A method for acquiring a gene sequencing signal intensity

Abstract

The application provides a method for obtaining gene sequencing signal intensity, and is particularly suitable for the process of multi-base sequencing. Through the steps of image acquisition, image preprocessing, extraction of data point positions, and calculation of data intensity, the intensity data of high-throughput gene sequencing can be obtained more accurately.

Description

TECHNICAL FIELD

[0001] The application discloses a method for acquiring gene sequencing signal intensity, and belongs to the field of gene sequencing. BACKGROUND

[0002] Gene sequencing is a rapidly developing industry in recent years. The process of gene sequencing is actually a process of synthesizing a double-stranded DNA molecule. In high-throughput gene sequencing, the signal intensity is extremely low, and the micro-pits are extremely small, so the reaction is generally carried out in the range of 0.2-2 microns, and therefore the extraction of the signal is very difficult.

[0003] With the rapid development of the second-generation sequencing technology, the sequencing throughput of a sequencer is higher and higher. The higher throughput means that the same size of a sequencing image needs to carry more reaction micro-pits, and therefore the number of pixels represented by each reaction micro-pit or data point is smaller, and therefore a pixel may belong to two or more reaction pits. In some conventional sequencers, it is often only necessary to detect whether a reaction has occurred, that is, signal 0 and 1, and the requirement for intensity is not high, but in a high-throughput sequencer for multi-base sequencing, it is necessary to accurately obtain an intensity signal, and a reaction pit may have only a few pixels, and a difference of one pixel may lead to complete distortion of the signal. The application discloses a method for extracting gene sequencing signal intensity, which more accurately defines the sequencing signal. SUMMARY

[0004] The application discloses a method for acquiring gene sequencing signal intensity, and belongs to the field of gene sequencing.

[0005] (1) obtaining a sequencing result image through gene sequencing;

[0006] (2) image preprocessing, removing the influence of inconsistent intensities of regions of the image and removing the influence of background signals of the image;

[0007] (3) estimating the position of a data point;

[0008] (4) planning a graph matched to the size of the data point with the center of the data point as the center; the graph comprises a plurality of pixel values of the data point;

[0009] (5) calculating the pixel intensity value in the graph to obtain light intensity information corresponding to the gene to be measured; wherein a pixel at the edge of the graph is split into a plurality of pixels through a high-resolution method, and the part of the split pixel inside the graph is proportionally attributed to the pixel value inside the graph.

[0010] According to a preferred embodiment, the estimation of the position of the data point refers to the calculation of the position of the data point through the size of the data point in the previously known physical size of the sequencing chip and the data point information in the sequencing result image.

[0011] Splitting a pixel into several pixels by high resolution method refers to splitting the pixel at the edge of the figure; the part inside the figure is calculated as the pixel inside the figure; when calculating the intensity value, multiply the pixel at the edge by the proportion of its pixel inside the figure to obtain the intensity of the split pixel.

[0012] According to the preferred embodiment, the removal of the influence of the inconsistency of the intensity of each region of the image refers to eliminating the influence of the non-uniformity of the light field on the image.

[0013] According to the preferred embodiment, the estimation of the position of the data point refers to first obtaining the position of the pit by the local maximum value method, and then using the weighted calculation method to represent the center of the pit.

[0014] According to the preferred embodiment, the background signal is obtained by image morphological opening operation or adding a non-fluorescent reagent.

[0015] According to the preferred embodiment, the sequencing refers to a 3-end non-closed gene sequencing method.

[0016] According to the preferred embodiment, further comprising a second removal of the influence of the background signal, further distinguishing the bright data points and the dark data points in the pixel intensity value of the figure in step (5), and further removing the second background signal of the dark data points, the method for removing the second background signal is obtained by statistical information of isolated dark data points; wherein the isolated dark data point information refers to that the data point does not participate in the reaction in the process of this gene sequencing, and is not directly adjacent to the bright data points that react.

[0017] The present application discloses a method for extracting the intensity of sequencing data, which is particularly suitable for the process of multi-base sequencing. In the present application, the shape is used to define the range of data points, and the method of dividing pixels is used to more accurately define the intensity of sequencing data. This is beneficial to optimize the data of gene sequencing, and makes the extraction of signal intensity in the process of multi-base sequencing easier and more accurate. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 . Flow chart of obtaining the intensity of sequencing signal;

[0019] Figure 2 . Shape template for extracting pit intensity;

[0020] Figure 3 . Example diagram of pixels across the edge of the shape;

[0021] Figure 4 . Shape definition of data points to define the range thereof;

[0022] Figure 5 . Signal intensity histogram, horizontal axis is intensity, vertical axis is normalized percentage;

[0023] Figure 6 . Bright data points selected. DETAILED DESCRIPTION

[0024] In order to further illustrate the core of the present application, the present application will be described with the following examples. The examples are to further explain the summary of the invention and do not limit the present application.

