Method and apparatus for determining copy number variation profiles by read-depth correction in whole-genome analysis

Abstract

The present disclosure provides a copy number variation prof ile determination method executed by at least one processor. Solution The copy number variation profile determination method comprises the following steps: obtaining a whol e-genome analysis result associated with a target sample collected from a subject; based on the obtained whole-genome analysis result, calculating, for each of a plurality of predetermined bins on a genome, a read depth associated with the target sample; correcting the read depth associated with the target sample; and determining, using the corrected read depth, a copy number variation (Copy Number Variation) profile associated with the target sam ple.

Description

Technical Field

[0001] The present disclosure relates to a method and apparatus for determining a copy number variation profile through read depth correction Specifically, the present disclosure relates to a method and apparatus for determining a copy number variation profile by calculating and correcting the read depth associated with a target sample collected from a subject based on the whole genome analysis result associated with the target sample to determine the copy number variation profile.

Background Art

[0002] Genetic information analysis technology is widely used in the medical field, for example it is used for determining what characteristics or temperaments an organism has by grasping the genetic information of the organism Recently, medical practices for understanding the causes of and treating various diseases such as tumors have shifted from a traditional prescription-centered approach to precision medicine , that is, a customized treatment form that takes into account individual patient's genetic information, health records, and the like, and has been evolving into such customized treatment form. In the field of precision medicine, the main practice is to obtain a huge amount of personal genetic information and conduct relevant clinical analysis, and such main element corresponds to the core element that accelerates the development speed of precision medical technology.

[0003] In particular, when performing whole genome analysis on a tissue collected from a patient or the like, the so-called "Fresh Frozen (FF)" treatment method, in which the tissue is frozen immediately after being collected from the human body, is widely used FF-treated tissue is frozen immediately after collection, so that DNA damage to cells in the tissue is less, which is known to be the optimal tissue treatment method for whole genome analysis. However, when FF treatment is performed on a collected tissue, or the FF-treated tissue To store the textiles, nitrogen tanks and other equipment are not typically available in a clinical setting. Alternatively, there are issues with the requirement for facilities or equipment that are difficult to provide.

[0004] On the other hand, tumor tissue is removed or biopsies are performed in order to analyze the genetic information of patients, etc. In such cases, medical institutions use FFPE (Formula Prescription) to process tissue samples taken from patients (such as tumor tissue). (Alin-Fixed, Paraffin-Embedded) treatment for long-term storage, It is common practice to utilize FFPE-treated tissue for follow-up examinations or academic research purposes. When processing tissue collected through the FFPE method, the processing and storage of the collected tissue Not only does it require little expense or effort, but it also preserves most of the genetic information within the tissue. Because the tissue can be stored for a long period of time, the collected tissue can be used later (for example, re-examination of the tissue). It has the advantage of being easy to examine or re-analyze.

[0005] However, during the process of long-term storage of tissue samples taken from the human body using FFPE treatment, the interaction of DNA Cross-linking is a chemical entanglement between different parts. , fragmentation, in which DNA is cut into smaller pieces. n) Various DNA mutations in tissues, such as DNA base mutations due to other non-biological causes. This type of damage may occur.

[0006] Due to the DNA damage described above, whole-genome analysis was performed on FFPE-treated tissues. In such cases, copy number variations (CNV) in the analysis data (raw data) may be considered. (Variation) Noise occurs in the profile, resulting in an inaccurate and distorted analysis. Results can be derived. Such noise is present in whole-genome analysis data of FF-treated tissues. It is generally not found. Therefore, distortion occurs in FFPE-treated tissue. To derive accurate analysis results, effectively process the noise on the copy number variation profile. Or it needs to be removed. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Korean Published Patent Publication No. 10-2018-0098438 [Overview of the project] [Problems that the invention aims to solve]

[0008] This disclosure provides a method for determining copy number mutation profiles to solve the problems described above. We provide computer programs and devices (systems) stored on a recording medium. [Means for solving the problem]

[0009] This disclosure is a computer-readable recording of a method, system (apparatus) or instructions. It can be realized in various ways, including recording media.

[0010] A read / decoded process executed by at least one processor according to one embodiment of the present disclosure. In a method for determining copy number variation profiles using read depth correction, Steps to obtain whole-genome analysis results associated with target samples taken from subjects. Based on the obtained whole-genome analysis results, predetermined on the genome. For each of the multiple bins, the target sample and the associated read depth ( calculating read depth), and a target-sample-related read dept h correction step, and determining a copy number variation profile associated with the target sample using the corrected read depth Copy Number Variation profile associated with the target sample comprising the step of determining the profile.

[0011] There is provided a computer-readable non-transitory recording medium having recorded thereon instructions for causing a computer to execute the method for determining a copy number variation profile according to an embodiment of the present disclosure. There is provided a computer-readable non-transitory recording medium having recorded thereon instructions for causing a computer to execute the method for determining a copy number variation profile according to an embodiment of the present disclosure.

[0012] In a computing device according to an embodiment of the present disclosure, the computing device comprises a communication module, a memory , and at least one processor coupled to the memory and configured to execute at least one computer-readable program contained in the memory, wherein said at least one processor comprises at least one processor configured to execute at least one computer-readable program contained in the memory, wherein said at least one program obtains whole genome analysis results associated with a target sample collected from a subject, and based on the obtained whole genome analysis results, for each of a plurality of predetermined bins on a genome , calculates a read depth associated with the target sample for each of the plurality of predetermined bins on the genome ), calculates a read depth associated with the target sample for each of a plurality of predetermined bins on the genome , corrects the read depth associated with the target sample, and uses the corrected read depth to obtain a read depth associated with the target sample , corrects the read depth associated with the target sample, and uses the corrected read depth to obtain a read depth associated with the target sample the copy number variation pro file) comprises instructions for determining. Effects of the Invention

[0013] According to various embodiments of the present disclosure, by grouping the genome into a plurality of bins, , noise (random noise or non-random noise) when calculating read depth The instability of the target sample's read depth due to noise can be significantly reduced. .

