introduces quality control parameters to ensure the scientific nature and reliability of the scoring.

[0126] Specifically, the principle of this invention is based on the cross-integration of systems biology and deep learning. From a systems biology perspective, there are intrinsic connections between the phenotypic characteristics, growth properties, and metabolic activities of engineered algal strains. Cell morphology reflects the impact of gene integration on cell structure, growth characteristics reflect the impact of exogenous gene expression on cell physiology, and photosynthetic characteristics reflect changes in the metabolic network. By analyzing the correlations between these characteristics, the impact of gene integration can be systematically assessed.

[0127] The introduction of deep learning models makes the evaluation of such complex biological systems possible. The improved residual network structure effectively extracts the nonlinear relationships between multidimensional features, while the biological characteristic prediction module incorporates biological knowledge into the model, improving prediction accuracy. This model learns the intrinsic patterns between different features through extensive training data, thereby accurately assessing the overall state of engineered algal strains.

[0128] Based on this, the present invention establishes a complete evaluation index system. Robustness assessment values ​​reflect the overall stability of the engineered algal strains, screening criteria values ​​assess the uniformity of gene expression, and transcriptional level fluctuation rate reflects genetic stability. These indicators are calculated using rigorous mathematical models to ensure the objectivity and reliability of the evaluation results. A comprehensive scoring equation organically combines these indicators to achieve a comprehensive evaluation of the engineered algal strains. By setting reasonable weights and thresholds, this method can effectively screen engineered algal strains with good stability and high uniformity.

[0129] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0130] The specific implementation of step S01 involves using a microscopic image acquisition system to analyze the morphological characteristics of the target algal strain. This is achieved through the following sub-steps: Morphological observation of the target algal strain is performed using a high-resolution microscope, with an image resolution of 1920×1080 pixels and a magnification of 400x; image data of no less than 100 cells are acquired for each sample to ensure statistical significance; single-cell contour information is extracted using an image segmentation algorithm based on a hybrid strategy of region growth and edge detection, which can accurately identify cell boundaries; the Hough circle transform algorithm is applied to calculate the cell diameter, achieving accurate detection of circular targets with a measurement accuracy of 0.1 micrometers by transforming the image space to the parameter space; and a gradient-based edge detection algorithm is used to calculate the cell edge integrity, quantifying edge integrity by calculating the image grayscale gradient and edge continuity. The edge integrity value ranges from 0 to 1, and its calculation uses the formula... Where E is the edge integrity value, N is the number of edge pixels, and I is the image grayscale value. and The gray-level gradients in the x and y directions are respectively, L cont For continuous edge length, L total The total edge length is given by α, a weighting coefficient ranging from 0.3 to 0.7, with an edge integrity threshold of at least 0.85. A region growing algorithm is used to segment the nucleus and cytoplasm regions, and the nucleus-cytoplasm ratio is calculated; this ratio is within the normal range of 0.15 to 0.25. A density clustering algorithm is used to analyze the spatial distribution characteristics of chloroplasts and calculate their distribution density using the formula... Where D chloro A represents the chloroplast distribution density. cell w is the cell area, n is the number of chloroplasts, and w is the cell surface area. i Let x be the weight of the i-th chloroplast, (x i y i ) represents the coordinates of the i-th chloroplast, (μ) x μy ) represents the coordinates of the center of chloroplast distribution, (σ) x , σ y The standard deviations in the x and y directions are given; finally, these feature parameters are organized into a morphological feature data matrix M. morph The matrix structure is Where d ij e represents the cell diameter value of the i-th sample, in micrometers. ij Let r be the cell edge integrity value of the i-th sample, dimensionless. ij Let c be the nucleus-to-cytoplasm ratio of the i-th sample, dimensionless. ij Let represent the chloroplast distribution density of the i-th sample, in units of chloroplasts per square micrometer, and n be the number of samples.

