Pulmonary nodule screening model building method and device and pulmonary nodule screening method and device

Abstract

The invention relates to a pulmonary nodule screening model building method and device and a pulmonary nodule screening method and device. Specifically, the invention relates to a method for establishing a pulmonary nodule screening model, and the method comprises the following steps: screening a predetermined number of regions with the maximum difference of weighted fragment distribution difference (WFDD) values in benign nodule crowd samples and malignant nodule crowd samples in a training set as feature data in the whole range of a human reference genome; and establishing the pulmonary nodule screening model by using the feature data. According to the method, non-invasive detection of human body nodule types is achieved, invasive examination is avoided, the judgment accuracy higher than that of existing CT scanning is provided, the accuracy of the method is close to that of invasive tissue biopsy, the method is developed based on cf DNA sequencing data, only blood drawing or other body fluid collection or other operations are needed, and the radioactive exposure risk is avoided.

Description

Method and device for building a pulmonary nodule screening model and method and device for pulmonary nodule screening Technical Field

[0001] The present invention relates to the technical field of bioinformatics, and in particular to a method and device for establishing a lung nodule screening model and a lung nodule screening method and device. Background Art

[0002] Cancer has become the leading cause of death in China, and its incidence continues to rise. According to the 2019 National Cancer Report, released by the National Cancer Center, malignant tumors account for 23.91% of all deaths among Chinese residents, and annual medical expenses attributable to malignant tumors exceed 220 billion yuan. Lung cancer ranks first among malignant tumors in my country, ranked by incidence.

[0003] Pulmonary sarcoidosis is a multi-system, multi-organ granulomatous disease of unknown etiology, often affecting the lungs, bilateral hilar lymph nodes, eyes, skin, and other organs. The chest invasion rate is as high as 80% to 90%. Pulmonary sarcoidosis generally has a good prognosis. Early lung cancer, on the other hand, often presents as small nodules in the lungs. Therefore, distinguishing nodule types is particularly important for early lung cancer screening. Currently, the most reliable method for determining lung nodule type relies primarily on invasive surgical sampling and tissue biopsy.

[0004] The World Health Organization states that early detection and treatment are key to effective cancer treatment. Therefore, developing early cancer screening and detection technologies is crucial. Currently, the noninvasive diagnosis of nodules (such as lung nodules) in clinical practice primarily relies on low-dose spiral CT. Low-dose spiral CT uses the smallest scanning area, lowest dose, and minimal X-ray volume to diagnose lesions. Compared to conventional CT scans, it emits less radiation and can more clearly visualize small nodules. However, there is ongoing controversy regarding the increased risk of cancer with low-dose spiral CT. Epidemiological studies have shown that even with just two or three CT scans, the radiation dose can lead to an increased risk of detectable cancer, particularly in children (Computed Tomography - An Increasing Source of Radiation Exposure. N Engl J Med 2007;357:2277-2284 DOI:10.1056 / NEJMra072149).

[0005] Cell-free DNA (cfDNA), also known as circulating DNA (cfDNA), is DNA that exists outside cells in peripheral fluids such as blood and urine. cfDNA derived from tumors is also known as ctDNA (circulating tumor DNA). The value of cfDNA as a new diagnostic marker has been demonstrated in a variety of solid tumors. Existing technologies have developed early cancer screening methods based on whole-genome sequencing or methylation sequencing of cfDNA. Currently, literature (Integrating Genomic Features for Non-Invasive Early Lung Cancer Detection. Nature; Volume 580, Pages 245-251 (2020); (https: / / doi.org / 10.1038 / s41586-020-2140-0)) has reported non-invasive cancer screening based on cfDNA fragments or protein markers in peripheral fluid. However, these methods are almost all used to screen healthy people, aiming to detect cancer patients from the healthy population, but it is difficult to accurately distinguish a tumor in a certain part from other non-cancerous diseases in that area. In addition, for patients with inflammation, the above methods may misjudge them as cancer patients.

[0006] Wenhua Liang et al. proposed a method for diagnosing lung nodules based on cfDNA methylation sequencing data (Non-Invasive Diagnosis of Early-Stage Lung Cancer Using High-Throughput Targeted DNA Methylation Sequencing of Circulating Tumor DNA (ctDNA). Theranostics; 2019; 9(7): 2056-2070. DOI: 10.7150). The results showed that in a validation set of 39 patients with malignant lung nodules and 27 patients with benign lung nodules, the area under the receiver operating characteristic curve (ROC) of this method was 0.816. This article compared malignant and benign lung lesions to understand the DNA methylation characteristics of tissues and established a diagnostic model for benign / malignant nodules. This model was applied to the identification of tumor-specific ctDNA in the plasma of patients with lung nodules, and it had certain sensitivity and specificity for early lung cancer. However, the accuracy of this non-invasive diagnostic method for lung nodules is low and cannot meet clinical requirements.

[0007] Copy number variation (CNV) is caused by genomic rearrangement, generally referring to the increase or decrease in the copy number of large genomic fragments with a length of more than 1kb, mainly manifested as deletions and duplications at the submicroscopic level. CNV is an important component of genomic structural variation (SV). The mutation rate of CNV sites is much higher than that of single nucleotide polymorphisms (SNPs), and it is one of the important pathogenic factors of human diseases. Existing literature (Maternal Malignancies Detected With Noninvasive Prenatal Testing Reply. Jama the Journal of the American Medical Association; 2015 Nov 24; 314(20): 2192-3 [DOI: 10.1001 / jama.2015.12922]) reported that copy number variation was detected using blood cfDNA sequencing data of cancer patients, which means that in the area where CNV occurs, the distribution of DNA fragments in cancer patients is different from the baseline representing healthy people.

[0008] Therefore, it is urgent to develop products that can meet clinical applications and detect the types of human nodules with high accuracy, such as determining whether a patient's lung nodules are benign or malignant through non-invasive methods, thereby avoiding invasive examinations.

[0009] Summary of the Invention

[0010] The main purpose of the present invention is to provide a method for assembling a lung nodule screening model, a screening model, a screening method and a screening device, which can distinguish between malignant tumors in a certain part and other non-cancerous diseases, especially the type of lung nodules.

[0011] To achieve the above objectives, according to a first aspect of the present invention, a method for establishing a lung nodule screening model is proposed, the method comprising the following steps:

[0012] Within the entire range of the human reference genome, a predetermined number of regions with the largest differences in weighted fragment distribution difference values ​​(WFDD) between benign nodule population samples and malignant nodule population samples in the training set are screened as feature data; and a lung nodule screening model is established using the feature data.

[0013] Furthermore, the step of screening a predetermined number of regions with the largest differences in weighted fragment distribution difference (WFDD) values ​​between benign nodule population samples and malignant nodule population samples in the training set as feature data includes:

[0014] Connect the base sequences of all autosomes in the reference genome together, and divide the connected base sequences into a series of windows of fixed length, each window corresponding to a segment of base sequence; calculate the window base depth and weight of each window, where the window base depth is the average value of the depth values ​​of the samples in the training set in the window, and the weight is the square of the variance of the depth values ​​of the samples in the training set in the window. The window depth value is the number of base sequence fragments that can be aligned to the base sequence corresponding to the window in the sequencing data of the sample; calculate the window sample depth of each window for a specified sample in the training set ; Calculate the difference between the window sample depth and the window reference depth; multiply the difference by the weight to obtain the weighted difference of the window; combine an indefinite number of windows to form different areas, and sum the weighted differences of all windows in the specified area to obtain the total weighted difference; perform a numerical transformation on the total weighted difference to obtain the weighted fragment distribution difference (WFDD) of the specified sample in the specified area; and calculate the difference in weighted fragment distribution difference values ​​in each area between the benign nodule population samples and the malignant nodule population samples in the training set, and screen a predetermined number of areas with the largest differences as feature data.

[0015] Furthermore, before calculating the difference between the window sample depth and the window reference depth, the method further includes a step of performing a normalization process on the window reference depth and the window sample depth.

[0016] Furthermore, after the normalization process, the average value of the window reference depth of each window is 0, and the standard deviation is 1.

