Prediction method and device for transfected cell screening, electronic equipment and storage medium

Abstract

The invention provides a prediction method and device for cell screening after transfection, electronic equipment and a storage medium, and relates to the technical field of computers. Comprising the following steps: acquiring data of a high-quality cell screening project as an original data set, preprocessing the data set, and checking the rationality of a data preprocessing strategy to obtain feature data; and constructing a fusion sequence self-attention network model, and inputting the obtained feature data into the fusion sequence self-attention network model to obtain a prediction result. According to the method, data of a high-quality cell screening project is adopted as an original data set, and the problem of unbalanced data labels in the original data set is solved by adopting SMOTE; rich feature representation is provided for an Encoder module through an FSAN model in combination with the long-term dependence processing capability of LSTM and the local mode capture capability of CNN; and identifying and paying attention to a complex relationship between data features extracted by the LSTM and the CNN through a self-attention layer.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a prediction method and device, electronic equipment, and storage medium for post-transfection cell screening. Background Art

[0002] In the field of modern biomedicine, the development of biopharmaceutical technology has greatly promoted the progress of treatment of various diseases. As an important means of biopharmaceuticals, cell engineering pharmaceuticals have formed a very large industrial scale. Cell engineering pharmaceuticals are a method of using genetic engineering or cell culture technology to transform cells so that they can produce drugs. Therefore, it is crucial to obtain high-quality modified cells, which will directly determine the quality and efficiency of drug production. The most commonly used cell modification method is transfection, which enables cells to produce specific drugs by introducing specific exogenous nucleic acids into host cells. However, due to the randomness of the gene sequence position of exogenous nucleic acids inserted into host cells, high-quality cells with good growth, high expression levels and stable expression levels need to be screened from a large number of cells after transfection.

[0003] The currently widely used screening method mainly includes two stages. The first stage is the deep-well plate stage. The large number of cells obtained by transfection are grouped and then cultured in deep-well plates. During the culture period, cells with poor growth are screened out, and only cells with good performance are retained. The second stage is the shake flask stage. The cells screened out in the deep-well plate stage are placed in shake flasks for continued culture. During this period, the culture data such as cell concentration and average cell particle size will be measured regularly, and cells that do not meet the conditions will continue to be screened out based on the measured data. Finally, only a few groups of cells with excellent performance will be retained for the subsequent process.

[0004] The above screening methods rely on time-consuming culture, which has become a bottleneck limiting the efficiency of cell engineering pharmaceuticals. Long-term culture and screening require a large amount of professional equipment and well-trained operators, resulting in high costs. Frequent intervention of human operations also greatly increases the risk of cell contamination and experimental errors. Cell screening is determined by the operator, and the screening results are easily affected by the operator's subjective judgment, which greatly weakens the reliability and consistency of the screening results. Summary of the invention

[0005] The purpose of the present invention is to develop an automated and efficient cell screening prediction method and device, electronic equipment, and storage medium to address the problems of long screening process, reducing labor costs and subjective human bias in the screening process, and improving screening quality. The specific technical solution is as follows.

[0006] The present invention provides a prediction method for post-transfection cell screening, comprising the following steps: Step 1: Obtain data from high-quality cell screening projects as the original data set; Step 2: Preprocess the data in the original data set, and use correlation analysis and significance to test the rationality of the data preprocessing strategy to obtain feature data; Step 3: Construct a fusion sequence self-attention network model, input the obtained feature data into the fusion sequence self-attention network model, and obtain the prediction result.

[0007] Furthermore, the original data set includes multiple data, each of which includes a time series with a length of 10 days and six key characteristic indicators; the six key characteristic indicators are viable cell concentration, total cell concentration, total cell number, average particle size, average roundness and agglomeration rate.