[0025] The gene sequencing described in the present application refers to the gene sequencing method without 3' end closure. When the 3' end is closed, the obtained gene sequencing signal can be represented as 1 and 0. When the 3' end is not closed, since the sequencing occurs continuously, multiple bases are reacted at one time, and therefore, a signal greater than 1 can be obtained, such as signal intensity of 2, 3, 4, 5, etc.

[0026] The gene sequencing signals 1 and 0 described in the present application represent the relative intensity values of the gene sequencing signals after removing the background values. For example, when a base is elongated, the signal intensity relative value can be considered as 1, and when N bases are elongated, the signal intensity is N. In actual sequencing, the intensity of the signal cannot be a specific integer value. Generally, for example, the signal intensity relative value of one base is 1, and other signal values are the comparison results with this value, such as the signal intensity of N times of one base.

[0027] When the signal is only 1 and 0, the method is not very meaningful. Because it only needs to judge the presence and absence of the signal.

[0028] When the signal is greater than 1, such as 3, 6, etc., and the sequencing throughput is higher, it is relatively difficult to accurately determine the signal value in the microwell due to the influence of the common pixels between the microwells. The method described in the present application has great advantages.

[0029] Sequencing is a process of detecting unknown or known sequences. In the process of gene sequencing, it is generally a chemical reaction process. Since the gene is very small, it is difficult to detect a single gene sequencing signal. The current single molecule detection technology is not suitable for large-scale gene detection. Therefore, the second-generation sequencing technology is the basic and commonly used sequencing technology. The second-generation sequencing technology is to amplify the gene fragments to be detected into more sequences on the basis of PCR and the like, so as to facilitate detection. Since the data volume is very large, the size of each data point is also very small.

[0030] The spatial structure of the gene sequencing chip is regular, such as the micro-pit with the characteristics of square or hexagonal arrangement, and cannot be a random micro-pit. The conventional gene sequencing chip is also regular. Therefore, the method is suitable for the common gene sequencing chip with micro-pits. The gene sequencing chip with micro-pits generally has an array of regular micro-pit structures on one or two bottom surfaces of the sequencing chip. Each micro-pit structure is a detection point.

[0031] In high-throughput DNA sequencing, ideally, one pit corresponds to one sequence signal intensity, but in practice, due to the influence of optical PSF, etc., as shown in the following formula, the intensity of each pit will affect the intensity of the surrounding pits, that is, the intensity of each pit will interfere into the surrounding pits. Figure 1 In some conventional sequencers (such as illumina, etc.), the detection of DNA signals is only a qualitative detection, that is, only the presence or absence of the signal is judged, and in these instruments, the influence of the crosstalk is small and can be ignored. However, in some new sequencing implementations, it is necessary to quantitatively detect the signal intensity of each reaction pit, and at this time, the real reaction intensity of each pit needs to be accurately measured, especially when high DPL signals exist, the crosstalk will seriously affect the sequencing quality. If the influence of the crosstalk is not removed, the intensity extracted from a single pit cannot actually reflect the light intensity produced by the reaction of a single pit, resulting in errors in the final results. Therefore, the present application designs a method for removing the crosstalk in the sequencing signal, which effectively removes the influence of the crosstalk of the surrounding pits.

[0032] The word threshold is mentioned in the present application. The word threshold is not a new term. For example, the use of threshold is also mentioned in patent CN102834828B. The threshold can also be called critical value. In the present application, the concept of threshold is used, which means that the signal higher than the threshold corresponds to the micro-pit that occurs sequencing reaction, and the signal lower than the threshold corresponds to the micro-pit that does not occur sequencing reaction. Usually, this value is a statistical value or a calculated value. The selection of the threshold is not a complex procedure. The signals of the reacted micro-pits and the unreacted micro-pits are obviously different. Therefore, the selection of the threshold is relatively simple, and there are many choices. Therefore, it is not complex. Based on the understanding of the principle and in combination with the previous technology, the threshold is easy to select.

[0033] The gene sequencing described in the present application refers to the method of 3-end unblocked gene sequencing. When the 3-end is blocked, the obtained gene sequencing signal can be represented as 1 and 0. When the 3-end is not blocked, since the sequencing occurs continuously, multiple bases are reacted at a time, and therefore, a signal greater than 1 can be obtained, such as signal intensity of 2, 3, 4, 5, etc.

[0034] Here, signal 1 means that the sequencing reaction is simply considered to be extended by one base. In the process of gene sequencing, it can be considered that all nucleic acid molecules on a data point are extended by one base. Of course, in the actual reaction, not all nucleic acids on the data point may be reacted, which is also an important reason for the error of sequencing data.

[0035] The gene sequencing signals 1 and 0 described in the present application represent the relative intensity values of the gene sequencing signals after background values are removed. For example, the signal intensity is considered to be 1 when one base is extended, and the signal intensity is N when N bases are extended. In actual sequencing, such signal intensity is not simply a multiple relationship, and may be related to many influencing factors. The signal value obtained in actual sequencing is not an integer, but its integer signal intensity can be obtained by comparison with single-base sequencing signals.