[0017] Furthermore, screening a predetermined number of regions with the largest differences as feature data includes:

[0018] For a specific region, the WFDD value of each sample in the benign nodule population sample and the malignant nodule population sample is calculated; the average WFDD value of the benign nodule population sample and the average WFDD value of the malignant nodule population sample are calculated respectively; the discrimination value of the specific region is calculated according to the following formula, and a predetermined number of regions with the largest discrimination value are selected as feature data:

[0019]

[0020] Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0021] Furthermore, screening a predetermined number of regions with the largest differences as feature data includes:

[0022] a.) Generate the initial region:

[0023] - Divide the genome into a series of windows;

[0024] -numbering a series of windows by their order;

[0025] - Use a series of window numbers to number the area obtained by combining a series of windows;

[0026] - Combine the n consecutive windows at window i and its downstream 2 j At n windows, another n consecutive windows are combined to form an initial region:

[0027] x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2j n+1,i+2 j n+2,...,i+2 j n+n-1}

[0028] i=1, n, 2n,..., N; j=1, 2,..., 8; i+(2 i+1 )n≤N

[0029] Where n is the number of given continuous windows, and N is the total number of divided windows;

[0030] b.) Combination and division of regions:

[0031] Using genetic algorithms, two parents combine to exchange information and generate offspring; all initial regions are placed in the region pool and randomly selected to generate offspring;

[0032] -The probability that region i is selected as one of the parents is:

[0033]

[0034] Among them, N is the total number of divided windows, t i is the t value of the i-th window;

[0035] -After region x is selected as the first parent, the probability that another region i is selected as the second parent is:

[0036]

[0037] Among them, N is the total number of divided windows, m i is the average value of the window numbers contained in region i; and

[0038] -After selecting the parents, take the union of the windows included in the parents and randomly delete some of them as offspring, and the random selection method is sampling with replacement;

[0039] - After obtaining the offspring, put the offspring into the region pool for the next round of selection; the parents are not deleted in this operation. Among them, the offspring generated by the combination of region P1 and region P2 are:

[0040] child(P1,P2)=P1∪P2-S(p,P1∪P2)

[0041] Where S(p,s) is a subset obtained by extracting elements of proportion p from the set s with replacement;

[0042] - Repeat the process of producing offspring;

[0043] - Calculate the discrimination value of the generated offspring according to the following formula, and select a predetermined number of regions with the largest discrimination value as feature data:

[0044]

[0045] Where: t is the discrimination value of a specific area; and are the mean values ​​of the WFDD values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; n1 and n2 are the number of values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; the conditions are:

[0046] When the sample of the benign nodule population corresponds to the subscript 1, the sample of the malignant nodule population corresponds to the subscript 2; or when the sample of the benign nodule population corresponds to the subscript 2, the sample of the malignant nodule population corresponds to the subscript 1.

[0047] Furthermore, the number n of consecutive windows is 1-100, preferably 5-50, and more preferably 5.

[0048] Furthermore, after selecting the parents, the union of the windows included in the parents is taken and 1%-99%, more preferably 5%-50%, and even more preferably 20% of the windows are randomly deleted as offspring, and sampling with replacement is performed.

[0049] Furthermore, the process of generating offspring is repeated 1 to 1 million times, more preferably 100 to 100,000 times, and even more preferably 300,000 times.

[0050] Furthermore, the connected base sequence is divided into a series of windows according to the length of 100 bp-100 kbp, preferably 10 kbp-50 kbp, more preferably 30 kbp.

[0051] Furthermore, the windows combined to form different regions are continuous or discontinuous.

[0052] Furthermore, the predetermined number of regions is 1-500, more preferably 10-100, and even more preferably 50.

[0053] According to a second aspect of the present invention, a pulmonary nodule screening method is provided, comprising the following steps: calculating the sum of weighted fragment distribution difference (WFDD) values ​​in a predetermined number of selected regions of a sample to be tested to obtain a total WFDD value; inputting the total WFDD value of the sample to be tested into a pulmonary nodule screening model constructed according to the method described in the first aspect of the present invention; and outputting a screening result for the sample to be tested; wherein the predetermined number of selected regions are the same as the predetermined number of regions with the largest difference in weighted fragment distribution difference (WFDD) values ​​between samples from a population of benign nodules and samples from a population of malignant nodules.

[0054] Furthermore, the total WFDD value of the sample to be tested is input into the lung nodule screening model, and the lung nodule screening model determines the lung nodule type of the sample to be tested based on the comparison between the total WFDD value of the sample to be tested and a predetermined threshold.

[0055] Furthermore, the predetermined threshold is obtained by the following method: calculating the total WFDD value of each sample in the training set in a predetermined number of areas with the largest WFDD value difference; calculating the optimal segmentation point based on the total WFDD values ​​of the benign nodule population and the malignant nodule population in the training set, and the optimal segmentation point is the predetermined threshold.

[0056] Furthermore, the optimal split point was calculated using the roc function of the pROC package in the R language analysis platform.

[0057] According to the third aspect of the present invention, the present invention proposes a device for assembling a lung nodule screening model, comprising: a feature data screening module, configured to screen a predetermined number of regions with the largest difference in weighted fragment distribution difference (WFDD) values ​​between benign nodule population samples and malignant nodule population samples in a training set within the entire range of the human reference genome as feature data; and an assembly module, configured to utilize the feature data to assemble a lung nodule screening model.

[0058] Furthermore, the feature data screening module includes: a window division module, which is configured to connect the base sequences of all autosomes in the human reference genome, and divide the connected base sequences into a series of windows according to a fixed length, each window corresponding to a base sequence; a first calculation module, which is configured to calculate the window reference depth and weight of each window, wherein the window reference depth is the average value of the depth value of the samples in the training set in the window, the weight is the square of the variance of the depth value of the samples in the training set in the window, and the window depth value is the number of base sequence fragments that can be compared to the base sequence corresponding to the window in the sequencing data of the sample; a second calculation module, which is configured to calculate the window sample depth of each window of the specified sample in the training set; a third calculation module, which is configured to calculate the window sample depth of each window of the specified sample in the training set; The invention is configured to calculate the difference between the window sample depth and the window reference depth; the fourth calculation module is configured to multiply the difference by the weight to obtain the weighted difference of the window; the fifth calculation module is configured to combine an indefinite number of windows to form different areas, and sum the weighted differences of all windows in the specified area to obtain the total weighted difference; the numerical transformation module is configured to perform a numerical transformation on the total weighted difference to obtain the weighted fragment distribution difference (WFDD) of the specified sample in the specified area; and the sixth calculation module is configured to calculate the difference in weighted fragment distribution difference values ​​in each area between the benign nodule population samples and the malignant nodule population samples in the training set; and the feature data screening submodule is configured to screen a predetermined number of areas with the largest differences as feature data.

[0059] Furthermore, the feature data screening module also includes a uniformity processing module.

[0060] Furthermore, after the normalization process, the average value of the window reference depth of each window is 0, and the standard deviation is 1.

[0061] Furthermore, the feature data screening submodule includes: a first calculation unit configured to calculate the WFDD value of each sample in the benign nodule population sample and the malignant nodule population sample for a specific region; a second calculation unit configured to calculate the average WFDD value of the benign nodule population sample and the average WFDD value of the malignant nodule population sample, respectively; a third calculation unit configured to calculate the discrimination value of the specific region according to the following formula; and a selection unit configured to select a predetermined number of regions with the largest discrimination values ​​as feature data:

[0062]

[0063] Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0064] Furthermore, the feature data screening submodule includes: an initial region generation unit and a region combination splitting unit, wherein the initial region generation unit includes: a window division element, which is configured to divide the genome into a series of windows; a window coding element, which is configured to number a series of windows in the order of a series of windows; a region numbering element, which is configured to number a region obtained by combining a series of windows using a series of window numbers; a window combination element, which is configured to combine the n consecutive windows at window i and the 2 windows downstream thereof. j At n windows, another n consecutive windows are combined to form an initial region:

[0065] x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2 j n+1,i+2 j n+2,...,i+2 j n+n-1},

[0066] i=1, n, 2n,..., N; j=1, 2,..., 8; i+(2 i+1 )n≤N

[0067] Where n is the number of given continuous windows, and N is the total number of divided windows;

[0068] The region combination and splitting unit includes: a first offspring selection component, which is configured to use a genetic algorithm to combine two parents to exchange information and generate offspring; wherein, all initial regions are placed in a region pool, and offspring are randomly selected; wherein:

[0069] -The probability that region i is selected as one of the parents is:

[0070]

[0071] Among them, N is the total number of divided windows, t i is the t value of the i-th window;

[0072] -After region x is selected as the first parent, the probability that another region i is selected as the second parent is:

[0073]

[0074] Among them, N is the total number of divided windows, m i is the average value of the window numbers contained in region i;

[0075] The second offspring selection component is configured to select a parent, take the union of the windows included in the parent, and randomly delete some of the windows as offspring, and the random selection method is sampling with replacement; and

[0076] The third child selection element is set to place the child into the region pool for the next round of selection after obtaining the child. In this operation, the parent is not deleted from the region pool. The child generated by combining regions P1 and P2 is:

[0077] child(P1, P2)=P1∪P2-S(p, P1∪P2),

[0078] where S(p,s) is a subset obtained by extracting proportion p of elements from the set s with replacement; and

[0079] A progeny repetitive generation element is configured to repetitively generate a progeny process;

[0080] - Calculate the discrimination value of the generated offspring according to the following formula, and select a predetermined number of regions with the largest discrimination value as feature data:

[0081]

[0082] Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0083] Furthermore, the number n of consecutive windows is 1-100, preferably 5-50, and more preferably 5.