[0131] The specific implementation of step S02 involves determining the growth characteristic parameters of the target algal strain under standard culture conditions. This is achieved through the following sub-steps: placing the algal strain under standard culture conditions of 200 micromoles per square meter per second, 25 degrees Celsius, and pH 7.0; sampling and measuring the optical density value every 4 hours to obtain growth curve data; and calculating the specific growth rate by fitting the optical density versus time curve using the following formula: (1-e -γt The formula is: β + δ, where μ is the specific growth rate, OD1 and OD2 are the optical density values ​​at time points t1 and t2 respectively, β is the growth regulation coefficient, γ is the time constant, and δ is the environmental correction term. Cell density was measured using a hemocytometer, with each sample measured three times and the average value taken. Algae were collected by centrifugation and dried at 60°C for 24 hours to determine the dry weight biomass. Finally, these parameters were organized into a growth characteristic data matrix M. growth The matrix structure is Where μ ij ρ represents the specific growth rate of the i-th sample, expressed in hours. ij B represents the cell density value for the i-th sample, in units of cells per milliliter. ij This represents the biomass value for the i-th sample, in grams per liter.

[0132] The specific implementation of step S03 involves determining the photosynthetic characteristics of the target algal strain, specifically through the following sub-steps: using a quantum light sensor to measure light intensity values ​​at three different locations on the surface of the culture flask and taking the average value; and using a spectrophotometer to determine the chlorophyll content value, specifically by extracting chlorophyll with 90% acetone and measuring absorbance at wavelengths of 663 nm and 645 nm. The chlorophyll content is calculated using the formula C. chl =20.2D 645 +8.02D 663 +λ·T+η·pH, where Cchl D represents the chlorophyll content value. 645 and D 663 The absorbance values ​​are at wavelengths of 645 nm and 663 nm, respectively, where T is temperature, pH is acidity / alkalinity, and λ and η are correction coefficients for temperature and pH, respectively. The photosystem II quantum yield was determined using a pulse-modulated chlorophyll fluorometer. Samples were dark-adapted for 15 minutes before measurement. The saturated pulse intensity was set to 3000 μmol / m² / s, and the pulse duration was 0.8 seconds. The calculation formula is as follows: Where Φ PSII F represents the quantum yield value of the optical system II. m F' represents the maximum fluorescence value under photoadaptive conditions. s The steady-state fluorescence value, F o ′ represents the minimum fluorescence value under photoadaptive conditions, I represents the light intensity, k represents the photoinhibition coefficient, and ξ represents the quantum efficiency correction term; simultaneously, the non-photochemical quenching coefficient is measured, and the calculation formula is: Where NPQ is the non-photochemical quenching coefficient, and F... m F represents the maximum fluorescence value under dark-adapted conditions. m ′ represents the maximum fluorescence value under photosynthetic conditions, ΔpH represents the proton gradient, t represents the illumination time, θ represents the pH sensitivity coefficient, and ω represents the time decay constant; finally, these parameters are organized into a photosynthesis data matrix M. photo The matrix structure is Among them I ij Chl represents the light intensity value of the i-th sample. ij Let Φ be the chlorophyll content value of the i-th sample. ij Let NPQ be the quantum yield value of the optical system II for the i-th sample. ij denoted as the non-photochemical quenching coefficient of the i-th sample.

[0133] The specific implementation of step S04 involves analyzing chromosome integration using fluorescence in situ hybridization (FISH), specifically through the following sub-steps: preparing chromosome slides, including trypsin digestion of cells, hypotonic treatment, and fixation with methanol and acetic acid; selecting DNA fragments complementary to the target gene sequence to prepare probes, and labeling them with biotin or digoxigenin; performing a hybridization reaction at 42 degrees Celsius for 16 hours; detecting the chromosomes using streptavidin-fluorescein labeling or anti-digoxigenin antibody; observing and photographing the chromosomes under a fluorescence microscope; and using image analysis software to statistically analyze the number and distribution of chromosome integration sites, obtaining an integration site data matrix M. int The matrix structure is Where N ij D represents the number of chromosomal integration sites in the i-th sample. ijdenoted as the numerical value of the chromosome integration site distribution for the i-th sample.