[0036] When the signal is only 1 and 0, the method has little meaning. Because it only needs to judge the presence and absence of the signal.

[0037] When the signal is greater than 1, for example, 3, 6, etc., it is relatively difficult to accurately determine the signal value in the micro-pit due to crosstalk between micro-pits and other influencing factors. The method described in the present application has great advantages.

[0038] The micro-pit described in the present application refers to the micro-reaction chamber used in the sequencing reaction. In the process of sequencing reaction, a certain intensity of sequencing reaction signal is needed for detecting the sequencing signal. At present, the second-generation sequencing is carried out by using a method similar to PCR to amplify the gene fragments, and when the number of gene fragments reaches a certain degree in a certain small range, it can be detected by an objective lens and a CCD. In the present application, the micro-pit refers to the small reaction chamber in which the gene fragments are located during the sequencing reaction. The size of the general reaction chamber is 0.2-3 microns in diameter; preferably, 0.3-2 microns in diameter; more preferably, 0.35-1.8 microns in diameter. The depth of the sequencing reaction chamber is generally 0.2-3 microns; preferably, 0.3-2 microns; more preferably, 0.35-1.8 microns. The micro-pit belongs to the common words in the art. The micro-pit described in the present application is consistent with the conventional meaning in the art.

[0039] In the present application, sequencing refers to high-throughput second-generation gene sequencing.

[0040] The micro-pit described in the present application refers to a regular arrangement of micro-pits formed on the surface of glass, silicon wafer, etc. by certain technical means. The use of micro-pits belongs to the common technology in the art.

[0041] In the process of gene sequencing, different sequencing methods used will affect the type of signal. However, the sequencing signal of the 3-end not being closed belongs to the data type range of the present application.

[0042] The present application discloses a method for extracting intensity from a sequencing image to obtain reaction intensity information corresponding to the information of a gene to be tested.

[0043] The present application discloses a method for extracting intensity from a sequencing image, characterized in comprising the following steps,

[0044] (1) obtaining a sequencing result image by a 3-end unclosed gene sequencing method;

[0045] (2) image preprocessing, removing the influence of inconsistent intensity of each region of the image, and removing the influence of background signal of the image;

[0046] (3) estimating the position of a data point;

[0047] (4) planning a pattern matching the size of the data point with the center of the data point as the center; the pattern comprises a plurality of pixel values of the data point;

[0048] (5) calculating the pixel intensity value in the pattern to obtain light intensity information corresponding to the gene to be tested; wherein the pixels at the edge of the pattern are split into a plurality of pixels by a method of increasing resolution, and the part of the split pixels inside the pattern is proportionally attributed to the pixel value inside the pattern;

[0049] The estimation of the position of the data point refers to the calculation of the position of the data point by the size of the data point in the pre-known physical size of the sequencing chip and the data point information in the sequencing result image.

[0050] The splitting of one pixel into a plurality of pixels by the method of increasing resolution refers to the splitting of the pixels at the edge of the pattern; the part inside the edge of the pattern is calculated and attributed to the pixel inside the pattern; when the intensity value is calculated, the pixel at the edge is multiplied by the proportion of its pixel inside the pattern to obtain the intensity of the split pixel.

[0051] According to a preferred embodiment, the removal of the influence of inconsistent intensity of each region of the image refers to the elimination of the influence of non-uniform light field on the image.

[0052] According to a preferred embodiment, the background signal is removed by an image morphological opening operation or by adding a non-fluorescent reagent to remove the influence of the background signal.

[0053] The flowchart of the present application is shown in Figure 1 .

[0054] From another point of view, the present application mainly comprises the following steps

[0055] 1. Image preprocessing

[0056] The real sequencing image signal refers to the sequencing signal without considering the external factors of the light field and the background.

[0057] The signal of the sequencing image is actually a combination of the real signal, the background signal and the light field effect.

[0058] The sequencing image signal can be simply represented as

[0059] Y_obs = K (X_Ori + B)

[0060] Where X_Ori is the real sequencing image, which can also be called the real signal intensity or the real signal; Y_obs is the observed sequencing image, which can also be called the sequencing image or the sequencing image signal, or the sequencing signal intensity; K is the light field image, which can also be called the light field coefficient. B is the background image, which can also be called the background signal intensity or the background signal. Now we have the observed signal Y_obs, in order to obtain the real sequencing image signal

[0061] X_Ori = Y_obs / K - B

[0062] We need to remove the effects of the background signal B and the light field K.

[0063] Background B removal: based on the image morphological opening operation method, after the opening operation of the image according to the given template, each pixel value is the local minimum value, and the background signal can be obtained.

[0064] Light field K removal: the intensity of light in space decays according to the square of the distance, and since the distance between each position of the chip in the shooting area and the light source is different, the intensity of light at each position on the picture will be different, resulting in different intensities of the same intensity of the fluorescent DPL signal in different areas. The light field removal is to remove the influence of this factor.