[0014] According to various embodiments of this disclosure, the detailed wavelet coefficients are hierarchical thresholds By being subjected to hierarchical thresholding, The higher the level of detail, including macroscopic morphological information of Null, the more detailed the wavelet coefficients are, the more they are replaced with 0. The degree of reduction and attenuation is low, and the macroscopic copy number variation profile is preserved to the maximum extent. It is possible.

[0015] According to various embodiments of this disclosure, whole-genome analysis is performed on FFPE-treated tissues. By correcting the noise or errors that may occur in the process, the entire data of the FF-processed tissue is corrected. Unbiased analysis results, similar to those obtained from Nomu's analysis data, can be derived.

[0016] According to various embodiments of this disclosure, healthcare institutions and biobanks High-precision whole-genome analysis will be performed on the vast amount of FFPE tissue that has been secured and accumulated by these organizations. Not only can this be done, but the tissue sample processing procedures at medical institutions can be performed using only the facilities normally available in clinical settings. Whole-genome analysis of tissue samples from patients and other sources can proceed without any changes to the plan.

[0017] The effects of this disclosure are not limited to those mentioned above, and any other effects not mentioned are subject to change. Based on the description of the scope of the request, this disclosure is intended for persons with ordinary skill in the art to which it pertains. This can be clearly understood by the "surgeon." [Brief explanation of the drawing]

[0018] Embodiments of the present disclosure will be described with reference to the accompanying drawings described below, and similar references thereto The numbers indicate similar elements, but are not limited to them. [Figure 1] This figure shows the detailed steps of a method for determining a copy number variation profile according to one embodiment of the present disclosure. [Figure 2] This is a schematic diagram showing a configuration in which an information processing system is connected to multiple user terminals in a communicative manner in order to provide a copy number variation profile determination service according to one embodiment of the present disclosure. [Figure 3] This is a block diagram showing the internal configuration of a user terminal and an information processing system according to one embodiment of the present disclosure. [Figure 4] This figure shows an example of a corrected dataset according to one embodiment of the present disclosure. [Figure 5] This figure shows the first-order correction process for the target sample and associated read depth for each of the multiple bins according to one embodiment of the present disclosure. [Figure 6] This figure shows the process by which the value of the correction variable that minimizes the result value of the standard deviation function according to one embodiment of the present disclosure is determined. [Figure 7] This figure shows the process by which a secondary correction is performed to remove noise from a primary corrected read depth according to one embodiment of the present disclosure. [Figure 8] This figure shows an example of a graph and coefficient set used during secondary correction according to one embodiment of the present disclosure. [Figure 9] This figure shows examples of a first heatmap associated with an FF-treated sample and a second heatmap associated with an FFPE-treated sample according to one embodiment of the present disclosure. [Figure 10] This figure shows an example of a distribution graph of the read depth ratio for each sample before and after correction of the read depth associated with the FFPE-treated sample according to one embodiment of the present disclosure. [Figure 11] This figure shows an example of the distribution of read depth by chromosome according to one embodiment of the present disclosure. [Figure 12]This figure shows an SNR graph for evaluating the degree of qualitative improvement in the copy number variation profile by read depth correction according to one embodiment of the present disclosure. [Figure 13] This flowchart shows a method for determining a copy number variation profile according to one embodiment of the present disclosure. [Modes for carrying out the invention]

[0019] <Summary of the Invention> In one embodiment of this disclosure, the target sample is a normal cell taken from a subject. A step that includes samples and abnormal cell samples to determine the copy number variation profile. This involves read depth associated with normal cell samples and read depth associated with abnormal cell samples. S, and the allele frequency associated with the target sample. This includes a step of determining the copy number variation profile based on ncy).

[0020] In one embodiment of this disclosure, the target sample is FFPE (Formalin- This is a sample that has undergone fixed (paraffin-embedded) processing.

[0021] In one embodiment of this disclosure, the read depth associated with the target sample is corrected. The step involves each of the multiple samples taken from the subject and multiple other subjects. Using a dataset containing read depth data for each of multiple connected bins This includes a step of performing a primary correction of the target sample and associated read depth.

[0022] In one embodiment of this disclosure, the dataset is collected from each of a plurality of subjects. Data associated with normal cell samples, or abnormal cells taken from each of multiple subjects. Includes at least one data point associated with a congenital cell sample.

[0023] In one embodiment of the present disclosure, normal cell samples collected from each of a plurality of subjects Each of the abnormal cell samples taken from the subject and from each of the multiple subjects was FFP Includes E-treated and FF (Fresh Frozen) treated samples. nothing.

[0024] In one embodiment of this disclosure, the dataset is the read depth of the target sample To be used for correction, the correction direction and relationship of the read depth for each of the multiple bins. It also includes related information.

[0025] In one embodiment of this disclosure, the correction direction and related information are provided for each of the multiple bins. Then, the distribution of the entire FFPE-processed sample and associated read depth within the dataset. , and for each of the multiple bins, the entire FF-processed sample in the dataset. It is determined based on the distribution of associated read depths.

[0026] In one embodiment of this disclosure, the dataset is the read depth of the target sample The correction size for the read depth for each of the multiple bins, used for correction. This includes further related information.

[0027] In one embodiment of this disclosure, the correction size and related information are provided for each of the multiple bins. Then, the average of all FFPE-processed samples in the dataset and the associated read depth. , and for each of the multiple bins, the entire FF-processed sample in the dataset and It is determined based on the average of the associated read depths.

[0028] In one embodiment of this disclosure, the step of performing a primary correction involves the correction direction and related information. Based on the correction size and related information, the target sample and associated read depth Includes a correction step.

[0029] In one embodiment of this disclosure, the read depth associated with the target sample is corrected. The next step is to perform a second-order correction to remove noise from the first-order corrected read depth. It also includes.