[0084] Furthermore, after selecting the parents, the union of the windows included in the parents is taken and 1%-99%, more preferably 5%-50%, and even more preferably 20% of the windows are randomly deleted as offspring, and sampling with replacement is performed.

[0085] Furthermore, the process of generating offspring is repeated 1 to 1 million times, more preferably 100 to 100,000 times, and even more preferably 300,000 times.

[0086] Furthermore, the connected base sequence is divided into a series of windows according to the length of 100 bp-100 kbp, preferably 10 kbp-50 kbp, more preferably 30 kbp.

[0087] Furthermore, the windows combined to form different regions are continuous or discontinuous.

[0088] Furthermore, the predetermined number of regions is 1-500, more preferably 10-100, and even more preferably 50. According to the fourth aspect of the present invention, the present invention proposes a lung nodule screening device, comprising: a first calculation module, configured to calculate the sum of weighted fragment distribution difference values ​​of the sample to be tested in a selected predetermined number of regions to obtain a total WFDD value; an input module, configured to input the total WFDD value of the sample to be tested into the lung nodule screening model constructed by the construction device according to the third aspect of the present invention; and an output module, configured to output the screening result of the sample to be tested; wherein the selected predetermined number of regions are the same as the predetermined number of regions with the largest difference in weighted fragment distribution difference values ​​(WFDD) between the benign nodule population samples and the malignant nodule population samples.

[0089] Furthermore, the input module includes: an input unit, which is configured to input the total WFDD value of the sample to be tested into the lung nodule screening model; and a judgment unit, which is configured to judge the lung nodule type of the sample to be tested based on the comparison of the total WFDD value of the sample to be tested with a predetermined threshold by the lung nodule screening model.

[0090] Furthermore, the screening device also includes a predetermined threshold acquisition module, which includes: a first calculation unit, which is configured to calculate the total WFDD value of a predetermined number of areas with the largest WFDD value difference for each sample in the training set; and a second calculation unit, which is configured to calculate the optimal segmentation point based on the total WFDD values ​​of the benign nodule population and the malignant nodule population in the training set, and the optimal segmentation point is the predetermined threshold.

[0091] Furthermore, the optimal split point was calculated using the roc function of the pROC package in the R language analysis platform.

[0092] According to the fifth aspect of the present invention, the present invention proposes a computer-readable storage medium, which includes a stored program. When the program is running, the program executes the assembly method according to the first aspect of the present invention or the lung nodule screening method according to the second aspect of the present invention.

[0093] According to the sixth aspect of the present invention, the present invention proposes a processor, which is used to run a program, wherein when the program is running, it executes the method for building a lung nodule screening model according to the first aspect of the present invention or the lung nodule screening method according to the second aspect of the present invention.

[0094] Using the technical solution of this invention, a method has been developed to differentiate between malignant tumors and other non-cancerous conditions (such as nodules) in a specific area. Results show that the method is significantly more effective than existing CT scans and approaches the performance of invasive tissue biopsies. This method has resulted in a product that can non-invasively detect the type (benign / malignant) of human nodules. For example, a blood test can be used to determine whether a patient's lung nodule is malignant, thus avoiding invasive examinations. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0096] FIG1 shows a flow chart of a method for constructing a lung nodule screening model according to the present invention.

[0097] Figure 2 illustrates a method for calculating the weighted fragment distribution difference (WFDD) of a pattern according to the present invention. A shows the method for calculating the weighted difference of a sample in a specified window; B shows the method for calculating the cumulative value of the i-th window; and C shows the method for calculating the WFDD of a sample in a specified area.

[0098] FIG. 3 shows an example of a benign correlation distribution pattern obtained by modeling according to an embodiment of the present invention.

[0099] FIG. 4 shows an example of a malignancy-related distribution pattern obtained by modeling according to an embodiment of the present invention.

[0100] FIG5 shows experimental results obtained by modeling and predicting validation set samples according to an embodiment of the present invention.

[0101] FIG6 shows a flow chart of the lung nodule screening method according to the present invention.

[0102] FIG7 shows a device for assembling a lung nodule screening model according to the present invention.

[0103] FIG8 shows a feature data screening module in the device for building a lung nodule screening model according to the present invention.

[0104] FIG9 shows a characteristic data screening submodule in a characteristic data screening module in a device for building a lung nodule screening model according to the present invention.

[0105] FIG10 shows a characteristic data screening submodule in a characteristic data screening module in a device for building a lung nodule screening model according to the present invention.

[0106] FIG11 shows a lung nodule screening device according to the present invention.

[0107] FIG12 shows an input module of the lung nodule screening device according to the present invention. DETAILED DESCRIPTION

[0108] The following description is presented to enable a person of ordinary skill in the art to make and use the various embodiments. The descriptions of specific devices, techniques, and applications are provided as examples only. Various modifications to the examples described herein will be apparent to a person of ordinary skill in the art, and the general principles defined herein may be applied to other examples and applications without departing from the scope of the various embodiments. Therefore, the various embodiments are not intended to be limited to the examples described and illustrated herein, but are consistent with the scope of the claims. It should be noted that the embodiments in this application and the features in the embodiments may be combined with each other unless they conflict. The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0109] As described in the background technology section, the application value of free DNA as a new diagnostic marker has been confirmed in a variety of solid tumors. Therefore, more and more studies have developed early screening methods for cancer through whole genome sequencing or methylation sequencing based on free DNA. Although these methods have proliferated, there are still many problems. First, the methods of the prior art can almost only detect cancer patients from healthy people, and it is difficult to ensure that a certain part of the tumor can be accurately distinguished from other non-cancer diseases in that part, especially for patients with inflammation, who may be misjudged as cancer patients by the above methods. Secondly, the accuracy of the prior art non-invasive diagnosis method of lung nodules based on tumor-specific ctDNA in patient plasma is low and cannot meet clinical requirements.

[0110] Due to the above problems, traditional early screening programs need to be modified to be able to distinguish between certain malignant tumors and other non-cancer types of diseases.

[0111] Now referring to FIG1 , FIG1 shows a flow chart of a method for constructing a lung nodule screening model according to the present invention. The method for constructing a lung nodule screening model according to an embodiment of the present invention constructs a model based on the distribution characteristics of DNA fragment sequence reads in the second generation sequencing data on the reference genome (reference), thereby distinguishing different types of nodules (benign / malignant). Existing literature (Maternal Malignancies Detected With Noninvasive Prenatal Testing Reply. Jama the Journal of the American Medical Association; 2015 Nov 24; 314(20): 2192-3 [DOI: 10.1001 / jama.2015.12922]) reports that copy number variation was detected using blood cfDNA sequencing data of cancer patients, which means that in the area where CNV occurs, the distribution of DNA fragments in cancer patients is different from the baseline representing healthy people. We assume that patients with benign / malignant nodules have differences in DNA fragment distribution in certain areas. Therefore, to better characterize the distribution of cfDNA fragments within a region, we proposed the concept of weighted fragment distribution difference (WFDD). CNV only focuses on the difference between the total number of fragments in a sample within a region and the baseline, while WFDD focuses on the details of the fragment distribution differences across the region. We developed a lung nodule screening model to describe this detail. Specifically, the method for building the lung nodule screening model includes:

[0112] In Figure 1, the sequencing data for all autosomes in the reference genome were first concatenated and then partitioned into a series of windows of fixed length. Here, the fixed length ranged from 100 bp to 100 kbp, preferably from 10 kbp to 50 kbp, and more preferably from 30 kbp. Because the partitioning was performed by concatenating the sequencing data for the autosomes, windows that spanned chromosomes were possible. To eliminate potential interference from sex factors, sex chromosome data were not used.

[0113] 2 to 6 of FIG1 show a method for calculating the weighted fragment distribution difference (WFDD) according to an embodiment of the present invention. In FIG1 2, the window reference depth and weight are calculated for each window: the average value of the depth values ​​of the samples of the training set in the window (each sample provides a depth value) is used as the window reference depth of the window, and the square of the variance of the depth values ​​of these samples is used as the weight of the window. Thereafter, the window sample depth in each window is calculated for the specified sample in the training set. In a preferred embodiment, the window reference depth of the windows in the specified area is normalized to obtain the normalized window reference depth, so that the average value of the reference depth of each window is 0 and the standard deviation is 1. The reason for this is that WFDD only focuses on the depth difference at the window level, and in order to eliminate the influence of the total depth difference of different samples in the region on WFDD, it is necessary to normalize the reference depth of the windows contained in a region before calculating WFDD to obtain the normalized window reference depth. The same normalization operation is also performed on the depth of the specified samples in these windows to obtain the normalized window sample depth. In 3 of FIG. 1 , the difference between the window sample depth and the window reference depth is calculated. In a preferred embodiment, the difference between the normalized window sample depth and the normalized window reference depth is calculated. In 4 of FIG. 1 , the difference is multiplied by a weight to obtain a weighted difference of the window. In a preferred embodiment in which normalization is performed, the difference between the normalized window sample depth and the normalized window reference depth is multiplied by a weight. Obviously, a window with a greater fluctuation in different sample depths will have a greater weight. Multiplying the difference between the depth of the sample in the window and the reference depth by its weight will amplify the difference, that is, amplify the distribution difference signal. Subsequently, in 5 of FIG. 1 , an indefinite number of windows are combined to form different regions, and the weighted differences of all windows in the specified region are summed; the sum of the weighted differences (that is, the last accumulated value of the sum) is numerically transformed to obtain the weighted fragment distribution difference (WFDD) of the specified sample in the specified region. In 6 of Figure 1 , the difference in weighted segment distribution difference values ​​for each region of the benign nodule population samples and the malignant nodule population samples in the training set is calculated, and a predetermined number of regions with the largest differences are selected as feature data. In one embodiment, the selected regions are 1-500, preferably 10-100, and more preferably 50; the feature data is then used to construct a pulmonary nodule screening model.