[0008] Furthermore, the collection process of a single piece of data is as follows: ①. In the deep-well plate stage, data collection was performed only once after four days of culture to obtain one piece of data; ② In the shaking bottle stage, one piece of data is collected on the first day, third day, fifth day, seventh day, eighth day, ninth day, tenth day, twelfth day and fourteenth day of the shaking bottle stage, and nine pieces of data are obtained; and cells that do not meet the requirements are screened out according to the growth of cells, and the incomplete time series are filled with 0; ③. Combine one piece of data obtained in the deep well plate stage with nine pieces of data obtained in the shake bottle stage to obtain a time series with a length of 10 days.

[0009] Furthermore, the specific process of preprocessing the data in the original data set is as follows: Select any two data in the original data set with label 1, use the median interpolation method to standardize the data in the selected two original data sets with label 1, and generate a new label New sample data ; All data in the original data set are standardized to obtain a standardized original data set; Add ±0.2 random perturbations to the standardized data, and Generate four new sample data , get the preprocessed data set; The normalized original data set was subjected to Pearson correlation analysis with the preprocessed data set. Test to obtain feature data, which includes the correlation changes between different features and its corresponding characteristics value.

[0010] Furthermore, the data in the original data set is stored in a CSV file; Pearson correlation analysis and The test results are stored in CSV files.

[0011] Furthermore, the process of constructing the fusion sequence self-attention network model is as follows: The LSTM neural network model is configured as a two-layer structure, each layer of which contains a variable configuration of 16 to 64 hidden units, and the dropout function is set to 0.5; The convolutional neural network is set to a variable configuration containing 1 to 2 convolutional layers, using a convolution kernel of size 2 and no pooling layer to capture the dynamic changes of cells in a short period of time; The Encoder module based on the self-attention mechanism uses single-head attention, followed by a single-layer feedforward neural network. The Encoder module uses the single-head attention mechanism to process the input information. The implementation method of single-head attention is to obtain the query, key, and value matrix through linear transformation of the input. The attention mechanism assigns different weights to inputs at different positions by calculating the similarity between the query and the key, thereby capturing the correlation and dependency between inputs at different positions. After the attention output, the data is further nonlinearly transformed through a single-layer feedforward neural network to enrich the feature expression. Concatenate the last layer output of LSTM with the flattened output of CNN as the input of Encoder; The output of the encoder passes through a Softmax activation function to produce the final binary classification prediction fusion sequence self-attention network model.

[0012] Furthermore, the specific process of inputting the obtained feature data into the fusion sequence self-attention network model for prediction is as follows: The preprocessed data set is divided into training set, validation set and test set in a ratio of 6:2:2, and all random seeds are set to 202; The Encoder module of the fused sequence self-attention network model is used to train the training set, and the Adam optimizer and cross entropy loss function are used when training the training set. The accuracy, precision, recall and F1 score are used as evaluation indicators of the fused sequence self-attention network model to obtain the prediction results.

[0013] The present invention also provides a post-transfection cell screening prediction device for implementing the post-transfection cell screening prediction method as described above, comprising: The original data set is imported into the unit to obtain the original data set of cell screening; The original data set preprocessing unit preprocesses the original data set to obtain feature data; The cell screening prediction unit inputs the preprocessed feature data into the fusion prediction model for prediction, obtains the prediction results, and exports the prediction result file.

[0014] The present invention also provides an electronic device, comprising a processor and a memory connected to each other; The memory is used to store a computer program that supports the electronic device to execute the prediction method for cell screening, and the computer program includes program instructions; The processor is configured to call the program instructions to execute the prediction method for cell screening as described above.

[0015] The present invention also provides a storage medium storing a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the prediction method for cell screening as described above.

