Deep learning-based gene sequencing result visualization method, product, device and medium

Abstract

Provided are a deep learning-based gene sequencing result visualization method, a product, a device and a medium. The method comprises: acquiring fluorescence images under test which are acquired for a sequencing chip and contain base signal acquisition units; using said fluorescence images as inputs of a base calling model trained on the basis of a deep learning model and, by means of the base calling model, performing feature extraction and recognition on said fluorescence images to output base calling results corresponding to each of said fluorescence images; sampling said fluorescence images to determine a reference fluorescence image; and, by means of a preset base calling algorithm, performing brightness information extraction and processing on the reference fluorescence image to output sequencing metric information corresponding to the reference fluorescence image. The sequencing metric information can be used for representing the degree of reliability of the base calling results of the current base calling model, thereby achieving the purpose of visually displaying deep learning-based gene sequencing results.

Description

Methods, products, equipment and media for visualizing gene sequencing results based on deep learning

[0001] The present invention claims priority to Chinese patent application number 202410379233.0, filed with the Patent Office of China on March 29, 2024, entitled “Methods, products, devices and media for visualizing results of gene sequencing based on deep learning”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] The present application relates to the field of gene sequencing technology, and in particular to a method for visualizing gene sequencing results based on deep learning, a computer program product, a gene sequencing device, and a computer-readable storage medium. Background Art

[0003] Gene sequencing refers to the process of identifying and analyzing genetic samples at the microscopic level through certain technical means.

[0004] A second-generation sequencer is a widely used instrument for genome sequencing, capable of quickly and accurately determining DNA sequences. The entire sequencing process primarily involves acquiring a fluorescent image of the sample through an optical system, performing a basecalling process, and obtaining and evaluating the sequencing results. The basecalling process (base recognition algorithm) uses the brightness information of the sample fluorescence image generated by laser light (sequencing signals corresponding to different base classes) to determine whether the sample belongs to one of the four base classes: A, C, G, or T. After basecalling, the sample information is converted from a digital image into sequence information, representing the sample's DNA sequence, for subsequent analysis and evaluation.

[0005] Base-calling algorithms are crucial for the accuracy and quality of base sequence sequencing. Due to its powerful fitting capabilities, deep learning has been widely applied to base call analysis in second-generation sequencers. While deep learning solutions can significantly improve base call accuracy, the processing of deep learning models is often uninterpretable. Sequencing metrics are crucial for subsequent troubleshooting and experimental improvement during gene sequencing, and thus, uninterpretable sequencing results significantly hinder the widespread application of deep learning models in the field. Summary of the Invention

[0006] The present application provides a method for visualizing gene sequencing results based on deep learning, a computer program product, a gene sequencing device, and a computer-readable storage medium that can visualize the gene sequencing results of deep learning.

[0007] In a first aspect of an embodiment of the present application, a method for visualizing gene sequencing results based on deep learning is provided, comprising:

[0008] Acquire a fluorescence image to be tested including a base signal acquisition unit collected from a sequencing chip;

[0009] The fluorescent images to be tested are used as input to a base recognition model obtained after training based on a deep learning model, and the base recognition model is used to extract and recognize features of the fluorescent images to be tested, and a base recognition result corresponding to each of the fluorescent images to be tested is output; wherein the base recognition result includes an identification result of the base type corresponding to the position of each base signal acquisition unit;

[0010] Determining a reference fluorescent image based on sampling the fluorescent image to be measured;

[0011] The brightness information of the reference fluorescence image is extracted and processed using a preset base recognition algorithm, and sequencing index information corresponding to the reference fluorescence image is output as visual reference information for the base recognition model; wherein the sequencing index information includes at least one of the following: information on the number of fluorescent bright spots, information on the brightness values ​​of fluorescent bright spots, information on the crosstalk ratio between channels corresponding to different base types, and information on the crosstalk ratio between cycles.

[0012] In a second aspect, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of the method for visualizing gene sequencing results based on deep learning as described in any embodiment of the present application.

[0013] In a third aspect, a gene sequencing device is provided, comprising a processor and a memory connected to the processor, wherein the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, the steps of the method for visualizing gene sequencing results based on deep learning as described in any embodiment of the present application are implemented.

[0014] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for visualizing the results of deep learning-based gene sequencing described in any embodiment of the present application are implemented.

[0015] In the above embodiment, in the gene sequencing process, a base recognition model trained based on a deep learning model is used to perform base recognition on the collected fluorescence image to be tested. While obtaining the base recognition result, the fluorescence image to be tested is sampled and selected as a reference fluorescence image. The reference fluorescence image is processed by a preset base recognition algorithm to output specified sequencing index information. These sequencing index information can be used to characterize the reliability of the base recognition result of the current base recognition model and serve as a guide for evaluating the quality of the sequencing result of the base recognition model, thereby achieving the purpose of visually displaying the results of gene sequencing based on deep learning. In addition, the sequencing index information can timely reflect the sequencing quality deviation of the base recognition model and facilitate the investigation of the causes of poor quality when the sequencing quality is poor, so that the sequencing results of the base recognition model can be controlled.

[0016] In the above embodiments, the computer program product, gene sequencing device, and computer-readable storage medium belong to the same concept as the corresponding embodiment of the gene sequencing result visualization method based on deep learning, and thus have the same technical effects as the corresponding embodiment of the gene sequencing result visualization method based on deep learning, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG1 is a flow chart of a method for visualizing gene sequencing results based on deep learning in one embodiment.

[0018] FIG2 is a schematic diagram of sequencing index information (the number of fluorescent bright spots corresponding to each base type) in one embodiment.

[0019] FIG3 is a schematic diagram of sequencing index information (brightness values ​​of fluorescent bright spots corresponding to each base type) in one embodiment.

[0020] FIG4 is a schematic diagram of sequencing index information (concentration of template spots corresponding to each base type) in one embodiment.

[0021] FIG5 is a schematic diagram of sequencing index information (concentration trends corresponding to each base type) in one embodiment.

[0022] FIG6 is a schematic diagram of sequencing index information (inter-cycle interference early reaction rate of each base type) in one embodiment.