[0114] Now referring to Figure 2, an example of a method for calculating pattern-weighted segment distribution differences according to an embodiment of the present invention is shown. Figure 2A shows an example of a method for calculating the weighted difference of a sample in a specified window. In this example, the average of the depth values ​​of the samples in the training set in the window (each sample provides a depth value) is used as the window base depth of the window, and the square of the variance of the depth values ​​of these samples is used as the weight of the window. The window sample depth in each window is calculated for the specified sample in the training set. The window base depth of the windows in the specified area is normalized to obtain the normalized window base depth, so that the average value of the window base depths in the same area is 0 and the standard deviation is 1. The depths of the specified samples in these windows are also normalized in the same way to obtain the normalized window sample depth. The difference between the normalized window sample depth and the normalized window base depth is calculated. The difference is multiplied by the weight to obtain the weighted difference of the window. Figure 2B shows an example of a method for calculating the cumulative value of the i-th window. Here, the indefinite number of windows that are combined to form different regions can be non-continuous (non-adjacent). And it should be noted that when calculating the WFDD of a sample in a region, we only normalize the sample depth and the reference depth based on the windows of this region, that is, we only require the average value of the reference depth of each window in the same region to be 0 and the standard deviation to be 1, rather than based on all windows on the genome. In C of Figure 2, the formula for numerical transformation of the last cumulative value of the sum is shown: WDFF = f(x) or WDFF = -1*f(x), and the forms of f(x) in different cases are specifically shown in Figures 3 and 4 respectively.

[0115] More simply (not explained from the perspective of cumulative value), in one embodiment, the empirical formula for calculating the WFDD of a sample in a specified area is:

[0116]

[0117] When the region belongs to the benign correlation region: WDFF=-1*f(x)

[0118] When the region belongs to the malignant related region: WDFF=f(x)

[0119] in,

[0120] x′ i The normalized window sample depth of the specified sample of the training set in the i-th window,

[0121] is the normalized base depth of the window for the i-th window, and

[0122] σ iThe variance of the depth value of the specified sample in the training set in the i-th window.

[0123] Now refer to Figure 3, which shows an example of a benign correlation distribution pattern obtained by modeling according to an embodiment of the present invention. For a given region, if the WFDD fluctuation of the samples of the benign pulmonary nodule population is larger than that of the malignant pulmonary nodule population, we call the fragment distribution pattern of the region a benign correlation pattern. In Figure 3, the region where the pattern is located includes 53 windows, and each broken line represents a sample. The left figure shows the weighted difference of cfDNA samples (40 in total) of 20 random benign / malignant pulmonary nodules patients on these windows; the middle figure shows the cumulative value of the weighted difference for each window; the right figure shows the correlation function and result (box plot) of the last cumulative value of the sample converted to the benign correlation pattern to obtain WFDD.

[0124] Now refer to Figure 4, which shows an example of a malignant correlation distribution pattern obtained by modeling according to an embodiment of the present invention. If the WFDD fluctuation of the samples of the malignant pulmonary nodule population is larger than that of the benign pulmonary nodule population, we call the fragment distribution pattern of this area a malignant correlation pattern. In Figure 4, the area where the pattern is located includes 231 windows. The meaning of each part in the figure is the same as (B), but the samples are not completely consistent. The left figure shows the weighted difference of cfDNA samples of 20 randomly selected patients with benign / malignant pulmonary nodules (40 in total) on these windows; the middle figure shows the cumulative value of the weighted difference for each window; the right figure shows the correlation function and result (box plot) of the last cumulative value of the sample in the benign correlation pattern to obtain the WFDD.

[0125] In one embodiment, a normalization operation may be performed on a set of values ​​(such as the depth values ​​of a sample in multiple windows) using the following formula:

[0126]

[0127] Where S is the standard deviation of the set of values, and is its average value.

[0128] As described above, we selected several regions with the largest difference in WFDD values ​​between the two types of samples to build the model. In order to evaluate the difference in WFDD between the two types of samples in a specific region, we calculated the discrimination value of the region based on the WFDD values ​​of the two groups of samples to evaluate its ability to distinguish the two types of samples. Specifically: for a specific region, calculate the WFDD value of each sample in the benign nodule population sample and the malignant nodule population sample; calculate the WFDD value of the benign nodule population sample separately.

[0129] The average value of the WFDD value and the average value of the WFDD value of the malignant nodule population sample; calculate the discrimination value of the specific area and select a predetermined number of areas with the largest discrimination value as feature data. In one embodiment, the formula for calculating the discrimination value of the specific area is:

[0130]

[0131] in:

[0132] t is the discrimination value of a specific area; and where n1 is the mean of the WFDD values ​​for samples from a population of benign nodules and n2 is the number of values ​​for samples from a population of malignant nodules; S1 is the standard deviation of the values ​​for samples from a population of benign nodules and malignant nodules, respectively. The condition is that when the benign nodule population corresponds to the subscript 1, the malignant nodule population corresponds to the subscript 2; or when the benign nodule population corresponds to the subscript 2, the malignant nodule population corresponds to the subscript 1. A higher t-value indicates that the WFDD values ​​of the two groups of samples have a greater difference in this region, that is, the region has a higher degree of discrimination.

[0133] Based on the divided windows, parts of them are combined into different regions. In theory, the number of regions that can be combined will exceed the scope that existing computers can handle. In order to more effectively find potential high-discrimination windows, in one embodiment, an improved genetic algorithm can be used to search. The improved genetic algorithm randomly merges and splits a series of regions (initial regions) obtained by a simple strategy, and guides the generation of regions with larger t values. Specifically, the steps of searching for potential high-discrimination windows include:

[0134] a.) Generate the initial region:

[0135] Divide the genome into a series of windows: Assume that we have divided the genome into N windows, each window is represented by its sequence number (such as 1, 2, ..., N), and the region obtained by combining these windows is represented by a series of window numbers (such as {1, 2, 3, 10, 11, 12}). We divide the n consecutive windows at window i and the 2 windows downstream of it into j At n windows, another n consecutive windows are combined to obtain a total of 2n windows combined into an initial region. That is:

[0136] x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2 j n+1,i+2 j n+2,...,i+2 j n+n-1}

[0137] i=1, n, 2n,..., N; j=1, 2,..., 8; i+(2 i+1 )n≤N

[0138] Where n is the number of given consecutive windows, ranging from 1 to 100, preferably from 5 to 50, and more preferably 5, and N is the total number of divided windows. Here, requiring a certain distance between two consecutive window sequences is to enable the region to span long distances.

[0139] b.) Combination and division of regions:

[0140] Using a genetic algorithm, two parents are combined to exchange information and produce offspring. All initial regions are placed in a region pool and randomly selected to produce offspring. Here, the probability that region i is selected as one of the parents is:

[0141]

[0142] Where N is the total number of divided windows.

[0143] When region x is selected as the first parent, the probability that another region i is selected as the second parent is:

[0144]

[0145] Among them, N is the total number of divided windows, m i is the average value of the window numbers contained in region i. Here, set |m x -m i The | item is to give priority to selecting the region with a close distance as the other parent.

[0146] After selecting the parents, the union of the windows included in the parents is taken and some of them are randomly deleted as offspring. In one embodiment, the range of randomly deleted windows is 1%-99%, preferably 5%-50%, and more preferably 20%, with replacement sampling. After the offspring are obtained, they are placed into the region pool for the next round of selection; the parents are not deleted in this operation. The offspring generated by the combination of region P1 and region P2 are:

[0147] child(P1,P2)=P1∪P2-S(p,P1∪P2)

[0148] Here, S(p,s) is a subset obtained by extracting elements of proportion p from the set s with replacement.

[0149] The process of generating offspring is repeated continuously. In one embodiment, the repetition range is 1 to 1 million times, in a preferred embodiment, the repetition range is 100 to 100,000 times, and in a more preferred embodiment, the repetition range is 300,000 times. Finally, several regions with the highest discrimination are selected as features to form a model and make predictions.