[0030] In one embodiment of this disclosure, a secondary correction is performed to remove noise from the primary corrected read depth. The correction step involves a predetermined wavelet and a first-order correction. Using the read depth, multiple wavelet coefficients are used. Steps to determine the icient, multiple wavelet coefficients thresholding (Thresholding) steps, and multiple thresholded steps The process includes determining a quadratic corrected read depth based on wavelet coefficients.

[0031] In one embodiment of this disclosure, a predetermined wavelet is a Haar wavelet It corresponds to category T.

[0032] In one embodiment of the present disclosure, a plurality of wavelet coefficients are thresholded. The step is determined by the absolute value of a particular wavelet coefficient among several wavelet coefficients. If the value is below a predetermined threshold, the step is to replace a specific wavelet coefficient with 0. And if the absolute value of a particular wavelet coefficient is greater than or equal to the threshold, the interval between the absolute value and the threshold The smaller the difference, the more the specific wavelet coefficient is attenuated (attenuated). This includes the step of causing it to happen.

[0033] In one embodiment of the present disclosure, a specific wavelet coefficient is at a high level of wavelet The lower the coefficient, the lower the threshold becomes.

[0034] <Details of the invention> The specific details for implementing this disclosure will be described below with reference to the attached drawings. Furthermore, the following explanation may unnecessarily obscure the gist of this disclosure, so it may not be widely known. I will omit a detailed explanation of the functions and configurations involved.

[0035] In the attached drawings, identical or corresponding components are assigned the same reference numeral. Furthermore, in the following descriptions of embodiments, identical or corresponding components will not be described redundantly. This may be omitted. However, even if the description of the constituent elements is omitted, such a composition It is not intended that any element is excluded from any embodiment.

[0036] The advantages and features of the disclosed embodiments, as well as methods for achieving them, are attached. This will become clearer with reference to the embodiments described later, along with the drawings. However, this disclosure is as follows: The embodiments are not limited to those disclosed herein and may be embodied in a variety of different forms. This embodiment merely ensures that the present disclosure is complete and that the present disclosure is within the scope of the invention for those of an ordinary skill. It is merely provided to fully inform you of the situation.

[0037] This document provides a brief explanation of the terms used herein and provides specific details about the disclosed embodiments. This will be explained below. The terms used herein are defined in consideration of their function in this disclosure. Therefore, we have selected common terms that are as widely used as possible, but this is because we are engaged in the relevant fields. It can change depending on the intentions of the engineers, precedents, the emergence of new technologies, etc. Also, in certain cases... Some terms may be arbitrarily selected by the applicant; in this case, their meaning should be explained in detail in the description of the relevant invention. The taste will be described. Therefore, the terms used in this disclosure are not simply the names of terms. Rather, the term must be defined based on its meaning and the overall content of this disclosure. I don't know.

[0038] In this specification, singular expressions refer to plural unless the context clearly indicates that they are singular. This includes expressions. Also, unless the context clearly indicates that multiple expressions are plural, they are treated as singular. This includes expressions. If the entire specification includes a certain component in a particular part, this does not specifically contradict Unless otherwise specified, it may include other components rather than excluding them. It means to do something.

[0039] Furthermore, the terms "module" or "part" used in the specification refer to software. "Module" or "part" refers to a hardware component, and "module" or "part" refers to a component that performs a certain role. However, the term "module" or "part" is not limited to software or hardware. It does not mean that a "module" or "part" is an addressable storage medium. It can also be configured to allow one or more processors to play back. It can also be done. Therefore, as an example, a "module" or "part" is a software Software components, object-oriented software components, class components and task Components such as components, processes, functions, attributes, procedures, and subroutines, Program code segments, drivers, firmware, microcode, circuits, At least one of the following: data, database, data structure, table, array, or variable It may include. The components and the functions provided within a “module” or “part” are fewer in number. Components and "modules" or "parts" are combined, or additional components are combined with "modules". It may be further separated into "section" or "part".

[0040] According to one embodiment of the present disclosure, the “module” or “part” is a processor and memory This can be embodied in: "Processor" refers to a general-purpose processor, central processing unit (CPU), microphone Microprocessors, digital signal processors (DSPs), controllers, microcontrollers It should be interpreted broadly to include things like state machines. In some environments, A processor is an application-specific semiconductor (ASIC), a programmable logic device ( It can also refer to PLDs, Field Programmable Gate Arrays (FPGAs), etc. A "processor" is, for example, a combination of a DSP and a microprocessor, or multiple microprocessors. A combination of processors, a combination of one or more microprocessors coupled with a DSP core. A combination of processing devices, such as a combination of any other such configuration. It can also refer to [something else]. Furthermore, "memory" refers to any electronic component capable of storing electronic information. It should be interpreted broadly to include "memory". "Memory" refers to random access memory (R). AM), read-only memory (ROM), non-volatile random access memory (NVRAM) ), programmable read-only memory (PROM), erasable programmable read Dedicated memory (EPROM), electrically erasable PROM (EEPROM), flash memory Processor-readable devices such as magnetic or optical data storage devices and registers. It can also refer to various types of memory media. For example, when a processor reads information from memory, And / or if it is possible to write information to memory, the memory communicates electronically with the processor. This state is called a state of electronic communication. The memory integrated into the processor is in a state of electronic communication with the processor. ru.

[0041] In this disclosure, “Subject” refers to the subject who will undergo copy number variation profiling in accordance with this disclosure. This could refer to an elephant patient or subject of study.

[0042] In this disclosure, “target sample” means a copy. Read depth is calculated to determine the numerical variation profile of the target subjects, and the sample size is adjusted. It can refer to a simple.