[0065] In order to remove the influence of the light field, it is necessary to estimate the light field image K. Since the intensity of light in space decays according to the square of the distance, the intensity reflected on the surface of the chip is approximately in the form of a quadratic surface, so we estimate the light field image K by quadratic surface fitting.

[0066] Sequencing, obtaining the sequencing image signal, and then removing the background signal and the light field effect to obtain the real sequencing image signal.

[0067] 2. Extract position

[0068] The chip used for sequencing has micro-wells arranged in a regular pattern. In theory, the position of all the micro-wells can be determined by knowing the position of a few of them. In practice, due to the precision of the manufacturing process and the conditions of the photograph, the exact position cannot be determined by calculation. During sequencing, the center of the micro-well is brighter than the surrounding area, so the rough position of the micro-well can be determined by a mathematical estimation method based on the approximate position. In short, the rough position of the micro-well is determined by the local maximum value.

[0069] The DNA being sequenced emits light in the micro-well, so each micro-well is illuminated, and the pixel intensity value is greater near the center of the micro-well. Based on this characteristic, the approximate center position of the micro-well can be determined by the local maximum value method. Since the accuracy we need is at the sub-pixel level, the position of the micro-well is adjusted by a weighted estimation method based on the approximate center position.

[0070] Coarse extraction: According to the sequencing principle, each sequence is adsorbed in a micro-well, so the position of the micro-well can be preliminarily determined by the local maximum value method. That is, the local maximum value image is obtained by closing the image, and the pixels equal to the original image in the local maximum value image are the rough positions of the micro-wells.

[0071] Fine adjustment: However, due to the point spread function, each light point will spread into an area or some pixels, so we cannot simply give a position or coordinate for each micro-well. In order to accurately represent the coordinates of a micro-well, we estimate its centroid by considering the gray scale information in the spread area of the micro-well. Specifically, for each micro-well, the gray scale value of each pixel is considered, and the weighted average calculation method is used to represent the center of the micro-well.

[0072] 3. Intensity extraction

[0073] After determining the center positions of the micro-wells, the intensity in each reaction well is extracted. There are two main parts: one is to split the pixels and distribute them to different reaction wells, and the other is to determine the template for intensity extraction, that is, to determine which pixels are used to calculate the intensity of each well.

[0074] For pixel splitting, we use the image upsampling method. For the intensity extraction template, several commonly used templates are as follows:

[0075] a. Draw a circle with the center position of the micro-well as the center and the size of the micro-well as the radius;

[0076] b. Draw a rectangle with the center position of the micro-well as the center;

[0077] c. Custom shape, such as selecting the largest number of pixels near the center according to the reaction characteristics, or giving a specific shape a priori;

[0078] Select the appropriate template based on the chip's pitting and optical characteristics. See [link / reference] Figure 2 In the process, circular templates, square templates, or templates of irregular shapes can be used.

[0079] After upscaling the image, a defined template is used to calculate the intensity, centered on the center of the extracted pits. Pixels within the template's coverage area are statistically accumulated and added to the final intensity. The final intensity divided by the number of pixels gives the final extracted intensity.

[0080] See Figure 3 A real-world example image showing pixels spanning multiple edges. When the defined shape, such as a circle, has pixels along its edge, these pixels need to be divided proportionally. In the image, the circular edge involves multiple pixel values, requiring proportional division.

[0081] When extracting locations, the defined shape is generally determined based on the characteristics of the micro-pits or data points. For example, circular micro-pits or near-circular data points can be defined using circles. The radius of the circle can generally be directly represented by the radius of the micro-pit, or it can be fine-tuned based on the data. Of course, this fine-tuning is generally determined by the characteristics of the data. A simple example is that when there are many high-intensity pixels within a micro-pit, the diameter of the circle can be defined to be slightly larger than the diameter of the micro-pit, for example, 1.1 times the diameter of the micro-pit. This fine-tuning is more conducive to data integrity. Therefore, the defined shape is adapted to the shape of the micro-pit or data point, and its size is 0.5-2 times the size of the micro-pit or data point, preferably 0.7-1.5 times, more preferably 0.8-1.2 times, more preferably 0.9-1.1 times, and more preferably 1 time.

[0082] The final extracted intensity information is not an integer. This process effectively removes the influence of external conditions, such as the light field and background, resulting in intensity information that may not perfectly reflect the actual number of bases sequenced. Further phase correction and other methods can then be used to obtain the true signal of the micro-pitted signal. This process can be referenced in another patent of the applicant, a method for reading sequence information from the raw signal of high-throughput DNA sequencing (CN201610899880.X). The final extracted intensity described in this invention is the raw signal from that patent.

[0083] In typical sequencing, the obtained signal can be transformed into a real signal in several ways. Simply by extracting the numerical value of a sequencing signal. When other influences are present, computational removal can also achieve a signal close to the real one. When experimental conditions are simple, the final extracted intensity signal can be obtained through a simple extraction method.