[0023] FIG7 is a schematic diagram of sequencing index information (inter-cycle interference hysteresis reaction rate of each base type) in one embodiment.

[0024] FIG8 is a schematic diagram of sequencing index information (inter-channel crosstalk between different base types) in one embodiment.

[0025] FIG9 is a flowchart of a method for visualizing gene sequencing results based on deep learning in another embodiment.

[0026] FIG10 is a schematic diagram of the structure of a device for visualizing gene sequencing results based on deep learning in one embodiment.

[0027] FIG11 is a schematic diagram of the structure of a gene sequencing device in one embodiment. DETAILED DESCRIPTION

[0028] The technical solution of the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0029] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0030] In the following description, the expression "some embodiments" is involved, which describes a subset of all possible embodiments. It should be noted that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.

[0031] In the following description, the terms "first, second, and third" are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understandable that "first, second, and third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0032] In the field of gene sequencing technology, based on the powerful fitting ability of deep learning models, more and more studies are applying deep learning to base recognition in gene sequencers, using various deep learning models to train base recognition models to predict base categories. In order to solve the application obstacles brought about by the unexplainability of the processing process of deep learning models, the inventors of this application summarized and analyzed the currently known methods of applying deep learning to gene sequencing, and mainly divided the input of base classification of base recognition models into the following three categories: fluorescence images collected in real time during the sequencing process, fluorescence brightness information extracted from fluorescence images (fluorescence brightness information of uncorrected fluorescence images), and fluorescence brightness information extracted from fluorescence images and corrected by optical and biochemical factors (fluorescence brightness information of corrected fluorescence images). These different types of inputs to the base recognition model are essentially based on the fluorescence brightness information of the base signal acquisition unit contained in the fluorescence image collected in real time during the sequencing process. The base recognition model identifies the base category of each base signal acquisition unit in different sequencing cycles through feature extraction and classification prediction. In the embodiment of the present application, the different types of inputs to the base recognition model: fluorescence image, fluorescence brightness information of uncorrected fluorescence image, and fluorescence brightness information of corrected fluorescence image are all classified within the data range of the fluorescence image to be tested containing the base signal acquisition unit.

[0033] Please refer to FIG1 , which shows a method for visualizing gene sequencing results based on deep learning according to an embodiment of the present application, including the following steps:

[0034] S101, obtaining a fluorescence image to be tested including a base signal acquisition unit collected from a sequencing chip.

[0035] The gene sequencing process is to amplify the DNA molecules to be tested through specific amplification technology, amplify each DNA fragment (single-stranded library molecule) to form a base cluster, use the base cluster detection results to construct the template points of the base cluster on the sequencing chip, and then perform base recognition and other operations through the base cluster template points to obtain the base sequence. It can be understood that the method for visualizing gene sequencing results based on deep learning provided in the embodiments of the present application obtains base recognition results by extracting and recognizing features of the fluorescent image to be tested based on the base recognition model obtained after training the deep learning model. This is a strategy for improving base recognition accuracy by using the deep learning model training strategy. The input of the base recognition model is a fluorescent image obtained based on the base clusters after amplification of the single-stranded library molecules on the sequencing chip, which performs base cluster positioning detection and base type identification. Here, each base cluster refers to a base signal acquisition unit, so it is not limited to the amplification technology used for the single-stranded library molecules. That is, the base type identification of the base signal acquisition unit for the sequencing chip in other massively parallel sequencing technologies is also applicable. For example, the base signal acquisition unit can refer to the base cluster obtained by the bridge amplification technology in the Illumina sequencing technology, and also includes the nanospheres obtained by the rolling circle amplification technology (RCA), etc. This application does not impose any restrictions on this.

[0036] It should be noted that the fluorescence image to be measured comprising the base signal acquisition unit may be a raw fluorescence image directly, or may be a fluorescence image that has undergone image preprocessing such as noise reduction and optical interference removal.

[0037] S103, using the fluorescent image to be tested as the input of a base recognition model obtained after training based on a deep learning model, performing feature extraction and recognition on the fluorescent image to be tested through the base recognition model, and outputting the base recognition results corresponding to each of the fluorescent images to be tested; wherein, the base recognition results include the recognition results of the base types corresponding to the positions of the base signal acquisition units.

[0038] Among them, the deep learning model can be various known neural network models.

[0039] The base recognition results can be presented in different forms, and the form of the recognition results of the base signal acquisition unit can be a data matrix that identifies the base type of each base signal acquisition unit in the current cycle, or an image that identifies the base type of each base signal acquisition unit. In an optional example, the base recognition result output by the base recognition model is a multi-channel output corresponding to different base types, and each channel output corresponds to the recognition result of a base signal acquisition unit of a base type. For example, different base types include four types: A, C, G, and T. Channel 1 can be a coordinate data matrix of the position information of the center of the base signal acquisition unit of the A base type, so that the coordinate data matrix output by channel 1 represents the recognition result of the base signal acquisition unit of the A base type in the current cycle; similarly, the coordinate data matrix of channel 2 corresponds to the recognition result of the base signal acquisition unit of the C base type, the coordinate data matrix of channel 3 corresponds to the recognition result of the base signal acquisition unit of the G base type, and the coordinate data matrix of channel 4 corresponds to the recognition result of the base signal acquisition unit of the T base type. In another optional example, the base recognition result of the base recognition model is a single-channel output of the recognition results of the base signal acquisition unit containing multiple base types in the same cycle, such as the recognition results of A, C, G, and T obtained by channels 1, 2, 3, and 4, forming a coordinate data matrix with base type labels marked at the corresponding positions of the center of the base signal acquisition unit containing all base types in the current cycle. It should be noted that although the output of the base recognition model is a coordinate data matrix containing the center of the base signal acquisition unit, it expresses the recognition of the base types belonging to different base signal acquisition units in the current cycle, and what is achieved is base type recognition.

[0040] The above-mentioned coordinate data matrix can adopt other forms that can characterize the base type at the center of each base signal acquisition unit, such as a probability data matrix indicating whether the pixel point is of a certain base type at the position of the center of the base signal acquisition unit. The probability value at the position of the center of the base signal acquisition unit represents the probability that the base signal acquisition unit belongs to the A, C, G or T base type.