[0134] Step S05 is implemented by analyzing the transcriptional level of the target gene using flow cytometry, specifically through the following sub-steps: Collect algal cells in the logarithmic growth phase, ensuring cell viability; extract total RNA using an RNA extraction kit, strictly controlling RNase contamination during the process; reverse transcribe the RNA into cDNA using a reverse transcription kit, controlling the reaction temperature at 42 degrees Celsius; design specific primers for the target gene, with a primer length of 18 to 25 bases and an annealing temperature of 58 to 62 degrees Celsius; select a suitable housekeeping gene as an internal control, typically a stable gene such as 18S rRNA or β-aCtin; perform real-time quantitative PCR amplification using the SYBR Green method, with three replicates per sample; calculate the relative expression level of the target gene using the negative ΔCt method, the calculation process including: first, calculating the ΔCt value of the target gene relative to the internal control gene: ΔCt = Ct 目标基因 -Ct 内参基因 Then calculate the ΔΔCt value of the treatment group relative to the control group: ΔΔCt = ΔCt 处理组 -ΔCt 对照组 Finally, the relative expression level is calculated: relative expression level = 2 -ΔΔCt Each sample was measured three times, and the mean and standard deviation were calculated. The coefficient of variation of transcription intensity was calculated by dividing the standard deviation by the mean. The coefficient of variation reflects the stability of gene expression, and the smaller the coefficient of variation, the more stable the expression.

[0135] The specific implementation of step S06 involves constructing a deep neural network model to evaluate the engineered algal strain. This is achieved through the following sub-steps: designing an improved residual network structure, adding a biological characteristic prediction module to the standard residual network; the input layer receives morphological feature data matrices, growth feature data matrices, photosynthetic data matrices, and integration site data matrices; the first convolutional layer uses 64 3×3 convolutional kernels with a stride of 1; the second convolutional layer uses 128 3×3 convolutional kernels with a stride of 2; the third convolutional layer uses 256 3×3 convolutional kernels with a stride of 2; batch normalization and ReLU activation functions are applied after each convolutional layer; the biological characteristic prediction module includes growth kinetics prediction equations and photosynthetic prediction equations; the growth kinetics prediction equation uses a modified Mono equation, the expression of which is: Where G is the predicted growth value, μ max The maximum specific growth rate is given by S, where S is the substrate concentration and K is the K value. s Let X be the half-saturation constant, and X be the biomass. max The maximum biomass is given by t, the culture time is given by α, and the time decay coefficient is given by β. i For the environmental factor weights, fi (E) represents the environmental factor response function, and ε is the random perturbation term; the photosynthesis prediction equation adopts an improved light response curve model, and its expression is: Where P is the predicted photosynthetic efficiency, P max I represents the maximum photosynthetic rate, and I represents the light intensity. k I is the light saturation parameter. h Here, Chl represents the light-inhibiting parameter, K represents the chlorophyll content, and K represents the light-inhibiting parameter. chl Φ is the chlorophyll half-saturation constant. PSII denoted as σ, where NPQ is the non-photochemical quenching coefficient, γ is the energy dissipation adjustment coefficient, and σ is the systematic error term. The number of neurons in the two fully connected layers are 512 and 256, respectively. Dropout is used to prevent overfitting with a probability of 0.5. The output layer uses the Softmax activation function. The number of training samples is no less than 10,000. A stochastic gradient descent optimizer is used with a learning rate of 0.001. The batch size is set to 64. The number of training epochs is 100.