[0150] Reference is now made to Figure 6, which illustrates a flow chart of a pulmonary nodule screening method according to the present invention. The pulmonary nodule screening method includes: selecting a certain number of regions. To determine the type of a particular sample based on these regions, the WFDD values ​​of the sample in these regions are calculated and summed to obtain a total WFDD value. This total WFDD value is then compared with a predetermined threshold, and the type of the sample is determined based on whether the total WFDD value is greater than or less than the threshold.

[0151] As used herein, the threshold is calculated based on the training set samples. First, the total WFDD value for each training set sample for these regions is calculated. Then, based on the total WFDD values ​​for the benign and malignant nodule populations in the training set, the optimal split point is calculated. This split point serves as the desired threshold. The roc function (from the pROC package in R) can be used to calculate the optimal split point.

[0152] Reference is now made to FIG7 , which illustrates an apparatus for constructing a lung nodule screening model according to the present invention. Specifically, the apparatus comprises: a feature data screening module configured to screen, within the entire human reference genome, a predetermined number of regions with the largest differences in weighted fragment distribution difference (WFDD) values ​​between benign nodule samples and malignant nodule samples in a training set as feature data; and a construction module configured to construct a lung nodule screening model using the feature data.

[0153] Now refer to Figure 8, which shows the feature data screening module in the assembly device of the lung nodule screening model according to the present invention. Specifically, the feature data screening module of the present invention includes: a window division module, which is configured to connect the base sequences of all autosomes in the human reference genome together, and divide the connected base sequences into a series of windows according to a fixed length, each window corresponding to a base sequence; a first calculation module, which is configured to calculate the window reference depth and weight of each window, wherein the window reference depth is the average value of the depth value of the sample in the training set in the window, the weight is the square of the variance of the depth value of the sample in the training set in the window, and the depth value of the window is the number of base sequence fragments that can be compared to the base sequence corresponding to the window in the sequencing data of the sample; a second calculation module, which is configured to calculate the window sample depth of each window of the specified sample in the training set; a third calculation module, is configured to calculate the difference between the window sample depth and the window reference depth; a fourth calculation module is configured to multiply the difference by a weight to obtain a weighted difference of the window; a fifth calculation module is configured to combine an indefinite number of windows to form different regions, and sum the weighted differences of all windows in the specified region to obtain a weighted difference sum; a numerical transformation module is configured to perform a numerical transformation on the weighted difference sum to obtain a weighted fragment distribution difference (WFDD) of the specified sample in the specified region; and a sixth calculation module is configured to calculate the difference in weighted fragment distribution difference values ​​between the benign nodule population sample and the malignant nodule population sample in each region in the training set; and a feature data screening submodule is configured to screen a predetermined number of regions with the largest difference as feature data. Optionally, as indicated by a dotted line in FIG8 , the feature data screening module further includes a normalization processing module, which is configured to perform normalization processing on the window reference depth and the window sample depth. Optionally, the average value of the window reference depth of each window after normalization is 0 and the standard deviation is 1.

[0154] Now refer to Figure 9, which shows a feature data screening submodule in the feature data screening module in the assembly device of the lung nodule screening model according to the present invention. Specifically, the feature data screening submodule of the present invention includes: a first calculation unit, which is configured to calculate the WFDD value of each sample in the benign nodule population sample and the malignant nodule population sample for a specific area; a second calculation unit, which is configured to calculate the average value of the WFDD values ​​of the benign nodule population sample and the average value of the WFDD values ​​of the malignant nodule population sample respectively; a third calculation unit, which is configured to calculate the discrimination value of the specific area according to the following formula, and a selection unit, which is configured to select a predetermined number of areas with the largest discrimination values ​​as feature data: Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0155] Now refer to Figure 10, which shows a feature data screening submodule in the feature data screening module in the assembly device of the lung nodule screening model according to the present invention. The feature data screening submodule includes: an initial region generation unit and a region combination decomposition unit, wherein the initial region generation unit includes: a window division element, which is configured to divide the genome into a series of windows; a window encoding element, which is configured to number the series of windows in the order of the series of windows; a region numbering element, which is configured to use a series of window numbers to number the regions obtained by combining the series of windows; a window combination element, which is configured to combine the n consecutive windows at window i and the 2 downstream windows thereof. j At n windows, another n consecutive windows are combined to form an initial region: x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2 j n+1,i+2 j n+2,...,i+2 j n+n-1},;i=1,n,2n,...,N;j=1,2,...,8;i+(2 i+1 )n≤N; where n is the number of given continuous windows and N is the total number of divided windows. The region combination and splitting unit includes: a first offspring selection component, which is configured to use a genetic algorithm to combine two parents to exchange information and generate offspring; wherein all initial regions are placed in a region pool and offspring are randomly selected; wherein: - the probability of region i being selected as one of the parents is: Among them, N is the total number of divided windows, t i is the t value of the i-th window; -After region x is selected as the first parent, the probability that another region i is selected as the second parent is: Among them, N is the total number of divided windows, m iis the average value of the window numbers contained in region i; the second child selection element is set to select the parents, take the union of the windows included in the parents and randomly delete some of the windows as children, and the random selection method is replacement sampling; and the third child selection element is set to, after obtaining the children, put the children into the region pool for the next round of selection; in this operation, the parents are not deleted from the region pool, wherein the children generated by the combination of region P1 and region P2 are: child(P1, P2) = P1∪P2-S(p, P1∪P2); wherein S(p,s) is the subset obtained by extracting a proportion p of elements from the set s with replacement; and the child regeneration element is set to repeat the process of generating children; - the discrimination value of the generated children is calculated according to the following formula, and a predetermined number of regions with the largest discrimination value are selected as feature data: Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0156] Reference is now made to FIG11 , which illustrates a pulmonary nodule screening device according to the present invention. Specifically, the screening device comprises: a first calculation module configured to calculate the sum of weighted fragment distribution difference values ​​of a sample to be tested in a predetermined number of selected regions to obtain a total WFDD value; an input module configured to input the total WFDD value of the sample to be tested into the pulmonary nodule screening model of the present invention; and an output module configured to output the screening result of the sample to be tested; wherein the predetermined number of selected regions is the same as the predetermined number of regions with the largest difference in weighted fragment distribution difference values ​​(WFDD) between the benign nodule population sample and the malignant nodule population sample.

[0157] Now refer to Figure 12, which shows an input module of the lung nodule screening device according to the present invention. Specifically, the input module includes: an input unit, which is configured to input the total WFDD value of the sample to be tested into the lung nodule screening model; and a judgment unit, which is configured to compare the lung nodule screening model with the total WFDD value of the sample to be tested and the predetermined threshold value to judge the type of lung nodule of the sample to be tested. As shown in the dotted box in Figure 12, in some embodiments, the screening device also includes a predetermined threshold acquisition module, which includes: a first calculation unit, which is configured to calculate the total WFDD value of a predetermined number of areas with the largest WFDD value difference for each sample in the training set; and a second calculation unit, which is configured to calculate the optimal segmentation point based on the total WFDD value of the benign nodule population and the malignant nodule population in the training set, and the optimal segmentation point is the predetermined threshold value; more preferably, the optimal segmentation point is calculated using the roc function of the pROC package from the R language analysis platform.

[0158] An embodiment of the present invention provides a computer-readable storage medium on which a program is stored. When the program is run, the program executes the method for assembling a lung nodule screening model of the present invention or the lung nodule screening method of the present invention.

[0159] An embodiment of the present invention provides a processor configured to execute a program, wherein when the program is executed, the method for constructing a pulmonary nodule screening model or the pulmonary nodule screening method of the present invention is executed. The present invention is further described in detail below with reference to specific embodiments, which are not to be construed as limiting the scope of the present invention.

[0160] Example 1: Establishing a lung nodule screening model

[0161] Blood samples were collected from 639 patients with untreated pulmonary nodules, including 484 patients with malignant nodules (85% of whom were stage I) and 155 patients with benign nodules. The malignant pulmonary nodules included only patients with non-small cell lung cancer. All patients were anonymized and consented to have their samples used for clinical research.

[0162] Collect whole blood in EDTA tubes and process immediately. If immediate processing is not possible, store at 4°C for no more than 1 day. Centrifuge at 1600 g for 10 minutes at 4°C to separate plasma from cellular components. Plasma should be further centrifuged at 16,000 g for 10 minutes at 4°C to remove any cellular residues and stored at -80°C until use.

[0163] cfDNA was extracted from 200ul of plasma using the MagPure Circulating DNA KF Kit (Magen). The obtained cfDNA was subjected to standard next-generation sequencing library construction using the MGIEasy Cell-free DNA Library Prep Set (MGI). Sequencing was performed using the MGISEQ-2000 platform, and whole-genome sequencing data with a sequencing depth of approximately 0.5-1.0x were obtained for each sample.