[0041] Other forms of characterizing the base type at each base signal acquisition unit can also be image forms, such as directly outputting the position of the center of the base signal acquisition unit of the A, C, G, and T base types obtained based on the coordinate data matrix and the probability data matrix, and directly outputting the fluorescent image marked with the base type label at the center of each base signal acquisition unit in the current cycle.

[0042] According to the various possible presentation forms of the base recognition results provided above, it can be seen that the base recognition result output by the base recognition model is a base recognition result obtained after the base recognition model processes the multiple fluorescent images to be tested collected in the current cycle, which can know the base recognition result corresponding to the position of each base signal acquisition unit in the current cycle. It may not be limited to a specific form and is not restricted here.

[0043] The base recognition model based on deep learning has powerful nonlinear modeling capabilities, automatic feature extraction capabilities and self-learning capabilities. The output of different base recognition results can be obtained by training the deep learning model with training samples with corresponding labels.

[0044] S105 , determining a reference fluorescent image based on sampling the fluorescent image to be measured.

[0045] In the gene sequencing process, the base recognition model uses the real-time acquired fluorescence image to be tested as input to perform base recognition and output the corresponding base recognition result. At the same time, a portion of the images from the fluorescence image to be tested, which is used as the input of the base recognition model, are randomly sampled as reference fluorescence images. The sampling ratio can be different at different stages of the gene sequencing process. In an optional example, the sampling ratio can be increased in the first few cycles of the gene sequencing process so that the number of reference fluorescence images can be increased in the first few cycles of the gene sequencing process to obtain sequencing index information. The sequencing index information is used as an explanatory indicator of the current base recognition result obtained by the base recognition model, which can assist in confirming the sequencing quality of the base recognition model, guide sequencing experiments, and troubleshoot the causes that affect sequencing quality, thereby visualizing the results of the deep learning-based base recognition model.

[0046] S107, extracting and processing brightness information of the reference fluorescence image using a preset base recognition algorithm, and outputting sequencing index information corresponding to the reference fluorescence image as visual reference information for the base recognition model; wherein the sequencing index information includes at least one of the following: information on the number of fluorescent bright spots, information on the brightness values ​​of fluorescent bright spots, information on the ratio of crosstalk between channels corresponding to different base types, and information on the ratio of crosstalk between cycles.

[0047] The preset base recognition algorithm can refer to a known traditional base recognition algorithm. Traditional base recognition algorithms process reference fluorescence images. Compared to deep learning-based base recognition models, traditional base recognition algorithms provide more stable and reliable base recognition quality and facilitate obtaining the desired type of calculation results through algorithmic formula settings. However, they are computationally intensive and have low recognition efficiency, which can easily affect gene sequencing efficiency. In this embodiment, a traditional base recognition algorithm is used to extract brightness information from sampled reference fluorescence images and process and output corresponding sequencing index information. This sequencing index information is used as visual reference information for base recognition quality assessment using the deep learning-based base recognition model.

[0048] Sequencing indicator information is associated with factors that characterize the accuracy of the base recognition results, such as the number, location, or base type of base signal acquisition units in the base recognition results. The factors that characterize the accuracy of the recognition results contained in the sequencing indicator information can be used to understand the current situation of base recognition accuracy and quality. For example, the number of fluorescent bright spots reflected in the sequencing indicator information can be used to know the number of effective base reactions in gene sequencing. If the number of effective base reactions is too low, it can be timely determined that the base recognition results of the deep learning-based base recognition model are inaccurate. The cause can be identified through the sequencing indicator and timely targeted solutions can be implemented. In this embodiment, the sequencing indicator information can be one or more of the following: the number of fluorescent bright spots, the brightness value of the fluorescent bright spots, the crosstalk ratio between channels corresponding to different base types, and the crosstalk ratio between cycles.

[0049] In the above embodiment, in the gene sequencing process, a base recognition model trained based on a deep learning model is used to perform base recognition on the collected fluorescent image to be tested. While obtaining the base recognition result, the fluorescent image to be tested is sampled and selected as a reference fluorescent image. The reference fluorescent image is processed by a preset base recognition algorithm to output specified sequencing index information. These sequencing index information can be used to characterize the reliability of the base recognition result of the current base recognition model and serve as a guide for evaluating the quality of the sequencing result of the base recognition model, thereby achieving the purpose of visually displaying the results of gene sequencing based on deep learning. In addition, the sequencing index information can timely reflect the sequencing quality of the base recognition model based on deep learning, and facilitates the investigation of the causes of poor quality when the sequencing quality is poor, so that the sequencing results of the base recognition model can be controlled.

[0050] In some embodiments, in step S107, outputting sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes:

[0051] The number of fluorescent bright spots in the reference fluorescent image is determined, and visual reference information representing the number of valid base signal acquisition units in the current sequencing process is obtained based on the sum of the numbers of fluorescent bright spots in multiple reference fluorescent images corresponding to different base types in the same cycle.

[0052] In this embodiment, the sequencing index information refers to the number of fluorescent bright spots. It should be noted that the information on the number of fluorescent bright spots can be displayed in different forms such as numerical values, images, and a combination of numerical values ​​and images. The reference fluorescent image is processed by a traditional base recognition algorithm to find the number of fluorescent bright spots in the reference fluorescent image. The number of fluorescent bright spots corresponding to the reference fluorescent images corresponding to the A, C, G, and T base types in the same cycle is used to obtain the number of valid base signal acquisition units on the sequencing chip in the current sequencing process. In this embodiment, the base signal acquisition unit specifically refers to the base cluster. The number of fluorescent bright spots can reflect the number of current valid base clusters on the chip. This indicator is also affected by factors such as image focus, balance, impurities, and background noise. By outputting the sequencing index information of the number of fluorescent bright spots, it is used as one of the explanatory indicators of the current base recognition result obtained by the base recognition model. It should be noted that the application interface for executing the deep learning-based gene sequencing result visualization method described in the embodiment of the present application includes a display area for sequencing index information. As shown in Figure 2, the fluorescence images collected in each cycle determined by sampling are processed by a traditional base recognition algorithm to obtain the number of fluorescent bright spots in the fluorescence images of different base types in each cycle, and a curve graph of the sequencing index information, namely the number of fluorescent bright spots (Peak Number) corresponding to base types A, C, G, and T, is obtained.