[0136] The specific implementation of step S07 involves using a trained deep neural network model to evaluate the robustness of the engineered algal strain. This is achieved through the following sub-steps: standardizing the morphological feature data matrix, growth feature data matrix, photosynthetic data matrix, and integration site data matrix using Z-score standardization to transform the data to the same scale; inputting the standardized data into the deep neural network model; extracting features through convolutional layers; calculating growth prediction values ​​and photosynthetic efficiency prediction values ​​through the biological characteristic prediction module; calculating the consistency between the predicted values ​​and the actual observed values; integrating all features through a fully connected layer; and calculating the robustness assessment value, using the following formula: Where R is the robustness assessment value, and CV g CV is the coefficient of variation of growth parameters. p S is the coefficient of variation of photosynthetic parameters. h denoted as Shannon entropy, ΔX as biomass change, X0 as initial biomass, w1, w2, w3, and w4 as weighting coefficients, and ψ as the system fluctuation term. The robustness assessment values ​​range from 0 to 1, with larger values ​​indicating better robustness of the engineered algal strain. A threshold of 0.8 is recommended.

[0137] The specific implementation of step S08 is based on constructing a screening equation according to the transcriptional intensity variation coefficient, which is achieved through the following sub-steps: analyzing the gene expression variation of wild-type algal strains to determine the baseline value range; considering the deviation between the variation coefficient and the baseline value, establishing a screening equation, the expression of which is: Where J is the selection criterion value, and CV exp The coefficient of variation (CV) for the experimental group. baseThe coefficient of variation for the control group is ΔE. i For environmental factor bias, λ i φ represents the environmental factor weights, t represents the culture time, τ represents the time decay coefficient, and φ represents the random fluctuation term. The value of the correction factor is related to the cell growth state and environmental conditions. The screening judgment value ranges from 0.1 to 0.3, which reflects the uniformity of target gene expression. The closer the value is to 0.2, the better the expression uniformity.

[0138] The specific implementation of step S09 is to evaluate the genetic stability of the engineered algal strain, which is achieved through the following sub-steps: The target algal strain is continuously passaged under standard culture conditions for 5 generations, with a 7-day interval between each generation to ensure consistency of culture conditions; the transcriptional level of the target gene is measured at each generation using real-time quantitative PCR, with 3 replicates per measurement; the average transcriptional level value for each generation is calculated; and the transcriptional level fluctuation rate is calculated by comparing the changes in transcriptional levels between different generations, using the following formula: Where F represents the transcriptional level fluctuation rate, and T... i The transcriptional level of generation i. The average transcription level is represented by n, the generation number is n, ΔG is the genomic instability index, κ is the genomic instability weight, and ω is the systematic error term. The fluctuation rate is calculated using the coefficient of variation method, which reflects the genetic stability of gene expression.

[0139] The specific implementation of step S10 involves calculating the comprehensive score of the engineered algal strain, specifically through the following sub-steps: Establishing a comprehensive scoring equation, comprising three main parameters: robustness assessment value, screening judgment value, and transcriptional level fluctuation rate value; setting weights for each parameter, where the robustness assessment value has a weight of 0.4, reflecting the importance of the growth and metabolic stability of the engineered algal strain to the final score; the screening judgment value also has a weight of 0.4, reflecting the importance of gene expression uniformity; and the transcriptional level fluctuation rate value has a weight of 0.2, representing the proportion of genetic stability in the score; substituting these parameters into the comprehensive scoring equation, its expression is: Where S is the engineered strain score, R is the robustness assessment score, J is the screening criterion score, F is the transcriptional level fluctuation rate, and Q is the score for the engineered strain. i α is a quality control parameter. i ρ represents the quality control weight, and ρ represents the system correction term; the score range is 0 to 100 points.