[0084] For 3-end closed sequencing signal, the data type is typical yes and no. That is, it can be expressed in the form of 0 and 1. In this case, the method described in the present application can also be used. For example, when the data is not good enough, such as the number of PCR rounds is relatively low, or better data results are needed; in order to make the sequencing more accurate, the method described in the present application can also be used. However, it can be known that this data type is relatively simple, such as the strength of crosstalk is considered to be 1% to 5%, so the improvement of data is limited, and the quality of sequencing data will not be substantially improved. According to simple experimental prediction, in this case, the data accuracy can reduce the error rate from 2% to 1.8% without using any crosstalk removal technology. Of course, this data will change according to the original sequencing error rate. It must be pointed out here that among all the accuracy data of the sequencing instrument currently disclosed, it is the result after the data calculation of the disclosed method, and cannot be directly treated as the error rate of the original data.

[0085] The processing method of the background signal is related to the principle. In the present application, when the background signal is processed, image morphological opening operation or adding non-fluorescent reagent to remove the influence of the background signal can be used.

[0086] When other existing technologies are used in combination, the data accuracy can reach the industry Q60 standard.

[0087] The extraction intensity process is a process of representing the sequencing signal of the data point in a reliable manner. The extraction of the intensity of a single data point is a relatively simple process. It can be obtained by using a simple range circumscription, such as pixel confirmation. However, the characteristics of multiple signal of gene sequencing determine that simply extracting the intensity of each micropit cannot obtain the actual or real sequencing signal of the micropit. The range circumscription is an image processing problem. Generally, the sequencing micropit is circular. A circle can be drawn to circumscribe the range of the micropit. Of course, in order to accurately calculate, the range can be set to a mode smaller than the period of the micropit. This is operated according to the relationship between the size of the micropit and the pixel value. It is an empirical value.

[0088] The applicant's previous disclosed technology (CN201610899880.X) belongs to post-processing of data. In combination with the process of data extraction of the present application, more realistic performance can be achieved for sequenced data. In addition, the applicant also discloses a data error correction method (CN201510944878.5), hereinafter referred to as the ECC method, which is suitable for multi-base sequencing data with mutual orthogonal relationship. Furthermore, the method described in the present application is suitable for multi-base sequencing, the method described in CN201610899880.X is suitable for all sequencing fields, and the ECC method is suitable for multi-base mutual orthogonal sequencing. Therefore, when multi-base mutual orthogonal sequencing is performed, the aspects described in the present application and the previous two methods can be used in combination to further increase the accuracy of data.

[0089] The DPL of each microwell is actually the pixel value of each microwell. In the present application, a general DPL such as 4-64 is used. The pixel value of a single microwell or data point is an important element to be considered in actual sequencing. When the pixel is too high, for example, greater than 1000, the total amount of data will be relatively low, and high-throughput sequencing will not have much significance.

[0090] In the present application, the microwell also refers to a single data point. Therefore, in the present application, the term microwell can also be replaced by data point. Therefore, the corresponding dark well is also a dark data point, and the bright well is also a bright data point.

[0091] After extracting the intensity, it can be clearly found that some data points are obviously brighter, and many data points are obviously darker. The microwells in which the brighter data points are located are microwells in which reactions occur; the microwells in which the darker data points are located are microwells in which no reaction occurs. After obtaining this data, a conventional method can be used to simply indicate whether a data point is bright or dark.

[0092] In general sequencing, the size of the microwell is about 0.5-10 microns. In the sequencing described in the present application, the microwell involved is 0.5-4 microns in diameter, preferably 1-3 microns in diameter, and more preferably 1.5-2.5 microns in diameter.

[0093] In general sequencing, the number of pixels when taking a picture of each microwell needs to exceed 4. In the sequencing described in the present application, the number of pixels of each microwell is 3-100, preferably 12-75, and more preferably 24-64. In some embodiments of the present application, 48 pixel points are used, and a lower or higher level does not affect the method described in the present application.

[0094] K is the light field coefficient, which represents the influence of the light field on the sequencing. Generally, the light field at different positions is not the same. Simply put, the light intensity at different positions is not the same. After correction by hardware, the light field still has certain differences, for example, the light field intensity at the edge of a certain field of view and at the middle position is different. This difference is often reflected in the improvement of the overall signal intensity. That is, the overall signal value of the area with strong light field is high, and the overall signal value of the area with weak light field is low. For example, when the number of fluorescent molecules of a certain data point is a fixed number, the number of photons in the area with strong light field is more, resulting in high light intensity collected by the CCD photographing device, and the number of photons in the area with weak light field is less, resulting in low light intensity collected by the CCD photographing device.

[0095] According to the above explanation, the formula is:

[0096] Y obs = K (X ort + B)

[0097] It can also be described as the following formula:

[0098] Y obs = KX Ori + b

[0099] In the two formulas, the meaning of B is different. The background signal B in the first formula is the background without the correction value of the light field intensity. The b (here, B is changed to b to distinguish) in the second formula is the actual detected background signal. Both formulas are correct. In actual application, both formulas are feasible, and there is no substantial difference in calculation method.