[0164] Use Sentieon software to process sequencing data (including comparison, sorting and de-duplication), and use software readCounter to count the reading value of each sample to the quantity of every 1kbp scope region on autosome, i.e. the reading value depth of every kb, then every 30 depth values ​​are added up to obtain the depth of the scope of 30kbp in length. In the present embodiment, 30kbp is the length of a window. In order to calculate the depth of all sites as much as possible, reduce the information loss caused by the window abandoning insufficient length, do not directly allow readCounter to be counted in units of 30kbp. All carry out the same operation to each sample, obtain a depth value matrix with a length of 95833 (number of windows) and a width of 639 (number of samples).

[0165] We randomly selected 31 samples from the malignant pulmonary nodule sample set and 30 samples from the benign nodule sample set as validation sets, respectively. We then used the remaining samples as training sets to extract relevant feature regions. In this example, we extracted the 10 most discriminative benign-related features and the 10 most discriminative malignant-related features to construct a model and predict the validation set samples. The prediction results are shown in Figure 5.

[0166] The results showed that the AUC of the model was approximately 0.954 (95% CI: 0.908-1.000) when distinguishing the two types of samples in the validation set, indicating that the model has excellent performance in distinguishing the type of lung nodules in patients based on their blood cfDNA.

[0167] In addition, according to the following formula:

[0168] Specificity = number of true negatives / (number of true negatives + number of false positives)*100% (the ratio of correctly identifying non-patients);

[0169] Sensitivity = number of true positives / (number of true positives + number of false negatives)*100% (ratio of correctly diagnosed patients);

[0170] The specificity and sensitivity of the method of the present invention were both about 0.8, and the result of specificity×sensitivity was about 0.64.

[0171] In addition, CT scans were also performed on the 639 patients with untreated lung nodules. The specificity of the CT scan-based approach was calculated to be approximately 0.3, meaning that approximately 70% of benign nodules were considered malignant or could not be diagnosed. The sensitivity of the CT scan-based approach was approximately 0.93, with the result of specificity multiplied by sensitivity being approximately 0.28.

[0172] It can be seen that compared with the CT scanning method of the prior art, the method of the present invention has significantly higher specificity and comparable sensitivity, thereby obtaining a significantly higher specificity × sensitivity result, indicating that the model has excellent performance in distinguishing the type of lung nodules in patients based on their blood cfDNA.

[0173] Example 2: Obtaining a lung nodule screening method

[0174] This embodiment provides a lung nodule screening method, comprising the following steps:

[0175] Calculate the sum of the weighted fragment distribution difference values ​​of the sample to be tested in a predetermined number of selected areas to obtain a total WFDD value; input the total WFDD value of the sample to be tested into the lung nodule screening model constructed by the construction method according to the first aspect of the present invention, or input it into the lung nodule screening model according to the second aspect of the present invention for screening; output the screening result of the sample to be tested; wherein the selected predetermined number of areas are the same as the predetermined number of areas with the largest difference in weighted fragment distribution difference values ​​(WFDD) between the benign nodule population samples and the malignant nodule population samples.

[0176] Optionally, the lung nodule screening method further includes: inputting the total WFDD value of the sample to be tested into a lung nodule screening model, and the lung nodule screening model determines the lung nodule type of the sample to be tested based on the comparison of the total WFDD value of the sample to be tested with a predetermined threshold.

[0177] Optionally, the predetermined threshold is obtained by the following method: calculating the total WFDD value of a predetermined number of regions with the largest WFDD value difference for each sample in the training set; calculating the optimal segmentation point based on the total WFDD values ​​of the benign nodule population and the malignant nodule population in the training set, and the optimal segmentation point is the predetermined threshold.

[0178] Alternatively, the optimal split point is calculated using the roc function from the pROC package in the R language analysis platform.

[0179] Example 3: Device for obtaining a lung nodule screening model

[0180] The present invention proposes a device for assembling a lung nodule screening model, which includes: a feature data screening module, which is configured to screen a predetermined number of regions with the largest differences in weighted fragment distribution difference (WFDD) values ​​between benign nodule population samples and malignant nodule population samples in a training set within the entire range of the human reference genome as feature data; and a assembling module, which is configured to use the feature data to assemble a lung nodule screening model.

[0181] Optionally, the feature data screening module includes: a window division module, which is configured to connect the base sequences of all autosomes in the human reference genome, and divide the connected base sequences into a series of windows according to a fixed length, each window corresponding to a base sequence; a first calculation module, which is configured to calculate the window reference depth and weight of each window, wherein the window reference depth is the average value of the depth value of the samples in the training set in the window, the weight is the square of the variance of the depth value of the samples in the training set in the window, and the depth value of the window is the number of base sequence fragments that can be compared to the base sequence corresponding to the window in the sequencing data of the sample; a second calculation module, which is configured to calculate the window sample depth of each window of the specified sample in the training set; a third calculation module, which is configured to calculate the window sample depth of each window of the specified sample in the training set; The invention is configured to calculate the difference between the window sample depth and the window reference depth; the fourth calculation module is configured to multiply the difference by the weight to obtain the weighted difference of the window; the fifth calculation module is configured to combine an indefinite number of windows to form different areas, and sum the weighted differences of all windows in the specified area to obtain the total weighted difference; the numerical transformation module is configured to perform a numerical transformation on the total weighted difference to obtain the weighted fragment distribution difference (WFDD) of the specified sample in the specified area; and the sixth calculation module is configured to calculate the difference in weighted fragment distribution difference values ​​in each area between the benign nodule population samples and the malignant nodule population samples in the training set; and the feature data screening submodule is configured to screen a predetermined number of areas with the largest differences as feature data.

[0182] Optionally, the feature data screening module also includes a uniformization processing module.

[0183] Optionally, after the normalization process, the average value of the window reference depth of each window is 0, and the standard deviation is 1.

[0184] Optionally, the feature data screening submodule includes: a first calculation unit, configured to calculate the WFDD value of each sample in the benign nodule population samples and the malignant nodule population samples for a specific region; a second calculation unit, configured to calculate the average WFDD values ​​of the benign nodule population samples and the average WFDD values ​​of the malignant nodule population samples, respectively; a third calculation unit, configured to calculate the discrimination value of the specific region according to the following formula; and a selection unit, configured to select a predetermined number of regions with the largest discrimination values ​​as feature data:

[0185]

[0186] Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0187] Optionally, the feature data screening submodule includes: an initial region generation unit and a region combination splitting unit, wherein the initial region generation unit includes: a window division element, which is configured to divide the genome into a series of windows; a window coding element, which is configured to number the series of windows in the order of the series of windows; a region numbering element, which is configured to number the regions obtained by combining the series of windows with the series of window numbers; a window combination element, which is configured to combine the n consecutive windows at the window i and the 2 windows downstream thereof. j At n windows, another n consecutive windows are combined to form an initial region: x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2 j n+1,i+2 j n+2,...,i+2 j n+n-1}; i=1, n, 2n,..., N; j=1, 2,..., 8; i+(2 i+1 )n≤N; wherein n is the number of given continuous windows, and N is the total number of divided windows; the region combination and splitting unit includes: a first offspring selection component, which is configured to use a genetic algorithm, and two parents are combined to exchange information and generate offspring; wherein all initial regions are placed in a region pool, and offspring are randomly selected; wherein:

[0188] -The probability that region i is selected as one of the parents is: Among them, N is the total number of divided windows, t i is the t value of the i-th window;

[0189] -After region x is selected as the first parent, the probability that another region i is selected as the second parent is:

[0190] Among them, N is the total number of divided windows, mi is the average value of the window numbers contained in region i;

[0191] The second child selection component is configured to select the parent, take the union of the windows included in the parent, and randomly delete some of the windows as the child, and the random selection method is replacement sampling; and the third child selection component is configured to, after obtaining the child, place the child into the region pool for the next round of selection; in this operation, the parent is not deleted from the region pool, wherein the child generated by the combination of region P1 and region P2 is: child(P1, P2) = P1∪P2-S(p, P1∪P2); where S(p, s) is the subset obtained by sampling a proportion p of elements from the set s with replacement; and the child regeneration component is configured to repeat the process of generating children;

[0192] - Calculate the discrimination value of the generated offspring according to the following formula, and select a predetermined number of regions with the largest discrimination value as feature data:

[0193] Where: t is the discrimination value of a specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; the condition is: when the benign nodule population samples correspond to the subscript 1, the malignant nodule population samples correspond to the subscript 2; or when the benign nodule population samples correspond to the subscript 2, the malignant nodule population samples correspond to the subscript 1.