[0140] The specific implementation of step S11 involves the final screening of engineered algal strains, which is achieved through the following sub-steps: All algal strains to be screened are sorted from highest to lowest based on their comprehensive scores, and a sorting database is established; the score distribution characteristics are analyzed, including statistical parameters such as mean, standard deviation, skewness, and kurtosis; a screening threshold is determined, with a suggested threshold of 80 points, as algal strains exceeding this threshold exhibit good growth stability, gene expression uniformity, and genetic stability; the algal strain with the highest score is selected as the final screening result; morphological observations are performed on the selected algal strains, including cell size, morphological integrity, nucleocytoplasmic ratio, and chloroplast distribution; functional verification is conducted, validating the functional expression of target genes, including protein expression levels, enzyme activity, or other functional indicators; a screening result database is established, recording the score, parameter values, and verification results for each algal strain; the selected dominant algal strains are preserved and propagated to establish a germplasm resource bank; a detailed screening report is written, including the screening process, scoring criteria, data analysis, and result verification.

[0141] The method of this invention is applicable to the following chromosomal integrated genes: hematopoietic cell differentiation antigen CD45 gene, T lymphocyte common antigen CD3 gene, helper T lymphocyte marker CD4 gene, cytotoxic T lymphocyte marker CD8 gene, natural killer cell marker CD16 gene, neural cell adhesion molecule CD56 gene, B lymphocyte marker CD19 gene, complement regulatory protein CD55 gene, complement regulatory protein CD59 gene, interleukin-2 receptor α chain CD25 gene, interleukin-7 receptor α chain CD127 gene, programmed death receptor 1 gene, T cell costimulatory molecule CD28 gene, immunoglobulin Fc receptor CD64 gene, complement receptor 3 gene, complement receptor 1 gene, and perforin gene. This method is also applicable to the following algae that have been domesticated in the laboratory and for which ribosome entry site sequences, chloroplast targeting sequences, and endoplasmic reticulum localization sequences have been established: *Chlamydomonas reinhardtii*, *Chlorella vulgaris*, *Dunaliella salina*, *Prorocentrum dinoflagellates*, *Alexandrium*, *Platycodon grandiflorus*, *Scenedesmus*, *Euglena*, *Rhodophyta*, and *Microcera*. The screening method provided by this invention can effectively screen for engineered algal strains with multi-copy chromosome integration, providing important technical support for algal genetic engineering breeding.

[0142] To better understand and implement this invention, a specific application scenario, Example 2, is provided below: Researchers selected *Chlamydomonas reinhardtii* as the host cell and the CD59 gene as the target integration gene for the experiment. First, the target gene was introduced into the algal strain via a ribosome entry site sequence. Then, the integration status of the engineered algal strain was evaluated using the screening method of this invention.

[0143] A batch of engineered algal strains were selected. First, morphological characteristics of the target algal strains were analyzed using a microscopic image acquisition system. High-resolution images of 150 cells were acquired at a magnification of 400x. Morphological feature parameters were extracted using image analysis software, as shown in Table 1.

[0144] Table 1. Statistical Table of Morphological Characteristics of Engineering Algal Strains

[0145] parameter average value Standard deviation coefficient of variation Cell diameter (μm) 8.45 0.62 0.073 Cell edge integrity 0.92 0.05 0.054 Nucleus-to-cytoplasm ratio 0.19 0.02 0.105 <![CDATA[Chloroplast distribution density (number / μm 2 )]]> 0.85 0.09 0.106

[0146] Under standard culture conditions, the growth characteristics of the engineered algal strains were analyzed. Samples were taken every 4 hours and monitored continuously for 72 hours to obtain growth curve data and calculate growth parameters. The results are shown in Table 2.

[0147] Table 2. Growth Characteristics Parameters of Engineered Algal Strains

[0148] Time (h) <![CDATA[Optical Density (OD 680 )]]> <![CDATA[Cell density (×10 6 / mL)]]> Biomass (g / L) 0 0.15 1.2 0.45 24 0.42 3.5 1.28 48 0.89 7.4 2.65 72 1.35 11.2 4.12

[0149] Figure 2 The growth curves of the engineered algal strain during a 72-hour culture process are shown, including optical density (OD). 680 The changes in light density and cell density were observed. Two curves showed similar growth trends but different numerical ranges, more clearly illustrating the slow growth phase of the algal strain in the early stage (0-24h), the rapid growth phase in the middle stage (24-48h), and the stable phase in the later stage (48-72h) with a gradually decreasing growth rate. The trends in light density and cell density remained consistent, indicating that the engineered algal strain had a stable growth state. Based on the growth curve data, the specific growth rate was calculated to be μ = 0.048h. -1 The photosynthetic characteristics of the engineered algal strains were measured. Photosynthetic parameters were measured under different light intensities, and the results are shown in Table 3.