[0016] The application of the technical solution of the present invention has at least the following beneficial effects: (1) The present invention uses the data of high-quality cell screening projects as the original data set, and adopts the Synthetic Minority Over-sampling Technique (SMOTE) to solve the problem of data label imbalance in the original data set, and then alleviates the problem of small data volume through data enhancement, completing the preprocessing of the original data set; then, the FSAN model combines the long-term dependency processing capability of LSTM and the local pattern capture capability of CNN to provide a rich feature representation for the Encoder module; through the self-attention layer, the model can identify and focus on the complex relationship between the data features extracted by LSTM and CNN. In order to ensure the effectiveness and practicality of the model, based on the consideration of model complexity and data characteristics, special attention was paid to the design and parameter selection of the fusion sequence self-attention network model during construction. The constructed fusion sequence self-attention network model can effectively process data of different time series lengths, providing researchers with a trade-off solution between prediction accuracy and shortening of cell screening time. This method highlights the potential of effectively using deep learning technology for time series analysis in cell screening.

[0017] (2) The present invention proposes a cell screening prediction device that can implement the cell screening prediction method.

[0018] (3) The present invention proposes an electronic device capable of storing and executing a computer program of the cell screening prediction method.

[0019] (4) The present invention proposes a storage medium capable of storing a computer program containing specific program instructions, which, when executed by a processor, causes the processor to execute the cell screening prediction method.

[0020] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 is a flow chart of a cell screening prediction method according to Example 1 of the present invention; Figure 2 It is a structural diagram of the FSAN fusion sequence self-attention network model in Example 1 of the present invention; Figure 3 It is a schematic diagram of the prediction process of the cell culture prediction device in Example 2 of the present invention. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned purposes, features and advantages of the present invention more clear and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings. It should be noted that the drawings of the present invention are all simplified and use non-precise proportions, which are only used to conveniently and clearly assist in explaining the implementation of the present invention; the "number" mentioned in the present invention is not limited to the specific number in the examples in the accompanying drawings; the directions or positional relationships indicated by "front", "middle", "back", "left", "right", "up", "down", "top", "bottom", "middle", etc. mentioned in the present invention are based on the directions or positional relationships shown in the drawings of the present invention, and do not indicate or imply that the devices or components referred to must have a specific direction, nor can they be understood as limitations on the present invention. Embodiment 1:

[0023] The present invention provides a prediction method for screening cells after transfection, see Figure 1 As shown, the following steps are included: Step 1: Obtain data from high-quality cell screening projects as the original data set.

[0024] Preferably, the original data set includes data from 9 high-quality cell screening projects that have been put into drug production.

[0025] Furthermore, the original data set includes 217 data of various product types such as monoclonal antibodies, recombinant proteins, vaccines, etc. Each data includes a time series with a length of 10 days and six key characteristic indicators; the six key characteristic indicators are viable cell concentration (F1), total cell concentration (F2), total cell number (F3), average particle size (F4), average roundness (F5) and agglomeration rate (F6).

[0026] Furthermore, the data in the original data set is stored in a CSV file. Specifically, the viable cell concentration (F1), total cell concentration (F2), total cell number (F3), average particle size (F4), average roundness (F5) and agglomeration rate (F6) are used as table headers, and the viable cell concentration (F1), total cell concentration (F2), total cell number (F3), average particle size (F4), average roundness (F5) and agglomeration rate (F6) in each row of data are separated by English commas as delimiters.

[0027] Preferably, the collection process of a single piece of data is as follows: ① In the deep-well plate stage, due to the small amount of cell culture fluid, sampling will cause a large loss. Therefore, data collection is only performed once after four days of culture to obtain one piece of data; ② In the shaking flask stage, one piece of data was collected on the first day, third day, fifth day, seventh day, eighth day, ninth day, tenth day, twelfth day and fourteenth day of the shaking flask stage, and nine pieces of data were obtained; and cells that did not meet the requirements were screened out according to the growth of cells, so some cells did not have complete culture data, and the incomplete time series was filled with 0; ③. Combine one piece of data obtained in the deep well plate stage with nine pieces of data obtained in the shake bottle stage to obtain a time series with a length of 10 days.

[0028] Step 2: Preprocess the data in the original data set and use correlation analysis and significance test to verify the rationality of the data preprocessing strategy to obtain feature data.