[0043] In this disclosure, "bin" refers to a small portion of the genome for gene / genome analysis. The genome is a unit used to group or classify a portion of something. e) This can refer to segments defined above. The position and number of base pairs, the number of bottles, etc., can be determined arbitrarily. For example, approximately 3 billion pairs For the genome of a subject containing bases, 30 groups of 1 million bases each can be formed. More than 00 bottles may be determined.

[0044] In this disclosure, “abnormal cells” are defined as cells that differ from normal cells in size, shape, structure, function, etc. This can refer to abnormal cells that possess certain characteristics. Abnormal cells can be caused by various factors such as genetic mutations, infection, and toxin exposure. It can be caused by various factors, including cancer cells, tumor cells, necrotic cells, senescent cells, and aneuploid cells (Aneuploid). (loid Cell), Hyperplastic Cell, Hypertrophic Cell It may include various types of abnormal cells, such as hypertrophic cells.

[0045] In this disclosure, "read depth ratio" is defined as normal. Read depth obtained as a result of sequencing cell samples and the sequence of abnormal cell samples This can refer to the ratio between the read depth obtained as a result of the measurement. Since the read depth can be calculated from the read depth ratio, "read depth" Unless the context clearly identifies it as read depth, "S" can refer to the read depth ratio. .

[0046] In this disclosure, the term "wavelet" or "wavelet function" is used. A "wavelet function" is a one-dimensional function that satisfies certain properties. It oscillates in a localized area with negative and positive values, with 0 as the base point. This can refer to a one-dimensional function whose integral value for all real numbers is 0. For example, "wavelet The wavelet function has a domain of (-∞, ∞) and its oscillations are expressed as follows: Outside the localized region, the region extending in both the negative and positive infinity directions can have a value of 0. Let or wavelet functions are the shape of the oscillation. Classified by Haar, Daubechies, Coiflet, Best-lo Various species such as calized Daubechies and Leastasymmetric Similar species may exist.

[0047] Figure 1 shows the detailed steps of a copy number variation profile determination method according to one embodiment of the present disclosure. This figure shows that in one embodiment, the copy number variation profile determination method is a learning step This may include a step (110) and an application step (160).

[0048] In the learning step (110), multiple FFPE sample data (120_1~120_ n) FFPE sample dataset (120) and multiple FF sample data ( FF sample dataset containing 130_1~130_n, where n is a natural number) Based on 130), corrected data set for correcting the target sample data (150) The set (140) can be determined / learned. FFPE sample dataset (120) and The FF sample dataset (130) is different from the subjects from whom the target sample was collected. This may include data calculated from multiple samples taken from multiple subjects. "Sample data" refers to the sample and its associated copy number variation (CopyNumberVa riation, CNV) profile or read depth The sample data can be any data related to the sample, and the FFPE sample data is FFPE(F Samples treated with ormalin-fixed or paraffin-embedded. The data related to the FF sample data is processed using FF (Fresh-Frozen) processing. This can refer to the sample and related data.

[0049] At least one of several FFPE sample data (120_1~120_n) And at least one of the multiple FF sample data (130_1~130_n) is identical This could be data related to samples taken from subjects. For example, the first FFPE Sample data (120_1) and the first FF sample data (130_1) are specific Samples were collected from subjects and processed using different methods (FFPE, FF) and related These are related data, and include multiple FFPE sample data (120_1~120_n) Multiple FF sample data (130_1~130_n) were collected from n subjects. These may be samples and related data processed in different ways.

[0050] FFPE sample dataset (120) and FF sample dataset (130) Each of these involved data associated with normal cell samples taken from each of multiple subjects. Abnormal cell samples taken from each of the subjects and associated data This may include the first FFPE sample data (120_1) and the first F F sample data (130_1) is a normal cell sample taken from a specific subject. Each of the multiple bins on the analyzed whole genome sequence This is the read depth data for the second FFPE sample data (120_2). The second FF sample data (130_2) consists of abnormal cells collected from the same subject. Read depth data for each of the multiple bins on the whole genome sequence analyzed from the sample could be.

[0051] The corrected dataset (140) is the FFPE sample dataset (120) and FF It may include all the data contained in the sample dataset (130). For example, corrected data. Set (140) consists of normal cells collected from multiple subjects, each different from the subject. This may include sample and abnormal cell samples and associated read depth data. , FFPE sample dataset (120) or FF sample dataset (130) ) includes copy number variation profiles rather than read depth data for each sample. If so, a separate process is used to determine read depth data from the copy number variation profile. S may be executed.

[0052] Target sample data (150) is corrected in accordance with this disclosure for the target of the subjects. Samples and associated data (e.g., copy number variation profiles, read depth data) It could be (e.g., a sample of 150). For example, target sample data (150) is collected from subjects. Based on the target sample and associated whole-genome analysis results, each of the multiple bins This could be the read depth associated with the target sample, calculated in relation to this. The whole genome analysis results were obtained using standard paired-end sequencing techniques. The read data is processed using various sequencing data processing tools to access the reference genome (re reference genome) (hg19, hg38, GRCh37, GRCh38 etc. Includes sequence data obtained by mapping and aligning to ( ). The target sample is a normal cell sample and / or a different cell sample taken from the subject. May contain normal cell samples.

[0053] In one embodiment, the target sample is FFPE (Formalin-Fixed (Paraffin-Embedded) sample may be a sample treated with a target sun When pulls are processed with FFPE, the copy number variation profile (ma Alternatively, noise occurring in the lead depth is removed by this disclosure, and the target sample It is possible to obtain substantially the same or similar analytical accuracy as that obtained after FF processing. can.

[0054] The correction dataset (140) is used to correct the read depth of the target sample. For use, the correction direction of the read depth for each of the multiple bins and related information It may also include information related to the correction size, namely the correction direction of the lead depth and The correction size for each bin / or read depth may differ for each of the multiple bins.