[0100] It should be noted that in the present application, B and b are not strictly distinguished. Strictly speaking, B = Kb. However, in actual operation, since b or B is unknown, the background signal needs to be determined by a certain method, for example, the average value of multiple points, therefore, B can be called the background signal, Kb can also be called the background signal, and b can also be called the background signal. These calculation methods will not affect the calculation results, but only the different explanations.

[0101] K value can be seen as a pre-test value. For example, K value can be obtained by taking a picture of a blank before sequencing. The meaning of K value is that the light field image represents the deviation of the light field to the image. Generally, due to the non-uniform illumination of the light source, the light intensity at different positions of the light field is different. This difference does not need to be accurately measured in many cases, but when the sequencing reaction is performed, the sample amount is small, and the sequencing signal is very weak compared to the common case; Therefore, the influence of the light field needs to be considered. In addition, the intensity of the light source may change over time, so it is necessary to measure the light field coefficient K. Various ways can achieve this measurement, for example, by correcting the blank sample, specifically, during the sequencing process, first take a picture to obtain the K value of different regions of the light field, or determine it according to the empirical light intensity attenuation coefficient.

[0102] The process of gene sequencing is also the process of translating the gene sequence. In current technology, the common method is the second-generation high-throughput sequencing method. In the process of synthesizing double-stranded DNA from single-stranded DNA, a fluorescent labeling method is used to identify the base involved in each reaction. It is very difficult to identify a single molecule, so generally a small amount of DNA fragments needs to be amplified to obtain a large number of molecules, and then during the reaction, the information of the reaction is obtained by taking a picture to record the reaction through fluorescent labeling or participation using CCD equipment.

[0103] The sequencing described in the present application refers to the second-generation sequencing technology, particularly the second-generation high-throughput sequencing technology.

[0104] In the second-generation sequencing technology, for example, illumina uses a 3-end closed sequencing method. After 3-end closing, only one base can be extended in each reaction. The result of the reaction can be represented by the number of extended bases, for example, one base extension is recorded as 1, and no reaction is recorded as 0. The actual situation is of course much more complex than the numbers 1 and 0. The gene sequencer uses a CCD camera device to take pictures, and then obtains signal values by extracting image intensity, and then according to certain rules, the specific value corresponding to the base extension number is obtained. Moreover, chemical reactions are affected by many factors, and it is very difficult to correspond the actual obtained image intensity to the number of base extensions. When illumina sequencing is performed, since it only extends one base, its signal is relatively regular and easy to extract signal intensity and correspond.

[0105] When sequencing, the 3' end is not blocked, and each extension can extend multiple bases, such as 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, etc. In this case, the corresponding signal intensity is more complex and needs to be more accurate, which is much more difficult than the signal extraction of illumina. The present application designs a signal processing method for 3' end unblocked sequencing. In order to more accurately extract the signal and obtain a better corresponding relationship, background removal, accurate estimation of data point position, pixel splitting and other steps are required.

[0106] It should be noted that the method described in the present application is also applicable to the 3' end blocked sequencing method, and can also improve the accuracy of the results, but from the necessity aspect, it is not very important.

[0107] It should be noted that in the present application, the default gene sequencing chip has micro-pits. Generally, each micro-pit has a gene fragment to be tested. During sequencing, depending on the different sequencing reagents, the gene fragments in the micro-pit can participate in the reaction or not. It should be noted that the micro-pit is a gene sequencing method, and some gene sequencing chips do not require micro-pits, so each gene fragment forms a cluster. When displayed on the sequencing image, the two are slightly different. However, their physical size or form has certain regularity. Therefore, in the present application, the micro-pit is also referred to as a data point. Or more accurately, in the present application, all descriptions of micro-pits can be considered as carriers of data points; the concept of micro-pit can be replaced with data point.

[0108] Further, the removal of background signal described in the present application is to remove the background signal of non-sequencing influence. During actual sequencing, due to the influence of sequencing fluorescence signal crosstalk, a second background signal is further generated. Generally, the second background signal is not significantly distinguished from the background signal (also referred to as the first background signal) described above. However, due to the influence of the calculation steps, all background signals are not strictly considered for removal in the method described in the present application. In this case, the second background signal can be further removed.

[0109] The following information describes the removal and principle of the second background signal:

[0110] The processing method of the second background signal has a principle. In the present application, when processing the second background signal, the signal value of the dark pit can be counted, and the signal value is taken as the average value of the background signal. The statistics described herein do not require statistics of all results, because the signal intensity of the sequencing chip in a small area is relatively parallel, and generally a part of the results can be counted. Therefore, the average value of the background signal can be obtained after distinguishing the bright and dark pits, or after screening out the isolated bright pits.