[0029] Preferably, the specific process of preprocessing the data in the original data set is as follows: S2.1. Use the SMOTE algorithm to perform label balancing on the data in the original data set to obtain a standardized original data set. The specific method is as follows: Select any two data with label 1, use median interpolation method to standardize the selected two data with label 1, and generate new label New sample data ; All data in the original data set are standardized to obtain a standardized original data set; The expression of the SMOTE algorithm is as follows: ; in, The new sample data is generated; is a sample in the minority class, where the minority class is a class without complete training data; for One of the k-nearest neighbor model samples; is a random number, taking the value in The interval is used to control the random offset of new samples.

[0030] S2.2. New sample data Perform data enhancement processing; the specific method is: Add ±0.2 random perturbations to the standardized data, and Generate four new sample data , and obtain the preprocessed data set; by expanding the original data set, the FSAN fusion sequence self-attention network model is improved for new sample data adaptability and generalization ability.

[0031] S2.3, Pearson correlation analysis was performed on the standardized original data set and the preprocessed data set. Test (unequal variances) to get the correlation changes between different features and its corresponding characteristics value to judge the rationality of preprocessing the data in the original data set.

[0032] Correlation changes between different features obtained using Pearson correlation analysis The expression is as follows: ; in, is any one of the key characteristic indicators, is any one of the key characteristic indicators, , yes The mean of for The mean of is the total number of correlation analyses of the six key feature indicators, is the number of key characteristic indicators (i.e., is 6); The expression for the test (unequal variances) is as follows: ; in, and are the mean of the original data set samples and the mean of the preprocessed data set samples, respectively. and are the variance of the original data set samples and the variance of the preprocessed data set samples, respectively. and are the sample size of the original dataset and the sample size of the preprocessed dataset, respectively.

[0033] Pearson correlation analysis and The results of the test (unequal variances) are stored in CSV files. Statistics, The value is the table header, in each row of data Statistics, The values ​​are separated by commas. The value represents the observed The probability of a statistic or more extreme situation occurring. Value and The statistics are used together to assess whether the correlations between different features are statistically significant.

[0034] The correlation between different features and Statistics, The corresponding relationship of the values ​​is shown in Table 1.

[0035] Table 1: Correlation between different features and Statistics, The corresponding relationship of values .

[0036] Step 3: Construct the FSAN fusion sequence self-attention network model, input the obtained feature data into the FSAN fusion sequence self-attention network model, and obtain the prediction result.

[0037] Preferably, see Figure 2 As shown in Figure 2, the process of constructing the FSAN fusion sequence self-attention network model is as follows: Step 3.1, configure the long short-term memory neural network model (i.e., LSTM) as a two-layer structure, each layer of which contains a variable configuration of 16 to 64 hidden units, and the dropout function is set to 0.5 to avoid overfitting; Step 3.2, set the convolutional neural network model (i.e., CNN) to a variable configuration containing 1 to 2 convolutional layers, using a convolution kernel of size 2, and no pooling layer to capture the dynamic changes of cells in a short period of time; Step 3.3, the Encoder module based on the self-attention mechanism uses single-head attention, followed by a single-layer feedforward neural network (FNN). Among them, the Encoder module uses a single-head attention mechanism to process the input information. The implementation method of single-head attention is to obtain the query, key, and value matrix through a linear transformation of the input. Subsequently, the attention mechanism assigns different weights to inputs at different positions by calculating the similarity between the query and the key, thereby capturing the correlation and dependency between inputs at different positions. After the attention output, the data is further nonlinearly transformed through a single-layer feedforward neural network to enrich the feature expression; Step 3.4: Concatenate the last layer output of LSTM (i.e., the latent vector of LSTM) with the flattened output of CNN (i.e., the latent vector of CNN) as the input of the Encoder module; Step 3.5: The output of the Encoder module passes through a Softmax activation function to generate the final binary classification prediction fused sequence self-attention network model. The Encoder module of the fused sequence self-attention network model (FSAN) mainly combines the self-attention mechanism and sequence feature extraction technology.