[0055] The correction dataset (140) can be implemented and utilized in various forms such as matrices, arrays, and vectors. It can be used. For example, rows (r ow) indicates each of the multiple bins, and column indicates the processing type (for example) For example, FFPE, FF) and / or the type of sample used (e.g., normal cell sample) Show samples from each of multiple subjects using pull (abnormal cell samples), or This indicates the correction direction and / or correction size of the read depth for each of the multiple bins. Obtain. A specific example of the corrected dataset (140) will be described in detail later in Figure 4.

[0056] In the application step (160), the correction dataset (140) and the first-order correction function (17 2) can be used to correct the target sample data (150). For example, corrected data Information related to the correction direction and correction size included in the taset (140) Based on this, the target sample and associated read depth are determined by the first-order correction function (172). The value can be corrected. This will be explained in detail later in Figures 5 and 6.

[0057] Subsequently, the noise in the first-order corrected data is removed by the second-order correction function (174). This generates corrected target sample data (FFPEX')(180). This is possible. This will be explained in detail later in Figures 7 and 8.

[0058] Subsequently, corrected target sample data (180) (e.g., corrected reads) Using depth, the target sample and its associated copy number variant profile (CNV) X')(190) can be determined. For example, the copy number variation profile (190) is - Read depth associated with normal cell samples from the get sample, and differences in the target sample. Read depth associated with congenital cell samples, and alleles associated with target samples. This can be determined based on allele frequency. Gene frequency is the purity of the target sample (cells within the tissue being sequenced). (The proportion of abnormal cells among them) and mean ploidy (the average number of DNA molecules in all regions). For example, normal human cells have DNA or genomes in pairs, so mean pl It is affected by oidy (corresponding to 2) and hardly affected by lead depth. Therefore, the noise that occurs when measuring allele frequencies in FFPE-treated samples is in the reads. In some cases, the results may not be as relevant as they would be in terms of distorting the analysis compared to depth analysis.

[0059] Corrected read depth or copy number determined using corrected read depth The mutation profile was added to the corrected dataset (140) and compared to other subjects. This can be utilized again when correcting for depth and determining copy number variation profiles.

[0060] Figure 2 shows a copy number variation profiling service provided according to one embodiment of the present disclosure. Therefore, the information processing system (230) has multiple user terminals (210_1, 210_2, 2 This is a schematic diagram showing a configuration that is connected to 10_3) in a manner that allows for communication. Information processing system (230 ) is a system (one or more) that can provide a copy number variation profiling service. or may include computing devices. In one embodiment, an information processing system (2 30) is a computer executable related to copy number mutation profiling services. Store and provide programs (e.g., downloadable applications) and data. and one or more server devices and / or databases, or cloud services that can run on them. One or more distributed computing devices based on computing services and / or This may include distributed databases.

[0061] Copy number variation profiling service provided by the information processing system (230) The system is installed on each of the multiple user terminals (210_1, 210_2, 210_3). It can be provided to users through a downloaded application or similar means.

[0062] Multiple user terminals (210_1, 210_2, 210_3) are connected to the network (220 It can communicate with the information processing system (230) via the network (220). The user terminals (210_1, 210_2, 210_3) and the information processing system (230) The network (220) can be configured to allow communication between them. For example, Ethernet, wired home network (Power (r Line Communication), telephone line communication equipment and RS-seria Wired networks such as communications, mobile communication networks, WLAN (Wireless LAN), W Wireless networks such as iFi, Bluetooth and ZigBee or It can be composed of such combinations. The communication method is not restricted and includes the network (220). Access to communication networks (for example, mobile communication networks, wired internet, wireless internet, broadcasting) In addition to communication methods that utilize networks, satellite networks, etc., user terminals (210_1, 210_2 This may also include short-range wireless communication between 210_3).

[0063] For example, multiple user terminals (210_1, 210_2, 210_3) are connected via a network. A request is sent to the information processing system (230) via (220), After receiving this, the information processing system (230) sends a response to the request. ) can be sent to multiple user terminals (210_1, 210_2, 210_3). For example, User terminal (210_1) requests whole genome analysis results related to the target sample and When a request for copy number mutation profile determination is sent to the information processing system (230) (request) The information processing system (230) processes the target sample and the associated copy number variation profile. The system may send messages such as a message to the user terminal (210_1) (response).

[0064] In Figure 2, a mobile phone terminal (210_1), a tablet terminal (210_2), and P Although terminal C (210_3) is shown as an example of a user terminal, it is not limited to this, The terminals (210_1, 210_2, 210_3) are capable of wired and / or wireless communication. Therefore, applications such as copy number mutation profiling applications are installed and run. It could be any computing device that can be used. For example, a user terminal could be a medical device, Smartphones, mobile phones, navigation systems, computers, laptops, digital Broadcast terminals, PDAs (Personal Digital Assistants), P MP (Portable Multimedia Player), tablet PC, game Game consoles, wearable devices e-devices, IoT (Internet of Things) devices, VR (Virtual reality) devices, AR (augmented reality) This may include devices such as ty. Also, Figure 2 shows three user terminals (210_1, 210 _2, 210_3) communicate with the information processing system (230) via the network (220). Although indicated as a trust, it is not limited to this, and different numbers of user terminals are networked It may also be configured to communicate with an information processing system (230) via a link (220). can.

[0065] Figure 2 shows that the user terminal transmits copy number variations via network communication with the information processing system. It is stated that a profile will be provided, but is not limited to this, and users will be able to use the information processing system Request for whole genome analysis results related to the target sample via an input device connected to the system. A request is made to determine the copy number variation profile, and in response, the system connects to the information processing system. The whole genome analysis results and copies associated with the target sample are output via a connected output device. Numerical variation profiles may be provided.