[0053] In some embodiments, in step S107, outputting sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes:

[0054] The brightness of the fluorescent bright spot in the reference fluorescent image is determined, and based on the average brightness of the fluorescent bright spot in multiple reference fluorescent images corresponding to different base types in the same cycle, visual reference information representing the signal intensity of the effective base signal acquisition unit in the current sequencing process is obtained.

[0055] In this embodiment, sequencing indicator information refers to the brightness of fluorescent bright spots. It should be noted that the brightness information of fluorescent bright spots can be displayed in various forms, such as numerical values, images, or a combination of numerical values ​​and images. A reference fluorescent image is processed using a traditional base calling algorithm to identify fluorescent bright spots within the reference fluorescent image and extract the brightness values ​​of the corresponding locations. The brightness values ​​of the fluorescent bright spots corresponding to the A, C, G, and T base types within the same cycle are used to determine the signal intensity of the effective base signal acquisition units on the sequencing chip during the current sequencing process. In this embodiment, the base signal acquisition units specifically refer to base clusters. In an optional specific example, the brightness value of the fluorescent bright spot refers to the grayscale value at the fluorescent bright spot location extracted from the corresponding reference fluorescent image using a traditional base calling algorithm. The average brightness of the fluorescent bright spot is calculated based on the grayscale values ​​of all fluorescent bright spot locations in the corresponding reference fluorescent image. Fluorescent bright spot brightness is affected by the sequencing signal laser intensity during the gene sequencing process. A more stable brightness value during the gene sequencing process generally indicates better sequencing result quality. Thus, the output of the fluorescent bright spot brightness value, as a sequencing indicator, serves as one of the interpretative indicators for the current base call results obtained by the base calling model. It should be noted that the application interface for executing the deep learning-based gene sequencing result visualization method described in the embodiment of the present application includes a display area for sequencing index information. As shown in Figure 3, the fluorescence images collected in each sampled cycle are processed by a traditional base recognition algorithm to obtain the average brightness of the fluorescent bright spots in the fluorescence images of different base types (A, C, G, T) in each cycle, and a curve graph of the sequencing index information, namely the fluorescence bright spot brightness (Raw Ave Intensity) corresponding to the base types A, C, G, and T, is obtained.

[0056] In some embodiments, in step S107, outputting sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes:

[0057] The position of the fluorescent bright spot in the reference fluorescent image is determined, and the average brightness of the fluorescent bright spot of the specified base type in the corresponding channel in the same cycle is calculated to obtain visual reference information representing the trend of the number of effective reaction cycles of the base in the current sequencing process.

[0058] In this embodiment, the sequencing index information refers to the brightness of the identified base represented by the average brightness of the fluorescent bright spots corresponding to different base types in their corresponding channels. It should be noted that the brightness of the identified base can be displayed in different forms such as numerical values, images, and a combination of numerical values ​​and images. The reference fluorescence image is processed by a traditional base recognition algorithm to find the positions of all fluorescent bright spots of the same base type in the reference fluorescence image. Taking channel 1 corresponding to base type A as an example, in the reference fluorescence image of channel 1, the positions of all base types A are counted to determine the template point position of base type A, and the brightness values ​​of the fluorescent bright spots of all base types A before normalization in channel 1 are extracted, and then the average value is calculated to represent the identification base brightness (Base Intensity) of base type A. Here, the brightness of the fluorescent bright spot used in the calculation of the base recognition brightness is corrected for channel crosstalk but not for cycle crosstalk (phasing / prephasing). It can be used to reflect the trend of base reaction efficiency. Generally, as the gene sequencing process progresses, the longer the reaction, the more severe the damage to the base cluster (that is, a base cluster undergoes multiple rounds of reaction and copying, and damage occurs as the number of reaction rounds increases). The faster the brightness of the recognized base decreases, the shorter the measurable read length, which can be used as a reference for sequencing estimation. By outputting the average brightness of the fluorescent bright spot of a specified base type in the corresponding channel, the recognized base brightness of the corresponding base type is represented, thereby correspondingly indicating the sequencing indicator information of the number of effective base reaction rounds in the current cycle, which serves as one of the explanatory indicators for the current base recognition result obtained by the base recognition model. It should be noted that the application interface for executing the deep learning-based gene sequencing result visualization method described in the embodiment of the present application includes a display area for sequencing index information. As shown in Figure 4, the fluorescence images collected in each sampled cycle are processed by a traditional base recognition algorithm to obtain the average brightness of the fluorescent bright spot of the specified base type in the corresponding channel in each cycle. The curve graphs of the sequencing index information, namely the brightness of the identified bases corresponding to the base types A, C, G, and T, can be obtained respectively to characterize the trend of the number of effective reaction rounds of the bases in each cycle.

[0059] In some embodiments, in step S107, outputting sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes:

[0060] The fluorescence brightness position in the reference fluorescence image is determined, and the average brightness of the fluorescent bright spot of the specified base type in the corresponding channel within the same cycle is calculated. Based on the ratio of the average brightness of the fluorescent bright spot of the specified base type in two consecutive cycles, visual reference information characterizing the power impact of the sequencing signal laser in the current sequencing process is obtained.

[0061] The number of fluorescent bright spots essentially reflects the number of valid base signal acquisition units. In this embodiment, the base signal acquisition units specifically refer to base clusters, and the fluorescent bright spot brightness values ​​are the basic data used to determine the number of fluorescent bright spots. By processing the reference fluorescence image using a traditional base recognition algorithm, all fluorescent bright spot locations of the same base type in the reference fluorescence image are found. Taking channel 1 corresponding to base type A as an example, in the reference fluorescence image of channel 1, the locations of all base type A are counted to determine the template point location of base type A. The brightness values ​​of all fluorescent bright spots of base type A in channel 1 before normalization are extracted and averaged to represent the recognized base brightness (Base Intensity) of base type A. The ratio of the average brightness value of the fluorescent bright spots of base type A in the subsequent cycle to the average brightness value of the fluorescent bright spots of base type A in the previous cycle is used to represent the brightness gradient trend (Intensity Trend) of base type A. The brightness gradient trend can be used to analyze the degree of decline and trend of base reaction efficiency in gene sequencing from a percentage dimension. A flatter trend indicates a better sequencing experimental environment.