[0111] After extracting intensity, it can be found that some data points are obviously brighter, and some data points are obviously darker. The micro-pits where the brighter data points are located are the micro-pits where the reaction occurs, and the micro-pits where the darker data points are located are the micro-pits where the reaction does not occur. After obtaining the data, a conventional method can be used to simply indicate whether a data point is bright or dark. Since the sequencing signals are relatively close to each other, the bright micro-pit can have a radiation effect on the surrounding, which is often referred to as signal crosstalk. Therefore, when there is no bright micro-pit within the range of, for example, two layers of micro-pits around a darker micro-pit, it can be considered that the darker micro-pit belongs to the background signal. By counting a large number of background signals and giving an average value, the average background signal can be obtained. In the process of calculating the background, the influence of the brighter micro-pit on the other unreacted micro-pit is also considered, which is equivalent to the process of calculating thermal radiation. The difference is that in the process of radiation, the surrounding of a bright point is not an average attenuation process, but a sharp attenuation in units of micro-pits. Generally, a bright point affects 1-2 layers of surrounding data points, and more range will not be affected.

[0112] Then, among all the bright pits, isolated pits are selected. As shown in Figure 6 The definition of an isolated pit is that there is no other bright pit within a certain range around the bright pit in the x direction and the y direction. The size of the range is determined according to experience in advance. In the implementation, for each bright pit, the distances of all other bright pits to it are counted. If the minimum distance among the distances is greater than the range set in advance, the bright pit is marked as an isolated pit.

[0113] In the present application, a first-order crosstalk micro-pit has been defined. Simply, the remaining 8 directly adjacent data points around a data point belong to the first-order crosstalk micro-pit or first-order crosstalk data point. The data point is a two-dimensional point type on the image.

[0114] In the present application, a second-order crosstalk micro-pit is defined. Simply, the second layer of micro-pits adjacent to the first-order crosstalk micro-pit belong to the second-order crosstalk micro-pit, and there are 16 in total.

[0115] According to the estimation of simple crosstalk, the isolated dark pit can be accepted for calculating the background signal.

[0116] According to the preferred embodiment, a certain number of unreacted micro-pits are selected for calculating the background signal, wherein the unreacted micro-pits are spaced apart from the reacted micro-pits by more than or equal to one micro-pit. That is, the first-order crosstalk micro-pit of a certain bright pit cannot be used as a background signal data point. That is, the unreacted micro-pit does not belong to the first-order crosstalk micro-pit of a certain bright pit.

[0117] Figure 6The selected bright pits are represented by the form "x" at their positions.

[0118] For each isolated bright spot, extract a 5x5 intensity matrix centered on it from the calculated intensity, and then...

[0119] An isolated bright pit is defined as a bright pit in which none of the surrounding micro-pits or its secondary micro-pits have undergone sequencing.

[0120] Simply put, if no reaction occurs in the first-order and second-order adjacent micro-pits around a bright pit, then the bright pit is defined as an isolated bright pit.

[0121] For each bright spot, its spot distance is defined as the minimum distance from all other bright spot points to that spot.

[0122]

[0123] dist i Let dot be the pit distance of the i-th bright pit. i Let |·| represent the position of the two points, and |·| represent the distance between them.

[0124] If dist > 4 * DL, then mark it as an isolated pit.

[0125] like Figure 6 As shown, x represents isolated pits that have been filtered out, and o represents dark pits.

[0126] Figure 6 The intensity matrix corresponding to the isolated bright pit is:

[0127] 158.9086 155.112 154.9645 154.5978 154.4302 158.8925 164.9705 170.1037 158.4598 155.6899 157.3772 176.9054 1219.421 179.7076 160.4912 163.0005 158.0009 181.5133 163.2679 161.3172 157.6925 158.0391 159.7521 154.154 160.4775

[0128] For each intensity matrix, the outermost 16 points represent the background pits. The median of these 16 points is taken as the background pit brightness. In the example above, the background pit brightness B = 157.8658. The background pit signal value at this location is the average value of the background signal.

[0129] It's understandable that the above describes the process of removing the second background signal. Although both can be referred to as background signals, the first background signal is removed through 1. image morphological opening operations; 2. imaging after using a seal wash buffer. The seal wash buffer refers to sealing a non-fluorescent liquid within the chip before imaging to obtain the background signal. This non-fluorescent reaction solution simply replaces the sequencing reaction solution, thus eliminating its fluorescence influence. Generally, this is an aqueous solution; water or other liquids that are helpful or auxiliary to the sequencing reaction can be considered.

[0130] When extracting the position of the pits or data points, the physical size of the pits or data points has a certain range due to the physical size or amplification multiple thereof. Through the range, the entire chip or a certain photographed area can be simply divided into a small data point area. The chip with pits is more likely to obtain specific data point position information. In the pits, the information obtained by the reaction is more concentrated and more regular. The specific position of the pits can be obtained by first roughly obtaining the position of the data points and then finely adjusting. There is no technical obstacle in image processing as long as the principle is understood. The present application applies various methods of image processing to the processing of gene sequencing data, which can better process the information of the gene sequencing image and obtain more real results.