[0038] The expression of the activation function is as follows: ; in, is the number of categories for the input data, For the The values ​​corresponding to the categories are For the The value in the category The value at time.

[0039] As a further solution of this embodiment, when the Encoder module of FSAN is used for training, the Adam optimizer and the cross entropy loss function are used. The preprocessed data set is divided into a training set, a validation set, and a test set in a ratio of 6:2:2. All random seeds are set to 202.

[0040] As a further solution of this embodiment, accuracy, precision, recall and F1 score are used as evaluation indicators of the FSAN model. Precision reflects the proportion of predicted positive classes that truly meet production requirements and is an intuitive measure of the model effect; recall measures the proportion of positive samples identified by the model in all positive samples, which is directly related to the omission rate of the screening process; their harmonic mean F1 score is used as the main evaluation indicator. The experimental results were processed and analyzed, and F1 was used as the main evaluation indicator. After more than 5 days of culture in the shake bottle stage, the F1 score of FSAN stabilized at more than 80%, and the best case reached 94.3%.

[0041] The experimental results of FSAN are shown in Table 2.

[0042] Table 2: Experimental results of FSAN . Embodiment 2:

[0043] The present invention also provides a prediction device for screening cells after transfection, see Figure 3 As shown, the prediction device comprises: The original data set is imported into the unit to obtain the original data set of cell screening; The original data set preprocessing unit preprocesses the original data set to obtain feature data; The cell screening prediction unit inputs the preprocessed feature data into the fusion prediction model for prediction, obtains the prediction results, and exports the prediction result file.

[0044] The original data set preprocessing unit comprises: A basic data file module, including a cell screening database and an import module, wherein the cell screening database obtains and stores basic data files, and the import module imports the basic data files into the data preprocessing module; the basic data files are original data sets; The data preprocessing module fills the incomplete time series with 0 and marks them, uses the SMOTE algorithm to balance the labels, normalizes the data, performs correlation analysis on the cell screening parameters, retains the data with a correlation coefficient less than 0.12 as feature data, and outputs the feature data to the cell screening prediction unit.

[0045] The cell screening prediction unit comprises: The prediction model generation module uses the feature data output by the original data set preprocessing unit to train and test multiple machine learning fusion prediction models; Prediction model storage module, which stores multiple trained machine learning fusion prediction models; The prediction module inputs the feature data obtained by preprocessing the original data set into the prediction model, outputs the prediction results, and exports the prediction result file. Embodiment 3:

[0046] The present invention also provides an electronic device, comprising a processor and a memory, wherein the processor and the memory are interconnected, wherein the memory is used to store a computer program that supports the electronic device to execute the prediction method for cell screening, the computer program comprises program instructions, and the processor is configured to call the program instructions to execute the prediction method for cell screening as described in any one of the above. Embodiment 4:

[0047] The present invention also provides a storage medium storing a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the prediction method for cell screening as described in any one of the above.

[0048] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A prediction method for screening cells after transfection, characterized in that: The following steps are involved: Step 1, obtaining cell screening data as an original data set; Step 2: Preprocess the data in the original data set, and use correlation analysis and significance to test the rationality of the data preprocessing strategy to obtain feature data; the specific process of preprocessing the data in the original data set is as follows: Select any two data in the original data set with label 1, use the median interpolation method to standardize the data in the selected two original data sets with label 1, and generate a new label New sample data ; All data in the original data set are standardized to obtain a standardized original data set; Add ±0.2 random perturbations to the standardized data, and Generate four new sample data , get the preprocessed data set; The normalized original data set was compared with the preprocessed data set by Pearson correlation analysis. Test to obtain feature data, which includes the correlation changes between different features and its corresponding characteristics value; Step 3: Construct a fusion sequence self-attention network model, input the obtained feature data into the fusion sequence self-attention network model, and obtain the prediction result.