[0062] In one optional example, the first cycle of the two successive cycles refers to the first cycle in the gene sequencing process. The intensity gradient trend (Intensity Trend) = n_cycle Base Intensity / first cycle Base Intensity. This means that the intensity of the first cycle is used as a benchmark to calculate the decreasing trend in base effective reflection efficiency as the gene sequencing process progresses, thereby guiding the gene sequencing experiment. The sequencing indicator information, which is the changing trend of the ratio of the brightness averages of the fluorescent highlights of a specified base type in the corresponding channel, is output as one of the explanatory indicators for the current base recognition result obtained by the base recognition model. It should be noted that the application interface for executing the deep learning-based gene sequencing result visualization method described in the embodiment of the present application includes a display area for sequencing index information. As shown in Figure 5, the fluorescence images collected in each sampled cycle are processed by a traditional base recognition algorithm to obtain the average brightness of the fluorescence bright spot of the specified base type in the corresponding channel in each cycle. The ratio of the average brightness of the current cycle to the average brightness of the first cycle is calculated, and curve graphs of the sequencing index information, namely the brightness gradient trends corresponding to the base types A, C, G, and T, can be obtained respectively to characterize the power impact of the sequencing signal laser in the current sequencing process.

[0063] In some embodiments, in step S107, outputting sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes:

[0064] The brightness of the fluorescent bright spots in the reference fluorescent image is determined, and the brightness of the fluorescent bright spots corresponding to multiple reference fluorescent images corresponding to different base types in the same cycle is calculated. Based on the change in the brightness of the fluorescent bright spots of the same base type in multiple adjacent cycles, the advance reaction rate / lag reaction rate of each cycle is calculated to obtain visual reference information characterizing the degree of crosstalk interference between cycles in the current sequencing process.

[0065] The prephasing rate is used to record the lagging reaction rate, such as a base that should react in cycle N but reacts in cycle N+n. The phasing rate is used to record the prephasing rate, such as a base that should react in cycle N but reacts in cycle N-n. The prephasing rate / lagging reaction rate of each cycle can be calculated based on the changes in the brightness of the fluorescent bright spots of the same base type in multiple adjacent cycles. The Markov model can be used for statistical calculation. Through the cumulative effect, the prephasing rate / lagging reaction rate can be calculated by the changes in the brightness of the fluorescent bright spots corresponding to each base type in multiple successive cycles. In one optional embodiment, the lead / lag reaction rates for different base types, A, C, G, and T, are calculated for each cycle, using the two cycles before and after the current cycle as a reference. That is, for each base type, the lead / lag reaction rate is calculated based on the change in brightness of the corresponding fluorescent bright spot over five cycles. Multivariate linear regression is then used to perform multifactor analysis, yielding a multivariate linear regression factor to measure the lead / lag reaction rate over the entire sequencing experiment. The sequencing indicator information, the lead / lag reaction rate for each base type, is output and used as one of the explanatory indicators for the current base recognition result obtained by the base recognition model.

[0066] It should be noted that the application interface for the method for visualizing the results of gene sequencing based on deep learning as described in the embodiment of the present application includes a display area for sequencing index information. As shown in Figures 6 and 7, the fluorescence images collected in each cycle determined by sampling are processed by a traditional base recognition algorithm. The advance reaction rate / lag reaction rate of each cycle is calculated based on the change in the brightness of the fluorescent bright spots of the same base type in multiple adjacent cycles. The curve graphs of the sequencing index information of the advance reaction rate and lag reaction rate corresponding to base types A, C, G, and T can be obtained respectively. The advance reaction rate / lag reaction rate can represent the degree of crosstalk interference between cycles, so that the sequencing quality of the base recognition results output by the base recognition model based on deep learning can be visualized, and the degree of crosstalk interference between cycles can be used to guide gene sequencing experiments and troubleshoot problems.

[0067] In some embodiments, in step S107, outputting sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes:

[0068] The brightness of the fluorescent bright spots in the reference fluorescent image is determined, and the ratio of the fluorescent bright spot brightness in multiple reference fluorescent images corresponding to different base types in the same cycle is calculated according to the base type, thereby obtaining visual reference information representing the degree of crosstalk interference between channels in the current sequencing process.

[0069] Crosstalk between sequencing signal channels corresponding to different base types can occur between any two-way combinations of base types. For example, base types A, C, G, and T correspond to four channels, each with a total of 12 channel crosstalk combinations. For example, AC represents the crosstalk of A on C, and CA represents the crosstalk of C on A. Similarly, this value is primarily related to the wavelength of the sequencing signal laser and is generally considered a fixed or constant value. The inter-channel crosstalk interference for each base type combination is obtained by taking the difference in the fluorescence bright spot brightness between two base type combinations within the same cycle. Thus, by performing a linear fit based on the fluorescence bright spot intensities corresponding to the base types, the ratio of the fluorescence bright spot brightness to each base type combination is calculated and output as a sequencing indicator representing the degree of inter-channel crosstalk (Crosstalk). This represents the degree of inter-channel crosstalk interference in the current sequencing process and serves as one of the explanatory indicators for the current base call results obtained by the base call model.

[0070] It should be noted that the application interface for the method for visualizing the results of deep learning-based gene sequencing described in the embodiment of the present application includes a display area for sequencing index information. As shown in FIG8 , the fluorescence images collected in each cycle determined by sampling are processed by a traditional base recognition algorithm. The degree of inter-channel crosstalk interference of different base types in each cycle is calculated based on the ratio of the brightness of the fluorescent bright spots in the base type combination in the same cycle. As the gene sequencing process progresses, the change in the degree of inter-channel crosstalk calculated based on the brightness of the fluorescent bright spots in different cycles can be used to obtain a curve graph of sequencing index information of the degree of inter-channel crosstalk interference between different base type combinations at different stages of the gene sequencing process. Finding base type combinations with a large degree of inter-channel crosstalk interference allows the sequencing quality of the base recognition results output by the deep learning-based base recognition model to be visualized, and the degree of inter-channel crosstalk interference can be used to guide gene sequencing experiments and troubleshoot problems, providing a reference for optical and biochemical optimization in gene sequencing experiments.