[0131] Embodiment 1

[0132] The image is obtained by photographing the chip using a conventional MK sequencing process. The sequencing process can refer to the previous patents CN201510822361.9 or CN201510815685.X. According to the data of the gene sequencing chip, the size of the pit is 2 microns in diameter, the period is 3 microns, the depth is 2 microns, and the shape is cylindrical.

[0133] According to the position and size of the pit, a series of circles (or rectangles or other shapes) are drawn as shown in Figure 4 , and the average value of all pixels in each circle is calculated as the signal intensity of the pit. Since the pits are distributed horizontally and vertically on the chip, the intensity matrix I can be obtained according to the geometric position distribution of the pits. Each element value of the matrix represents the intensity value of a certain pit.

[0134] When calculating the signal intensity, since the general pit is circular, the pit position is circled by drawing a circle. When the pit is of other shapes, the method can be changed.

[0135] When calculating the signal intensity of four adjacent pits, the common adjacent non-circular area can be considered as a non-calculation area. The influence of this area can be ignored during actual calculation, and no calculation is performed.

[0136] When the number of pixel points in the selected circle is not an integer, it only needs to be simply cut. For example, 31.4 pixel values only need to multiply the edge pixel value by 0.4 for simple operation.

[0137] A histogram of all pit intensities is shown in Figure 5 , and the lowest value between the dark pit and the bright pit peak is taken as the threshold value to distinguish between a dark pit and a bright pit. The threshold value is a signal intensity value, and the value is higher than the threshold value, and the value is lower than the threshold value.

[0138] Figure 5 Signal strength histogram, horizontal axis represents intensity, vertical axis represents percentage normalized to maximum number.

[0139] It should be noted that the specific embodiments of the present application are an explanation of the innovation. Special examples are selected. Therefore, the specific embodiments in the present application do not limit the protection scope of the present application. The principle of matrix calculation or other calculation methods in the specific embodiments does not deviate from the specific principles of the technical solutions claimed by the present application, and no other restrictions are added. The specific calculation method does not limit the protection scope of the present application.

[0140] The above is only the preferred specific embodiment of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for obtaining gene sequencing signal intensity, characterized in that... Includes the following steps, (1) Obtain sequencing result images through gene sequencing; (2) Image preprocessing: remove the influence of inconsistent intensity in different regions of the image and remove the influence of background signal; (3) Estimate the location of the data points; (4) Using the center of the data point as the center, plan a graphic that matches the size of the data point; the graphic contains multiple pixel values ​​of the data point; (5) Calculate the pixel intensity values ​​in the image to obtain the light intensity information corresponding to the gene to be tested; Pixels located at the edge of the graphic are split into multiple pixels by increasing the resolution, and the portion of the split pixels that is inside the graphic is proportionally incorporated into the pixels inside the graphic.

2. The method according to claim 1, characterized in that, The estimated location of the data point refers to calculating the location of the data point using the known size of the data point in the physical dimensions of the sequencing chip and the data point information in the sequencing result image.

3. The method according to claim 1, characterized in that, In step (5), a pixel is split into multiple pixels by increasing the resolution, which means splitting the pixels at the edge of the image; the part located inside the edge of the image is calculated and included in the pixels inside the image; when calculating the intensity value, the pixel at the edge is multiplied by the proportion of the pixels inside the image to that pixel to obtain the intensity of the split pixel.

4. The method according to claim 3, characterized in that, The removal of the influence of inconsistent intensity in different regions of the image refers to eliminating the influence of uneven light field on the image.

5. The method according to any one of claims 1-4, characterized in that, The step (2) of removing the influence of the image background signal refers to removing the influence of the background signal by performing image morphological opening operations or adding non-fluorescent reagents.

6. The method according to any one of claims 1-4, characterized in that, The sequencing was a multibase gene sequencing with open ends.

7. The method according to claim 6, characterized in that, The threshold is used to determine which data points react and which do not.

8. The method according to any one of claims 1-4, characterized in that, Satisfy the following formula: Y obs =K(X Ori +B) Among them, X Ori Y represents the signal intensity of the actual sequencing image. obs K represents the intensity of the sequenced image signal obtained from observation, K represents the light field image coefficient, and B represents the background signal intensity.

9. The method according to any one of claims 1-4 further includes a second removal of the influence of background signals, further distinguishing between bright and dark data points in the pixel intensity values ​​of the image in step (5), and further removing the second background signal of the dark data points, wherein the method for removing the second background signal is obtained by statistical analysis of isolated dark data point information; wherein isolated dark data point information refers to data points that did not participate in the reaction during the gene sequencing process and are not directly adjacent to bright data points that have participated in the reaction.

10. A gene sequencer, characterized in that, It includes a sequencing system and a computing system, wherein the image obtained by the sequencing system is used to obtain the true signal intensity of the sequencing data through the computing system; the method for obtaining the true signal intensity is the method described in any one of claims 1-9.

Images