[0150] Table 3. Photosynthetic characteristics parameters of engineered algal strains

[0151]

[0152] Figure 3 The relationship between photosynthetic parameters and light intensity is shown. This figure comprehensively illustrates the changing trends of three important photosynthetic parameters: photosystem II quantum yield (ΦPSII), non-photochemical quenching coefficient, and chlorophyll content under different light intensities. With increasing light intensity, the photosystem II quantum yield shows a decreasing trend, while the non-photochemical quenching coefficient gradually increases, indicating that the algae possess a good photoprotection mechanism. Chlorophyll content (shown as a bar chart) increases slightly with increasing light intensity but tends to stabilize under high light, reflecting the algae's photoadaptability. Fluorescence in situ hybridization was used to analyze chromosome integration, and the distribution of integration sites in 200 cells was statistically analyzed. The results are shown in Table 4.

[0153] Table 4. Statistics on Chromosomal Integration Sites

[0154] Number of integration sites Cell count frequency 1 24 0.12 2 86 0.43 3 62 0.31 4 28 0.14

[0155] The transcriptional level of the target gene CD59 was determined using real-time quantitative PCR, with 18S rRNA as an internal reference gene. The relative expression level was calculated using the negative ΔCt method (2). The results showed that the average Ct value for the target gene was 23.45, and the average Ct value for the internal reference gene was 18.32, resulting in a calculated ΔCt value of 5.13. The coefficient of variation for transcriptional intensity was calculated to be 0.156.

[0156] The above data was input into a deep neural network model for analysis. The model adopted a structure of 3 convolutional layers and 2 fully connected layers, and was trained using 10,000 sets of historical data. The biological characteristic prediction module used an improved Mono equation to predict the growth trend, and the calculated growth prediction value was 0.892; the light response curve model was used to predict the photosynthetic efficiency, and the calculated photosynthetic efficiency prediction value was 0.785.

[0157] Based on the model analysis results, the robustness assessment value was calculated to be 0.845, which is above the recommended threshold of 0.8, indicating that the engineered algal strain has good stability. A screening decision equation was constructed based on the transcription intensity coefficient of variation, and the calculated screening decision value was 0.183, which is within the ideal range of 0.1 to 0.3 and close to the optimal value of 0.2, indicating that gene expression has good uniformity.

[0158] The engineered algal strain was cultured for five consecutive generations, with a 7-day interval between each generation. Changes in the transcriptional level of the target gene were monitored, and the results are shown in Table 5.

[0159] Table 5. Analysis of Gene Expression Stability in Continuous Culture

[0160] generation relative expression level coefficient of variation Generation 1 1.00 0.156 2nd generation 0.97 0.162 3rd generation 0.95 0.158 4th generation 0.96 0.165 5th generation 0.94 0.160

[0161] The transcriptional level fluctuation rate was calculated to be 0.124, indicating that the gene expression exhibits good genetic stability. Substituting the robustness assessment value, screening determination value, and transcriptional level fluctuation rate into the comprehensive scoring equation, the engineered strain was scored 86.5 points, which is higher than the screening threshold of 80 points.