[0071] In some embodiments, the method for visualizing gene sequencing results based on deep learning further includes:

[0072] Outputting a base recognition reference result through the preset base recognition algorithm;

[0073] determining a difference between a base call result of the reference fluorescent image and a base call result of the corresponding fluorescent image to be tested;

[0074] If the difference exceeds the threshold, the base recognition model using the uncorrected fluorescence brightness information of the fluorescent image to be tested as input is replaced to execute the subsequent sequencing process; or, the base recognition model using the corrected fluorescence brightness information of the fluorescent image to be tested is replaced to execute the subsequent sequencing process.

[0075] A base recognition reference result refers to a recognition result that is identical or corresponds to the base recognition result output by a deep learning-based base recognition model. "Identical to the base recognition result" means that the base recognition result output by the base recognition model is presented in the same format. For example, if the base recognition result is a coordinate data matrix containing base type labels at corresponding positions in the center of a base signal acquisition unit that includes all base types, the base recognition reference result is also a coordinate data matrix containing base type labels at corresponding positions in the center of a base signal acquisition unit that includes all base types. "Corresponding to the base recognition result" means that the base recognition result is presented in a different format than the base recognition result output by the base recognition model, but also contains recognition result information that can determine the base type of each base signal acquisition unit in the current cycle.

[0076] Please refer to FIG9 , which is a flowchart of a method for visualizing gene sequencing results based on deep learning provided in an embodiment of the present application, comprising the following steps:

[0077] S21, the gene sequencer acquires fluorescence images in real time.

[0078] S22, using a base recognition model trained based on a deep learning model, performs feature extraction and recognition on the fluorescence image to obtain a base recognition result.

[0079] S23, output the base recognition result.

[0080] S24: Determine whether the fluorescence image is the reference fluorescence image determined by sampling. If so, execute S25.

[0081] S25, the reference fluorescence image is processed using a traditional base recognition algorithm to obtain a base recognition reference result and various key sequencing indicators under the traditional algorithm.

[0082] S26, outputs the base recognition reference results and key sequencing indicators under the traditional algorithm.

[0083] In the gene sequencing process, a base recognition model trained based on a deep learning model is used to perform base recognition on the collected fluorescent image to be tested. While obtaining the base recognition result, the fluorescent image to be tested is sampled and selected as a reference fluorescent image. The reference fluorescent image is processed by a preset base recognition algorithm to output a base recognition reference result. The base recognition reference result obtained by the traditional base recognition algorithm is compared with the base recognition result obtained by the base recognition model based on deep learning to determine the difference between the base recognition reference result of the reference fluorescent image and the base recognition result of the corresponding fluorescent image to be tested. This difference can assist in identifying the deviation of the recognition result of the base recognition model based on deep learning, thereby characterizing the reliability of the base recognition result of the current base recognition model and providing guidance for the quality evaluation of the sequencing results of the base recognition model. In this way, the traditional base recognition algorithm is retained, and the base recognition results are obtained by processing the fluorescent images through a preset algorithm formula, which can obtain the advantages of being relatively more stable, reliable and controllable. At the same time, the required key sequencing indicators and base recognition results used as a reference can be obtained at a relatively small cost (consumption of computing resources). Through these key sequencing indicators, sample quality problems, experimental environment problems, and generalization ability problems of deep learning models that may arise in the gene sequencing process can be reflected in a timely manner to guide gene sequencing experiments and troubleshoot problems.

[0084] It should be noted that under the technical guidance proposed by the inventors of the present application of processing the sampled reference fluorescence images in parallel based on the traditional base recognition algorithm to output sequencing index information, the quantity, type, display form, etc. of the sequencing index information can be subject to various changes.

[0085] In some embodiments, the sequencing index information also includes image quality index information of the fluorescent image to be tested, background brightness information of the fluorescent image to be tested, and a bright spot brightness distribution histogram of the fluorescent image to be tested for each cycle. The image quality index information can be various known parameter types capable of evaluating image quality, such as brightness uniformity and resolution. The image quality index information, image background brightness information, and bright spot brightness distribution histogram can be obtained by combining image processing algorithms known in the field of image processing technology, and are not further described here.

[0086] It should be noted that, in the embodiments of the present application, sequencing index information is not limited to being obtained by extracting and processing brightness information of a reference fluorescent image through a traditional base recognition algorithm, but may also include various index information that can characterize the environment and conditions of a gene sequencing experiment.

[0087] In some embodiments, the method for visualizing gene sequencing results based on deep learning further includes:

[0088] Obtain operating performance parameters of the sequencing device, and output other sequencing indicator information based on the operating performance parameters; the other sequencing indicator information includes chip rotation angle and chip offset.

[0089] The sequencing device may specifically refer to a gene sequencer. Using the operating performance parameters of the sequencing device as other sequencing indicator information can be used to troubleshoot problems with the test environment and conditions of the gene sequencing process. In this embodiment, the other sequencing indicator information includes chip rotation angle and chip offset.

[0090] It should be noted that the application interface for executing the method for visualizing gene sequencing results based on deep learning described in the embodiment of the present application may display a plurality of operation buttons to be selected corresponding to each of the sequencing indicator information. The method for visualizing gene sequencing results based on deep learning may further include:

[0091] An instruction for selecting sequencing index information of a target type is obtained, and detailed information of the sequencing index information of the target type is displayed in a display area of ​​the sequencing index information according to the instruction.

[0092] In this embodiment, in the gene sequencing process, while a deep learning-based base recognition model is used to identify and output the corresponding base recognition results, the application interface displays various sequencing indicator information output by processing the sampled reference fluorescence image determined by the traditional base recognition algorithm, providing the user with the sequencing indicator information they want to view, so as to assist in real-time understanding of the reliability of the base recognition results of the current base recognition model.

[0093] On the other hand, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for visualizing gene sequencing results based on deep learning as described in any embodiment of the present application.