[0066] Figure 3 shows a user terminal (210) and an information processing system (2) according to one embodiment of the present disclosure. This is a block diagram showing the internal configuration of (30). The user terminal (210) is a copy number variant programmer. File selection applications and other similar programs can be executed, and any wired / wireless connection is possible. This can refer to a computing device, for example, the mobile phone terminal (210_1) in Figure 2, a tablet This may include user terminals (210_2), PC terminals (210_3), etc. As shown in the diagram, The terminal (210) includes memory (312), processor (314), and communication module (316 ) and input / output interfaces (318) may also be included. Similarly, information processing systems (23 0) includes memory (332), processor (334), communication module (336) and input It may include an output interface (338). As shown in Figure 3, the user terminal (210) The information processing system (230) utilizes each of its communication modules (316, 336). It is configured to enable the communication of information and / or data over a network (220). This can be achieved. In addition, the input / output device (320) can be accessed via the input / output interface (318). Entering information and / or data into the user terminal (210), or the user terminal (210) It can be configured to output information and / or data generated from it.

[0067] The memory (312, 332) may include any computer-readable recording medium. According to one embodiment, the memory (312, 332) is a non-temporary computer read It can be a read-only recording medium, for example, ROM (read-only memory). Disk drives, SSDs (solid state drives), flash memory Non-erasable mass storage devices such as flash memory (permanent memory) This may include (nt mass storage device). Another example is ROM, Non-erasable mass storage devices such as SSDs, flash memory, and disk drives are A separate permanent storage device distinct from memory is used as a user terminal (210) or information processing system. It may be included in stem(230). Also, memory(312,332) contains operating A system and at least one program code (for example, copy number variation profiling) Code for applications, etc., may be stored there.

[0068] These software components are separate from the memory (312, 332) of the computer. It can be loaded from a reading storage medium. Such a separate computer can read it. The recording media that can be used are these user terminals (210) and information processing systems (230). It may include a recording medium that can be directly connected to it, for example, a floppy drive, disk, tape, Computer-readable recording media such as DVD / CD-ROM drives and memory cards. This may include media. As another example, software components are computer-readable. The memory (312, 332) is transmitted via a communication module (316, 336) which is not a recording medium. ) can also be loaded into. For example, at least one program can be developed by or A file distribution system that distributes application installation files over a network Computer programs installed via files provided through (220) Memory based on RAM (e.g., copy number variation profiling application) It can be loaded into (312, 332).

[0069] The processors (314, 334) perform basic arithmetic, logical, and input / output operations. This allows it to be configured to process instructions in a computer program. The instructions are The processor is controlled by Mori (312, 332) or communication module (316, 336) 314, 334) may be provided to the memory. For example, the processor (314, 334) may provide memory ( Life received according to program code stored in a recording device such as 312, 332) It may be configured to execute orders.

[0070] The communication modules (316, 336) connect to the user terminal via the network (220). 210) and the information processing system (230) provide a configuration or function for them to communicate with each other. If a user terminal (210) and / or information processing system (230) is connected to another user terminal Or a configuration for communicating with other systems (for example, a separate cloud system). or it can provide functionality. For example, the processor (314) of the user terminal (210) The necessary data generated according to the program code stored in a recording device such as Mori (312) Requests or data are transmitted to the network (220) by the control of the communication module (316). It can be transmitted to the information processing system (230) via this. Conversely, the information processing system (230) Control signals and instructions provided by the processor (334) are transmitted to the communication module ( 336) and the communication module (31) of the user terminal (210) via the network (220) 6) can be received by the user terminal (210).

[0071] The input / output interface (318) is for interfacing with the input / output device (320) This could be a means. For example, the input device could be an audio sensor and / or an image sensor. Devices such as cameras including sensors, keyboards, microphones, mice, and output devices The components include a display, speakers, and a haptic feedback device. This may include devices such as feedback devices. Another example is input / output devices. The interface (318) is used to perform input and output, such as a touchscreen. This could be a means for interfacing with a device in which the configuration or functions are integrated into a single unit. In Figure 3, the input / output device (320) is shown not to be included in the user terminal (210). However, it is not limited to this, and may consist of a user terminal (210) and a single device. The input / output interface (338) of the information processing system (230) is the information processing system ( Inputs or outputs that are connected to (230) or that the information processing system (230) may include It may be a means for interface with a device (not shown) for the purpose of input and output. The power interface (318, 338) is configured separately from the processor (314, 334). Although shown as an element, it is not limited to this, and includes input / output interfaces (318, 3 38) may be configured to be included in the processor (314, 334).

[0072] The user terminal (210) and the information processing system (230) are more numerous than the components shown in Figure 3. It may include many components. However, most conventional components do not need to be clearly illustrated. There is no such thing. In one embodiment, the user terminal (210) is the above-mentioned input / output device (320) It can be embodied to include at least a part of it. In addition, the user terminal (210) is a trans Transceiver, GPS (Global Positioning System) Other components such as the system module, camera, various sensors, and database. This may further include:

[0073] According to one embodiment, the processor (314) of the user terminal (210) has copy number variation An application or web browser application that provides a profile determination service. The process can be configured to work. In this case, the process associated with the application and Gram code can be loaded into the memory (312) of the user terminal (210). While the operation is running, the processor (314) of the user terminal (210) processes the input / output device ( Information and / or data provided from 320) to the input / output interface (318) It can be received via or from the information processing system (230) via the communication module (316). It can receive information and / or data, and the received information and / or data The data can be processed and stored in memory (312). Also, this information and / or data The data can be provided to the information processing system (230) via the communication module (316).

[0074] While the application is running, the processor (314) uses the input / output interface ( 318) A touchscreen, keyboard, audio sensor and / or connected to it Inputs are received or selected via input devices such as cameras including image sensors and microphones. It can receive audio data, text, images, videos, etc., and the received audio Voice data, text, images and / or video, etc. are saved to memory (312), Information processing system (23) via signal module (316) and network (220) 0) It may be provided to.