[0194] In the assembly device of the present invention, the number n of consecutive windows is 1-100, preferably 5-50, and more preferably 5. Preferably, after selecting the parents, the union of the windows included in the parents is taken, and 1%-99%, more preferably 5%-50%, and even more preferably 20% of the windows are randomly deleted as offspring, and sampling with replacement is performed. Preferably, the offspring generation process is repeated 1 to 1 million times, more preferably 100 to 100,000 times, and even more preferably 300,000 times. The connected base sequence is divided into a series of windows according to the length of 100 bp to 100 kbp, preferably 10 kbp to 50 kbp, and even more preferably 30 kbp. Preferably, the windows combined to form different regions are continuous or discontinuous. Preferably, the predetermined number of regions is 1-500, more preferably 10-100, and even more preferably 50.

[0195] Example 4: Obtaining a lung nodule screening device

[0196] The present invention provides a pulmonary nodule screening device, comprising: a first calculation module configured to calculate the sum of weighted fragment distribution difference values ​​of a sample to be tested in a predetermined number of selected regions to obtain a total WFDD value; an input module configured to input the total WFDD value of the sample to be tested into a pulmonary nodule screening model constructed by a construction device of the present invention; and an output module configured to output the screening result of the sample to be tested. The predetermined number of selected regions is the same as the predetermined number of regions with the largest difference in weighted fragment distribution difference (WFDD) values ​​between samples from a population of benign nodules and samples from a population of malignant nodules.

[0197] Optionally, the input module includes: an input unit, configured to input the total WFDD value of the sample to be tested into the lung nodule screening model; and a judgment unit, configured to judge the lung nodule type of the sample to be tested by comparing the total WFDD value of the sample to be tested with a predetermined threshold value.

[0198] Optionally, the screening device also includes a predetermined threshold acquisition module, which includes: a first calculation unit, which is configured to calculate the total WFDD value of a predetermined number of areas with the largest WFDD value difference for each sample in the training set; and a second calculation unit, which is configured to calculate the optimal segmentation point based on the total WFDD values ​​of the benign nodule population and the malignant nodule population in the training set, and the optimal segmentation point is the predetermined threshold.

[0199] Alternatively, the optimal split point is calculated using the roc function from the pROC package in the R language analysis platform.

[0200] In addition, it should be noted that, after simple changes, the lung nodule screening model assembly method, screening model, screening method and screening device of the present invention can also be applied to methylation sequencing data, RNA sequencing data and proteomics-related data.

[0201] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0202] The present application is described with reference to the flow chart and / or block diagram of the method, device (system), and computer program product according to the embodiment of the present application. It should be understood that each flow process and / or box in the flow chart and / or block diagram and the combination of the flow process and / or box in the flow chart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processing machine or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one flow chart flow or multiple flows and / or one box or multiple boxes of the block diagram.

[0203] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0204] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0205] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0206] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0207] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0208] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0209] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention. Industrial Applicability

[0210] From the above description, it can be seen that the present invention provides a product for detecting human nodule types with high accuracy based on low-depth sequencing data of extracellular free DNA whole genome. Specifically, the present invention proposes a method and device for establishing a lung nodule screening model, as well as a lung nodule screening method and device, which:

[0211] 1) Non-invasively detect the type of human nodules (benign / malignant), avoiding invasive examinations;

[0212] 2) Provide higher accuracy than existing CT scans, with accuracy approaching that of invasive tissue biopsies;

[0213] 3) This method was developed based on sequencing data of cfDNA and only requires blood drawing or collection of other body fluids, without the risk of radioactive exposure.

Claims

1. A method for establishing a lung nodule screening model, characterized in that: The assembly method comprises the following steps: In the entire range of the human reference genome, a predetermined number of regions having the largest differences in weighted fragment distribution difference (WFDD) values ​​between benign nodule population samples and malignant nodule population samples in the training set are selected as feature data; and The lung nodule screening model is constructed using the characteristic data.

2. The assembly method according to claim 1, wherein: The step of screening a predetermined number of regions with the largest weighted fragment distribution difference (WFDD) value differences between benign nodule population samples and malignant nodule population samples in the training set as feature data includes: Connect the base sequences of all autosomes in the human reference genome together, and divide the connected base sequences into a series of windows of fixed length, each window corresponding to a segment of base sequence; Calculate the window reference depth and weight of each window, where the window reference depth is the average of the depth values ​​of the samples in the training set in the window, the weight is the square of the variance of the depth values ​​of the samples in the training set in the window, and the window depth value is the number of base sequence fragments in the sequencing data of the sample that can be aligned to the base sequence corresponding to the window; Calculate the window sample depth of each window of the specified sample in the training set; Calculating the difference between the window sample depth and the window reference depth; Multiplying the difference by the weight to obtain a weighted difference of the window; An indefinite number of windows are combined to form different regions, and the weighted difference values ​​of all windows in the specified region are summed to obtain the weighted difference sum; Performing a numerical transformation on the weighted difference sum to obtain a weighted fragment distribution difference (WFDD) of the specified sample in the specified area; and Calculating the difference in the weighted segment distribution difference value in each region between the benign nodule population samples and the malignant nodule population samples in the training set, and selecting a predetermined number of regions with the largest differences as the feature data; Preferably, before calculating the difference between the window sample depth and the window reference depth, the method further comprises the step of performing a normalization process on the window reference depth and the window sample depth; More preferably, after the normalization process, the average value of the window reference depth of each window is 0, and the standard deviation is 1.

3. The assembly method according to claim 2, characterized in that: The predetermined number of regions with the largest screening differences as the feature data includes: - For a specific region, calculating the WFDD value of each sample in the benign nodule population sample and the malignant nodule population sample; - respectively calculating the average WFDD values ​​of the benign nodule population samples and the average WFDD values ​​of the malignant nodule population samples; - Calculate the discrimination value of the specific area according to the following formula, and - selecting the predetermined number of regions with the largest discrimination values ​​as the feature data: in: t is the discrimination value of the specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; the condition is: When the benign nodule population sample corresponds to the subscript 1, the malignant nodule population sample corresponds to the subscript 2; or When the samples of benign nodules correspond to the subscript 2, the samples of malignant nodules correspond to the subscript 1.

4. The assembly method according to claim 2, wherein: The predetermined number of regions with the largest screening differences as the feature data includes: a.) Generate the initial region: - Divide the genome into a series of windows; - numbering the series of windows by the order of the series of windows; - using a series of window numbers to number the area obtained by combining the series of windows; - Combine the n consecutive windows at window i and its downstream 2 j At n windows, another n consecutive windows are combined to form an initial region: x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2 j n+1,i+2 j n+2,...,i+2 j n+n-1} i=1,n,2n,...,N;j=1,2,...,8;i+(2 i+1 )n≤N Where n is the number of given continuous windows, and N is the total number of divided windows; b.) Combination and division of regions: Using a genetic algorithm, two parents combine to exchange information and generate offspring; all initial regions are placed in a region pool and offspring are randomly selected; in: -The probability that region i is selected as one of the parents is: Among them, N is the total number of divided windows, t i is the t value of the i-th window; -After region x is selected as the first parent, the probability that another region i is selected as the second parent is: Among them, N is the total number of divided windows, m i is the average value of the window numbers contained in region i; and -After selecting the parents, take the union of the windows included in the parents and randomly delete some of them as offspring, and the random selection method is sampling with replacement; - After obtaining the offspring, put the offspring into the regional pool for the next round of selection; in this operation, the parents are not deleted from the regional pool. Among them, the offspring generated by the combination of region P1 and region P2 is: child(P1,P2)=P1∪P2-S(p,P1∪P2) Where S(p,s) is a subset obtained by extracting elements of proportion p from the set s with replacement; - Repeat the process of producing offspring; - Calculate the discrimination value of the generated offspring according to the following formula, and select the predetermined number of regions with the largest discrimination value as the feature data: in: t is the discrimination value of the specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; the condition is: When the benign nodule population sample corresponds to the subscript 1, the malignant nodule population sample corresponds to the subscript 2; or When the samples of benign nodules correspond to the subscript 2, the samples of malignant nodules correspond to the subscript 1.

5. The assembly method according to claim 4, characterized in that: The number n of the continuous windows is 1-100, preferably 5-50, more preferably 5; Preferably, after selecting the parents, taking the union of the windows included in the parents and randomly deleting 1%-99%, more preferably 5%-50%, and even more preferably 20% of the windows as offspring, and performing sampling with replacement; Preferably, the process of repeatedly producing offspring is 1 to 1 million times, more preferably 100 to 100,000 times, and further preferably 300,000 times.

6. The assembly method according to claim 2, wherein: Dividing the connected base sequence into a series of windows according to the length of 100 bp-100 kbp, preferably 10 kbp-50 kbp, more preferably 30 kbp; Preferably, the windows combined to form different areas are continuous or discontinuous; Preferably, the predetermined number of regions is 1-500, more preferably 10-100, and even more preferably 50.

7. A method for screening pulmonary nodules, characterized in that: The following steps are involved: Calculate the sum of the weighted fragment distribution difference values ​​of the sample to be tested in a predetermined number of selected areas to obtain a total WFDD value; Inputting the total WFDD value of the sample to be tested into the pulmonary nodule screening model constructed by the construction method according to any one of claims 1 to 6; Outputting the screening result of the sample to be tested; The selected predetermined number of regions are the same as the predetermined number of regions having the largest difference in weighted fragment distribution difference (WFDD) between the benign nodule population sample and the malignant nodule population sample.