[0094] Among them, in the computer program product, an optional implementation form of the program module architecture of the computer program that implements each step of the method is a result visualization device for gene sequencing based on deep learning. Please refer to Figure 10. The result visualization device for gene sequencing based on deep learning includes: an acquisition module 311, which is used to obtain a fluorescent image to be tested that includes a base signal acquisition unit collected for the sequencing chip. A base recognition model 312 based on deep learning is used to take the fluorescent image to be tested as input, extract and recognize the features of the fluorescent image to be tested through the base recognition model, and output the base recognition results corresponding to each of the fluorescent images to be tested; wherein the base recognition results include the recognition results of the base types corresponding to the positions of each base signal acquisition unit. A sampling module 313 is used to determine a reference fluorescent image based on sampling the fluorescent image to be tested. The sequencing index module 314 is used to extract and process the brightness information of the reference fluorescence image using a preset base recognition algorithm, and output sequencing index information corresponding to the reference fluorescence image as visual reference information for the base recognition model; wherein the sequencing index information includes at least one of the following: information on the number of fluorescent bright spots, information on the brightness values ​​of fluorescent bright spots, information on the crosstalk ratio between channels corresponding to different base types, and information on the crosstalk ratio between cycles.

[0095] Optionally, the sequencing index module 314 is specifically used to determine the number of fluorescent bright spots in the reference fluorescence image, and obtain visual reference information representing the number of valid base signal acquisition units in the current sequencing process based on the sum of the number of fluorescent bright spots in multiple reference fluorescence images corresponding to different base types in the same cycle.

[0096] Optionally, the sequencing index module 314 is specifically used to determine the brightness of the fluorescent bright spots in the reference fluorescent image, and obtain visual reference information representing the signal intensity of the effective base signal acquisition unit in the current sequencing process based on the average of the brightness of the fluorescent bright spots in multiple reference fluorescent images corresponding to different base types in the same cycle.

[0097] Optionally, the sequencing index module 314 is specifically used to determine the position of the fluorescent bright spot in the reference fluorescent image, calculate the average brightness of the fluorescent bright spot of the specified base type in the corresponding channel in the same cycle, and obtain visual reference information that characterizes the trend of the number of effective reaction rounds of the base in the current sequencing process.

[0098] Optionally, the sequencing index module 314 is specifically used to determine the fluorescence brightness position in the reference fluorescence image, calculate the average brightness of the fluorescence bright spot of the specified base type in the corresponding channel in the same cycle, and obtain visual reference information characterizing the power impact of the sequencing signal laser in the current sequencing process based on the ratio of the average brightness of the fluorescence bright spot of the specified base type in two consecutive cycles.

[0099] Optionally, the sequencing index module 314 is specifically used to determine the brightness of the fluorescent bright spots in the reference fluorescent image, calculate the brightness of the fluorescent bright spots corresponding to multiple reference fluorescent images corresponding to different base types in the same cycle, and calculate the advance reaction rate / lag reaction rate of each cycle based on the change in the brightness of the fluorescent bright spots of the same base type in multiple adjacent cycles, so as to obtain visual reference information characterizing the degree of crosstalk interference between cycles in the current sequencing process.

[0100] Optionally, the sequencing index module 314 is specifically used to determine the brightness of the fluorescent bright spots in the reference fluorescent image, and calculate the ratio of the brightness of the fluorescent bright spots in multiple reference fluorescent images corresponding to different base types in the same cycle according to the base types, so as to obtain visual reference information characterizing the degree of crosstalk interference between channels in the current sequencing process.

[0101] Optionally, the sequencing index module 314 is also used to output a base recognition reference result through the preset base recognition algorithm; determine the difference between the base recognition reference result of the reference fluorescence image and the base recognition result of the corresponding fluorescence image to be tested; if the difference exceeds a threshold, replace the base recognition model with the uncorrected fluorescence brightness information of the fluorescence image to be tested as input to execute the subsequent sequencing process; or, replace the base recognition model with the corrected fluorescence brightness information of the fluorescence image to be tested as input to execute the subsequent sequencing process.

[0102] Optionally, the sequencing index information further includes image quality index information of the fluorescent image to be tested in each cycle, background brightness information of the fluorescent image to be tested, and a bright spot brightness distribution histogram of the fluorescent image to be tested.

[0103] Optionally, the sequencing index module 314 is further configured to obtain operating performance parameters of a sequencing device and output other sequencing index information based on the operating performance parameters; the other sequencing index information includes a chip rotation angle and a chip offset.

[0104] It should be noted that: the result visualization device of gene sequencing based on deep learning provided in the above embodiment uses a traditional base recognition algorithm to output sequencing index information to assist in visualizing the gene sequencing results of the deep learning model while performing gene sequencing through a deep learning model. The division of the above program modules is only used as an example to illustrate. In actual applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the method steps described above. In addition, the result visualization device of gene sequencing based on deep learning provided in the above embodiment and the result visualization method embodiment of gene sequencing based on deep learning belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0105] On the other hand, the present application also provides a gene sequencing device. Please refer to Figure 11, which is an optional hardware structure diagram of a gene sequencing device. The gene sequencing device includes a processor 212 and a memory 211 connected to the processor 212. The memory 211 stores a computer program for implementing the method for visualizing the results of gene sequencing based on deep learning provided in any embodiment of the present application, so that when the corresponding computer program is executed by the processor, the steps of the method for visualizing the results of gene sequencing based on deep learning provided in any embodiment of the present application are implemented. The gene sequencing device loaded with the corresponding computer program has the same technical effect as the corresponding method embodiment. To avoid repetition, it will not be repeated here.

[0106] On the other hand, the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, each process of the above-mentioned embodiment of the method for visualizing the results of deep learning-based gene sequencing is implemented, and the same technical effect is achieved. To avoid repetition, it is not described here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0107] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0108] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for a terminal (which can be a mobile phone, computer, server, spatial genomics sequencing platform, gene sequencer, or network equipment, etc.) to execute the methods described in each embodiment of the present invention.