[0075] The processor (314) of the user terminal (210) inputs and / or inputs information and / or data. It can be transmitted to the input / output device (320) via the interface (318) and output. For example The processor (314) of the user terminal (210) is a display output device (e.g., Output from devices capable of outputting sound (e.g., speakers), such as touchscreens and displays. The device (320) may output information and / or data that has been processed via the device.

[0076] The processor (334) of the information processing system (230) has multiple user terminals (210) and / or manage and process information and / or data received from multiple external systems. , and / or may be configured to be stored. Processed by processor (334) The information and / or data are transmitted to the communication module (336) and network (220 It may be provided to the user terminal (210) via ).

[0077] In one embodiment, the processor (334) processes target sunlight collected from a subject. Based on the pull and related whole-genome analysis results, a predetermined number of bins on the genome For each, the read depth associated with the target sample is calculated, and the target sample The lead depth associated with this is corrected, and the corrected lead depth is used for the target sample. The copy number variation profile associated with the code can be determined. The system communicates with user terminals via a communication module (336) and a network (220). It may be provided to 210).

[0078] Figure 4 shows an example of a corrected dataset (400) according to one embodiment of the present disclosure. In Figure 4, the corrected dataset (400) is shown in table format, This is for illustrative purposes only and is not limited to the corrected dataset (400). These can be embodied and used in various forms, such as matrices, arrays, and vectors.

[0079] The corrected dataset (400) consists of multiple samples taken from multiple subjects different from the subject. A first area (410) containing data related to each of the samples, and based on this, A second region (42) containing data to correct the read depth associated with the get sample. 0) can be classified into this. In contrast, the data in the first area (410) is in the second area ( After being used to calculate the data in 420), it is included in the corrected dataset (400). It may not be included, or it may be removed from the corrected dataset (400).

[0080] In the corrected dataset (400), each row is a bin. This may indicate that each of the multiple bins (Bin1~Binn) represents a gene / genome analysis. It is a unit used to group or classify at least a portion of the genome. Obtain. In one embodiment, the size of each of the multiple bins is 100kbps (kilo The speed can be selected within the range of basepairs to 1Mbps (megabasepairs). ru.

[0081] Each of the multiple bins can be partitioned so that they do not share the same portion of the genome with the others. For example, if each of the multiple bins is 1 Mbps in size, then Bin1 is within a specific chromosome. The bases from position 1 to 1,000,000 are grouped, and Bin2 is 1,000 within the chromosome. The bases from position 0,001 to 2,000,000 can be grouped together. In contrast, multiple Each of the bins can be divided so that they share some of the same parts of the genome with each other. Bin1 groups the bases from position 1 to 1,000,000 within a specific chromosome, and B in2 groups the bases from position 500,001 to 1,500,000 within the chromosome in question. To obtain. Through this configuration, genomes are grouped in multiple bins, Noise (random noise or non-random noise) during depth calculation The instability of the target sample's read depth due to OISE can be significantly reduced.

[0082] The column in the first area (410) is a multi-process by processing type (FFPE, FF). The sample for each of the number of subjects is shown, and the second column (420) is the target object. Indicates the correction direction and / or correction size of the effect. Values ​​within the correction dataset (400). (value) is the read depth for each of the multiple bins in the first region (410) In the second region (420), the correction direction and correction size for each of the multiple bins are shown. The correction factor applied to the read depth can be shown. For example, as shown in Figure 4. The resulting "Depth_111" is a sample from a specific subject (normal cells) processed using FFPE. Read depth to the first bin, calculated from the sample or abnormal cell sample. This can be shown. At this time, the read depth of each of the multiple bins (Depth_111~Dep th_2nn (one of the above) has multiple read depths for each of the multiple bins. That is, the read depth for multiple reads classified into multiple bins. It can be a representative value, and is the arithmetic mean of multiple read depths. This could be equivalent to the median, etc.

[0083] In one embodiment, the correction direction and related information are provided for each of the multiple bins, The distribution of the entire FFPE-treated sample and its associated read depth p within the dataset, and For each of the multiple bins, the entire FF-processed sample in the dataset and related This can be determined based on the distribution q of the read depth. For example, it can be related to the correction direction for Bin1. The related information includes the distribution of Depth_111~Depth_11n and Depth_21 This can be determined based on the distribution of 1 to Depth_21n.

[0084] For example, to determine whether the distributions p and q for that bin are statistically significantly different. Results of a statistical test (e.g., Student's t-test with a p-value of 0.05) , when distributions p and q are significantly different (i.e., the mean and median of the two distributions are ra If there is a difference despite considering noise, the bin in question will be considered to have noise. The region is determined to be such that the correction direction and related information for that bin is "1" or "-1". It can be set to ". In contrast, if the distributions p and q are not significantly different, the bin can be set to The correction direction and related information can be set to "0".

[0085] Based on the distributions p and q, the entire FFPE-treated sample and its associated read depth (Or, the mean or median of the distribution p, etc.) is related to the entire FF-treated sample. If it is judged to be statistically larger than the read depth (or the mean or median of the distribution q, etc.) In that case, the correction direction and related information for the bin will be set to "-1" if it is judged to be small. The value can be set to "1".

[0086] If the correction direction and related information for a specific bin is "-1", then the Lee for that bin DoDepth should be corrected in the (-) direction, and if it's "1", LeadDepth should be corrected in the (+) direction. This can indicate that a value of "0" means that read depth correction is not necessary. Referring to Figure 4, for example, the read depth of the target sample relative to Bin1 is The depth should be corrected in the direction of decreasing the target sample for Binn. This may indicate that the read depth should be corrected in the direction of increasing read depth. On the other hand, the read depth of the target sample relative to Bin2 does not need to be corrected. This could demonstrate that.

[0087] In one embodiment, the correction size and related information are provided for each of the multiple bins, The average of all FFPE-processed samples in the set and their associated read depths, and For each of the multiple bins, the entire FF-processed sample in the dataset and the associated This can be determined based on the average of the depths. For example, it can be related to the correction size for Bin1. The information to be obtained is the average of Depth_111~Depth_11n and Depth_211~D This could be the absolute value of the difference between epth_21n and its mean.