8. The screening method according to claim 7, wherein: Inputting the total WFDD value of the sample to be tested into the pulmonary nodule screening model, and the pulmonary nodule screening model determines the type of pulmonary nodule of the sample to be tested based on the comparison of the total WFDD value of the sample to be tested with a predetermined threshold; Preferably, the predetermined threshold is obtained by the following method: - calculating the total WFDD value of each sample in the training set in a predetermined number of regions where the WFDD value difference is the largest; - calculating the optimal segmentation point based on the total WFDD values ​​of the benign nodule population and the malignant nodule population in the training set, wherein the optimal segmentation point is the predetermined threshold; More preferably, the optimal split point is calculated using the roc function of the pROC package from the R language analysis platform.

9. A device for assembling a lung nodule screening model, characterized in that: The assembly device comprises: a feature data screening module configured to screen, within the entire human reference genome, a predetermined number of regions having the largest differences in weighted fragment distribution difference (WFDD) values ​​between benign nodule population samples and malignant nodule population samples in a training set as feature data; and The building module is configured to build the lung nodule screening model using the feature data.

10. The assembly device according to claim 9, characterized in that: The feature data screening module includes: The window division module is configured to concatenate the base sequences of all autosomes in the human reference genome and divide the concatenated base sequences into a series of windows of fixed length, each window corresponding to a segment of base sequence; A first calculation module is configured to calculate a window reference depth and a weight for each window, wherein the window reference depth is the average of the depth values ​​of the samples in the training set in the window, the weight is the square of the variance of the depth values ​​of the samples in the training set in the window, and the window depth value is the number of base sequence fragments that can be aligned to the base sequence corresponding to the window in the sequencing data of the sample; A second calculation module is configured to calculate the window sample depth of a specified sample in each window of the training set; a third calculation module, configured to calculate a difference between the window sample depth and the window reference depth; a fourth calculation module, configured to multiply the difference by the weight to obtain a weighted difference of the window; A fifth calculation module is configured to combine an indefinite number of windows to form different regions, and to sum the weighted differences of all windows in the specified region to obtain a weighted difference sum; a numerical transformation module configured to perform numerical transformation on the weighted difference sum to obtain a weighted fragment distribution difference (WFDD) value of the specified sample in the specified area; and a sixth calculation module, configured to calculate the difference in the weighted segment distribution difference value in each region between the benign nodule population samples and the malignant nodule population samples in the training set; and A feature data screening submodule is configured to screen a predetermined number of regions with the largest differences as the feature data; Preferably, the feature data screening module further includes a homogenization processing module; More preferably, after the normalization process, the average value of the window reference depth of each window is 0, and the standard deviation is 1.

11. The assembly device according to claim 10, characterized in that: The feature data screening submodule includes: A first calculation unit is configured to calculate, for a specific region, a WFDD value of each sample in the benign nodule population sample and the malignant nodule population sample; The second calculation unit is configured to respectively calculate an average value of the WFDD values ​​of the benign nodule population samples and an average value of the WFDD values ​​of the malignant nodule population samples; The third calculation unit is configured to calculate the discrimination value of the specific area according to the following formula, and The selection unit is configured to select the predetermined number of regions having the largest discrimination values ​​as the feature data: in: t is the discrimination value of the specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; the condition is: When the benign nodule population sample corresponds to the subscript 1, the malignant nodule population sample corresponds to the subscript 2; or When the samples of benign nodules correspond to the subscript 2, the samples of malignant nodules correspond to the subscript 1.

12. The assembly device according to claim 10, characterized in that: The feature data screening submodule includes: an initial region generating unit and a region combination splitting unit, wherein: The initial area generation unit includes: A window partitioning component is configured to partition the genome into a series of windows; a window coding element configured to number the series of windows using the order of the series of windows; An area numbering component is configured to use a series of window numbers as area numbers obtained by combining the series of windows; The window combination element is set to combine the n consecutive windows at window i and the 2 downstream windows j At n windows, another n consecutive windows are combined to form an initial region: x i ={i,i+1,i+2,...,i+n-1,i+2 j n,i+2 j n+1,i+2 j n+2,...,i+2 j n+n-1}, i=1,n,2n,...,N;j=1,2,...,8;i+(2 i+1 )n≤N Where n is the number of given continuous windows, and N is the total number of divided windows; The regional combination splitting unit includes: The first generation selection element is set to use a genetic algorithm, where two parents are combined to exchange information and generate offspring; wherein all initial regions are placed in a region pool and offspring are randomly selected; in: -The probability that region i is selected as one of the parents is: Among them, N is the total number of divided windows, t i is the t value of the i-th window; -After region x is selected as the first parent, the probability that another region i is selected as the second parent is: Among them, N is the total number of divided windows, m i is the average value of the window numbers contained in region i; The second offspring selection component is configured to select a parent, take the union of the windows included in the parent, and randomly delete some of the windows as offspring, and the random selection method is sampling with replacement; and The third child selection element is set to place the child into the region pool for the next round of selection after obtaining the child. In this operation, the parent is not deleted from the region pool. The child generated by combining regions P1 and P2 is: child(P1, P2)=P1∪P2-S(p, P1∪P2), where S(p,s) is a subset obtained by extracting proportion p of elements from the set s with replacement; and A progeny repetitive generation element is configured to repetitively generate a progeny process; - Calculate the discrimination value of the generated offspring according to the following formula, and select the predetermined number of regions with the largest discrimination value as the feature data: in: t is the discrimination value of the specific area; and are the average values ​​of the WFDD values ​​from the benign nodule population samples or the malignant nodule population samples, respectively; n1 and n2 are the number of values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; and S1 and S2 are the standard deviations of the values ​​from the benign nodule population sample or the malignant nodule population sample, respectively; the condition is: When the benign nodule population sample corresponds to the subscript 1, the malignant nodule population sample corresponds to the subscript 2; or When the samples of benign nodules correspond to the subscript 2, the samples of malignant nodules correspond to the subscript 1.

13. The assembly device according to claim 12, characterized in that: The number n of the continuous windows is 1-100, preferably 5-50, more preferably 5; Preferably, after selecting the parents, taking the union of the windows included in the parents and randomly deleting 1%-99%, more preferably 5%-50%, and even more preferably 20% of the windows as offspring, and performing sampling with replacement; Preferably, the process of repeatedly producing offspring is 1 to 1 million times, more preferably 100 to 100,000 times, and further preferably 300,000 times.

14. The assembly device according to claim 10, characterized in that: Dividing the connected base sequence into a series of windows according to the length of 100 bp-100 kbp, preferably 10 kbp-50 kbp, more preferably 30 kbp; Preferably, the windows combined to form different areas are continuous or discontinuous; Preferably, the predetermined number of regions is 1-500, more preferably 10-100, and even more preferably 50.

15. A pulmonary nodule screening device, characterized in that: The screening device comprises: The first calculation module is configured to calculate the sum of weighted fragment distribution difference values ​​of the sample to be tested in a predetermined number of selected areas to obtain a total WFDD value; an input module configured to input the total WFDD value of the sample to be tested into the pulmonary nodule screening model constructed by the construction device according to any one of claims 9 to 14; and An output module, configured to output the screening result of the sample to be tested; The selected predetermined number of regions are the same as the predetermined number of regions having the largest difference in weighted fragment distribution difference (WFDD) between the benign nodule population sample and the malignant nodule population sample.

16. The screening device according to claim 15, characterized in that The input module includes: an input unit configured to input the total WFDD value of the sample to be tested into the pulmonary nodule screening model; and A determination unit is configured to determine the type of lung nodules of the sample to be tested by comparing the total WFDD value of the sample to be tested with a predetermined threshold value according to the lung nodule screening model; Preferably, the screening device further includes a predetermined threshold value acquisition module, and the predetermined threshold value acquisition module includes: A first calculation unit is configured to calculate a total WFDD value of each sample in the training set in a predetermined number of regions where the WFDD value differences are largest; and A second calculation unit is configured to calculate an optimal segmentation point according to the total WFDD values ​​of the benign nodule population and the malignant nodule population in the training set, where the optimal segmentation point is the predetermined threshold; More preferably, the optimal split point is calculated using the roc function of the pROC package from the R language analysis platform.

17. A computer-readable storage medium, characterized in that The storage medium includes a stored program. When the program is run, the program executes the method for constructing a lung nodule screening model according to any one of claims 1 to 6 or the lung nodule screening method according to claim 7 or 8.

18. A processor, characterized in that: The processor is used to run a program, wherein when the program is run, the method for building a lung nodule screening model according to any one of claims 1 to 6 or the lung nodule screening method according to claim 7 or 8 is executed.

Images