[0162] The verification in this embodiment reveals the following main problems with traditional techniques for screening engineered algal strains: 1. Single-index evaluation is insufficient to comprehensively reflect the integration status of engineered algal strains; 2. There is a lack of dynamic assessment of growth stability and metabolic activity; 3. Long-term genetic stability cannot be accurately predicted; 4. The screening process is highly subjective and lacks a standardized evaluation system. In contrast, the method of this invention has the following advantages: 1. Comprehensive evaluation of engineered algal strains is achieved through multi-dimensional data analysis; 2. The introduction of a deep learning model improves the accuracy and objectivity of screening; 3. A standardized scoring system is established, making the screening results comparable; 4. Accurate prediction of the long-term stability of algal strains is achieved through a biological characteristic prediction module; 5. The comprehensive scoring equation organically combines multiple indicators, avoiding the limitations of single-index evaluation.

[0163] It should be noted that the variables involved in this invention are explained in detail in Tables 6 and 7 below.

[0164] Table 6. Variable Explanation Table (Part 1)

[0165]

[0166] Table 7. Variable Explanation Table (Part Two)

[0167]

[0168] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for screening multiple copy integration of chromosomes in engineered algal strains, characterized in that, include: Construct a data matrix of morphology, growth, photosynthesis, and integration sites; A deep neural network model was used to calculate the robustness assessment values; a screening decision equation was constructed to calculate the screening decision values. The transcriptional level fluctuation rate is calculated; the robustness assessment value, screening judgment value, and transcriptional level fluctuation rate value are input into the comprehensive scoring equation to calculate the engineered strain score; the engineered strains are sorted according to their scores and the strain with the highest score is selected; the deep neural network model sets a biological characteristic prediction module between the convolutional layer and the fully connected layer. This module includes a growth kinetics prediction equation and a photosynthesis prediction equation to improve the screening accuracy.

2. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 1, characterized in that, The steps for constructing the morphological data matrix specifically involve acquiring microscopic images of the target algal strain, measuring cell diameter, cell edge integrity, nucleocytoplasmic ratio, and chloroplast distribution density, and then constructing the morphological feature data matrix.

3. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 2, characterized in that, The step of constructing the growth characteristic data matrix specifically involves measuring the specific growth rate, cell density, and biomass of the target algal strain during the exponential growth phase, and then constructing the growth characteristic data matrix.

4. The screening method for multi-copy integration of chromosomes in engineered algal strains according to claim 3, characterized in that, The step of constructing the photosynthesis data matrix specifically involves measuring the photosynthetic characteristic parameters of the target algae, including light intensity, chlorophyll content, photosystem II quantum yield, and non-photochemical quenching coefficient, to construct the photosynthesis data matrix.

5. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 4, characterized in that, The step of constructing the integration site data matrix specifically involves using fluorescence in situ hybridization (FISH) to determine the number and distribution of chromosome integration sites, and then constructing the integration site data matrix.

6. The screening method for multiple copy integration of chromosomes in engineered algal strains according to claim 5, characterized in that, The step of constructing the screening and determination equation specifically involves using flow cytometry to determine the transcriptional level of the target gene, calculating the transcriptional intensity variation coefficient, constructing the screening and determination equation based on the transcriptional intensity variation coefficient, and calculating the screening and determination value. The threshold range of the screening and determination value is 0.1 to 0.

3.

7. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 6, characterized in that, The step of calculating the transcriptional level fluctuation rate involves continuously culturing for 5 generations, measuring the change in the transcriptional level of the target gene, and calculating the transcriptional level fluctuation rate to assess the genetic stability of the target chromosome integration.

8. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 7, characterized in that, The deep neural network model adopts an improved residual network structure, including an input layer, three convolutional layers, a biological characteristic prediction module, two fully connected layers, and an output layer.

9. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 8, characterized in that, The growth kinetics prediction equation in the biological characteristic prediction module has input parameters including specific growth rate, cell density, and biomass, and output parameters are growth prediction values.

10. The screening method for multi-copy chromosome integration of engineered algal strains according to claim 9, characterized in that, The input parameters of the photosynthesis prediction equation in the biological characteristic prediction module include light intensity, chlorophyll content, photosystem II quantum yield, and non-photochemical quenching coefficient, and the output parameter is the predicted photosynthetic efficiency.

Images