[0109] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for visualizing gene sequencing results based on deep learning, characterized in that: include: Acquire a fluorescence image to be tested including a base signal acquisition unit collected from a sequencing chip; The fluorescent images to be tested are used as input to a base recognition model obtained after training based on a deep learning model, and the base recognition model is used to extract and recognize features of the fluorescent images to be tested, and a base recognition result corresponding to each of the fluorescent images to be tested is output; wherein the base recognition result includes an identification result of the base type corresponding to the position of each base signal acquisition unit; Determining a reference fluorescent image based on sampling the fluorescent image to be measured; The brightness information of the reference fluorescence image is extracted and processed using a preset base recognition algorithm, and sequencing index information corresponding to the reference fluorescence image is output as visual reference information for the base recognition model; wherein the sequencing index information includes at least one of the following: information on the number of fluorescent bright spots, information on the brightness values ​​of fluorescent bright spots, information on the crosstalk ratio between channels corresponding to different base types, and information on the crosstalk ratio between cycles.

2. The method for visualizing gene sequencing results based on deep learning according to claim 1, wherein: The sequencing index information includes information on the number of fluorescent bright spots; and the outputting of the sequencing index information corresponding to the reference fluorescent image as visual reference information for the base recognition model includes: The number of fluorescent bright spots in the reference fluorescent image is determined, and visual reference information representing the number of valid base signal acquisition units in the current sequencing process is obtained based on the sum of the numbers of fluorescent bright spots in multiple reference fluorescent images corresponding to different base types in the same cycle.

3. The method for visualizing gene sequencing results based on deep learning according to claim 1, wherein: The sequencing index information includes fluorescent bright spot brightness value information; the outputting the sequencing index information corresponding to the reference fluorescent image as visual reference information of the base recognition model includes: The brightness of the fluorescent bright spot in the reference fluorescent image is determined, and based on the average brightness of the fluorescent bright spot in multiple reference fluorescent images corresponding to different base types in the same cycle, visual reference information representing the signal intensity of the effective base signal acquisition unit in the current sequencing process is obtained.

4. The method for visualizing gene sequencing results based on deep learning according to claim 1, wherein: The sequencing index information includes the brightness of the identified base represented by the average brightness of the fluorescent bright spots corresponding to different base types in their corresponding channels; the output of the sequencing index information corresponding to the reference fluorescent image as the visual reference information of the base recognition model includes: The position of the fluorescent bright spot in the reference fluorescent image is determined, and the average brightness of the fluorescent bright spot of the specified base type in the corresponding channel in the same cycle is calculated to obtain visual reference information representing the trend of the number of effective reaction cycles of the base in the current sequencing process.

5. The method for visualizing gene sequencing results based on deep learning according to claim 1, wherein: The sequencing index information includes the number of valid base signal acquisition units; the outputting of the sequencing index information corresponding to the reference fluorescent image as visual reference information of the base recognition model includes: The fluorescence brightness position in the reference fluorescence image is determined, and the average brightness of the fluorescent bright spot of the specified base type in the corresponding channel within the same cycle is calculated. Based on the ratio of the average brightness of the fluorescent bright spot of the specified base type in two consecutive cycles, visual reference information characterizing the power impact of the sequencing signal laser in the current sequencing process is obtained.

6. The method for visualizing gene sequencing results based on deep learning according to claim 1, wherein: The sequencing index information includes inter-channel crosstalk ratio information of channels corresponding to different base types; the outputting of the sequencing index information corresponding to the reference fluorescence image as visual reference information of the base recognition model includes: The brightness of the fluorescent bright spots in the reference fluorescent image is determined, and the brightness of the fluorescent bright spots corresponding to multiple reference fluorescent images corresponding to different base types in the same cycle is calculated. Based on the change in the brightness of the fluorescent bright spots of the same base type in multiple adjacent cycles, the advance reaction rate / lag reaction rate of each cycle is calculated to obtain visual reference information characterizing the degree of crosstalk interference between cycles in the current sequencing process.

7. The method for visualizing gene sequencing results based on deep learning according to claim 1, wherein: The sequencing index information includes inter-cycle crosstalk ratio information; the outputting the sequencing index information corresponding to the reference fluorescence image as visualization reference information of the base recognition model includes: The brightness of the fluorescent bright spots in the reference fluorescent image is determined, and the ratio of the fluorescent bright spot brightness in multiple reference fluorescent images corresponding to different base types in the same cycle is calculated according to the base type, thereby obtaining visual reference information representing the degree of crosstalk interference between channels in the current sequencing process.

8. The method for visualizing gene sequencing results based on deep learning according to any one of claims 1 to 7, wherein: Also includes: Outputting a base recognition reference result through the preset base recognition algorithm; determining a difference between a base call result of the reference fluorescent image and a base call result of the corresponding fluorescent image to be tested; If the difference exceeds the threshold, the base recognition model using the uncorrected fluorescence brightness information of the fluorescence image to be tested as input is replaced to execute the subsequent sequencing process.

9. The method for visualizing gene sequencing results based on deep learning according to any one of claims 1 to 7, wherein: Also includes: Outputting a base recognition reference result through the preset base recognition algorithm; determining a difference between a base call result of the reference fluorescent image and a base call result of the corresponding fluorescent image to be tested; If the difference exceeds the threshold, the base recognition model using the corrected fluorescence brightness information of the fluorescence image to be tested as input is replaced to execute the subsequent sequencing process.

10. The method for visualizing gene sequencing results based on deep learning according to any one of claims 1 to 7, wherein: The sequencing index information also includes image quality index information of the fluorescent image to be tested in each cycle, background brightness information of the fluorescent image to be tested, and a bright spot brightness distribution histogram of the fluorescent image to be tested.

11. The method for visualizing gene sequencing results based on deep learning according to any one of claims 1 to 7, wherein: The method for visualizing gene sequencing results based on deep learning further includes: Obtain operating performance parameters of the sequencing device, and output other sequencing indicator information based on the operating performance parameters; the other sequencing indicator information includes chip rotation angle and chip offset.

12. The method for visualizing gene sequencing results based on deep learning according to any one of claims 1 to 7, wherein: The reference fluorescence image includes an image determined by sampling the fluorescence images to be tested in the first few cycles of the current gene sequencing process.

13. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

14. A gene sequencing device comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which implements the steps of the method according to any one of claims 1 to 12 when executed by a processor.

Images