2. The prediction method for post-transfection cell screening according to claim 1, characterized in that: The original data set includes multiple data, each of which includes a time series with a length of 10 days and six key characteristic indicators; the six key characteristic indicators are viable cell concentration, total cell concentration, total cell number, average particle size, average roundness and agglomeration rate.

3. The prediction method for post-transfection cell screening according to claim 2, characterized in that: The process of collecting a single piece of data is as follows: ①. In the deep-well plate stage, data collection was performed only once after four days of culture to obtain one piece of data; ② In the shaking bottle stage, one piece of data is collected on the first day, third day, fifth day, seventh day, eighth day, ninth day, tenth day, twelfth day and fourteenth day of the shaking bottle stage, and nine pieces of data are obtained; and cells that do not meet the requirements are screened out according to the growth of cells, and the incomplete time series are filled with 0; ③. Combine one piece of data obtained in the deep well plate stage with nine pieces of data obtained in the shake bottle stage to obtain a time series with a length of 10 days.

4. The prediction method for post-transfection cell screening according to claim 1, characterized in that: The data in the original data set is stored in a CSV file; Pearson correlation analysis and The test results are stored in CSV files.

5. The prediction method for screening transfected cells according to any one of claims 1 to 4, characterized in that: The process of constructing the fusion sequence self-attention network model is as follows: The LSTM neural network model is configured as a two-layer structure, each layer of which contains a variable configuration of 16 to 64 hidden units, and the dropout function is set to 0.5; The convolutional neural network is set to a variable configuration containing 1 to 2 convolutional layers, using a convolution kernel of size 2 and no pooling layer to capture the dynamic changes of cells in a short period of time; The Encoder module based on the self-attention mechanism uses single-head attention, followed by a single-layer feedforward neural network. The Encoder module uses the single-head attention mechanism to process the input information. The implementation method of single-head attention is to obtain the query, key, and value matrix through linear transformation of the input. The attention mechanism assigns different weights to inputs at different positions by calculating the similarity between the query and the key, thereby capturing the correlation and dependency between inputs at different positions. After the attention output, the data is further nonlinearly transformed through a single-layer feedforward neural network to enrich the feature expression. Concatenate the last layer output of LSTM with the flattened output of CNN as the input of Encoder; The output of the encoder passes through a Softmax activation function to produce the final binary classification prediction fusion sequence self-attention network model.

6. The prediction method for post-transfection cell screening according to claim 5, characterized in that: The specific process of inputting the obtained feature data into the fusion sequence self-attention network model for prediction is as follows: The preprocessed data set is divided into training set, validation set and test set in a ratio of 6:2:2, and all random seeds are set to 202; The Encoder module of the fused sequence self-attention network model is used to train the training set, and the Adam optimizer and cross entropy loss function are used when training the training set. The accuracy, precision, recall and F1 score are used as evaluation indicators of the fused sequence self-attention network model to obtain the prediction results.

7. A post-transfection cell screening prediction device for implementing the post-transfection cell screening prediction method according to any one of claims 1 to 6, characterized in that: The prediction device comprises: The original data set is imported into the unit to obtain the original data set of cell screening; The original data set preprocessing unit preprocesses the original data set to obtain feature data; The cell screening prediction unit inputs the preprocessed feature data into the fusion prediction model for prediction, obtains the prediction results, and exports the prediction result file.

8. An electronic device, characterized in that: including interconnected processors and memories; The memory is used to store a computer program that supports the electronic device to execute the prediction method for cell screening, and the computer program includes program instructions; The processor is configured to call the program instructions to execute the prediction method for cell screening as described in any one of claims 1-6.

9. A storage medium, characterized in that: The storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor, the processor executes the prediction method for cell screening according to any one of claims 1 to 6.

Images