[0088] In one embodiment, the correction coefficient applied to the lead depth is the correction method This can be determined by the product of the direction and the correction size. The correction determined for each of the multiple bins The coefficient is a correction for the target sample and associated read depth for each of the multiple bins. This can be used for the following purposes. This will be explained in detail later in Figures 5 and 6.

[0089] Figure 5 shows the target sample for each of the multiple bins according to one embodiment of the present disclosure. Figure 6 shows the primary correction process of the lead depth, and is a standard according to one embodiment of the present disclosure. This diagram shows the process by which the value of the correction variable (532) that minimizes the result of the quasi-deviation function is determined. For primary correction, multiple bins (primary bins) are converted into multiple secondary bins. They can be further classified into bin, bin-of-bin, and sub-bin. For example, about 30 For a subject's genome, which contains hundreds of millions of base pairs, it is possible to group 100,000 base pairs from each individual genome. Over 30,000 primary bottles were selected, and each of these 30,000 bottles was grouped with over 300 other primary bottles. It can be classified into more than 100 secondary bins that can be piped. In one embodiment, the individual secondary bins The number can be determined to be between 50 and 150, but is not limited to this.

[0090] As shown in Figure 5, the read depth data (510) for the secondary bin is obtained from multiple primary bins. The lead depth for each of the n is arranged vertically around the first reference line (512). (Indicated by a bar) may include. The first baseline (512) is the primary line associated with the target sample. Ground truth or FF processed read depth for each bin The read depth calculated from the target sample is shown, and for each of the multiple primary bins... The degree to which the lead depth deviates from the first reference line (512) is determined by the ground trout. This can show the difference from the read depth calculated from the target sample after h or FF processing. ru.

[0091] The correction coefficient data (520) is used to correct the read depth relative to the primary bin. It may include a correction factor. The second reference line (522) indicates a correction factor of 0, and the arrow indicating the correction factor is This can indicate the correction direction and correction size of the read depth for each of the multiple primary bins. In one embodiment, the correction coefficient is independent of the type of target sample, the object to be extracted, etc. It may correspond to a fixed value.

[0092] Target sample and associated read depth for each of the multiple bins (primary bins) The data can be linearly corrected using the following equations 1 and 2.

number

number

[0093] Here, X'iRDR is the first-order corrected read depth data for the i-th secondary bin. (540) XiRDR is read depth data before primary correction for secondary bins (510) POFi and adjust are the correction coefficients, α is the correction variable (532), and s() is the standard deviation function. (530) This is possible. That is, the correction variable (α)(532) is any variable in the correction coefficient. Multiply by the value (α') and use this to determine the read depth for each primary bin contained in the secondary bin. The arbitrary variable value (α') multiplied by the correction coefficient that minimizes the standard deviation of the result after adding it to the data. ) can then be determined as follows. Then, the reads for each of the primary bins contained in the secondary bin. The depth data includes a correction constant, which is the product of the correction variable (α)(532) and the correction coefficient. This can be corrected to a first-order correction by doing so. That is, read depth for each of the first-order bins The correction constant for first-order correction of sample data consists of a fixed correction coefficient and a target sample. It can be calculated by multiplying correction variables specific to the character.

[0094] In one embodiment, a single correction variable (α) is applied to a plurality of primary bins contained in a secondary bin. (532) is determined, and the determined correction variable (α)(532) is the same as that of multiple primary bins. This can be used for the read depth data for each of them. That is, for each of the secondary bins It may not contain the same primary bin as the primary bin contained in other secondary bins, and one primary bin A correction constant can be determined for this. In contrast, each of the secondary bins is the other 2 The next bin overlaps with the primary bins that other secondary bins contain. By including this, one or more correction constants can be determined for a single primary bin. The mean, median, or other values ​​of one or more determined correction constants may be used as the final correction constant. ru.

[0095] The standard deviation function (530) is a convex function with respect to the correction variable (α'). Therefore, there can always be one minimum value. Thus, the correction variable (α)(532) is gr It can be determined as a single variable value using methods such as ID search and line search. For example, as shown in Figure 6, d is repeatedly added to the read depth for each of the primary bins. The standard deviation can be calculated while doing so. In Figure 6, the standard deviation appears in a form that decreases and then increases. It can be confirmed that the standard deviation is minimized when d is added for the third time. Therefore, The value of the correction variable (α)(532) can be determined to be α=3d.

[0096] In summary, the first-order correction increases read depth to a similar degree across multiple samples. And using the decreasing pattern, if the calculated read depth is higher than the baseline, This could be considered a method of correcting lead depth by decreasing it and then increasing it if it has decreased. In this way, regions where read depth commonly increases or decreases are specific to a particular sample (single or (These are not limited to only plural forms) but are universally defined, and such domains exist. There is a great deal of direct and indirect evidence that it exists. For example, the GC-richness of sequences. The strength and charge distribution of DNA double bonds can be affected by ss, and as a result The degree to which DNA attempts to maintain its structure can vary from one another. Furthermore, DNA actually has a 3D structure, and is wound into nucleosomes and other structures. It exists in a coiled state like a tree, and is involved in epigenetic processes such as DNA or nucleosomes. 3D defined by epigenetic modification Because of their different structures, each DNA sequence exhibits various internal properties and resistance to external changes. They are different. In FFPE-treated samples, the tissue is preserved using It is known that DNA crosslinking can occur depending on the solution used, but crosslinking Areas where pounding is more or less likely to occur are greatly influenced by the factors mentioned above. It can receive the ripple effect.

[0097] Contrary to what is shown and explained in Figures 5 and 6, the distribution of read depth is specific to a narrow excessively increase or decrease over a certain region (that is, the distribution variation of region-to-region read depth is greater than or equal to a threshold value) by removing read depth data within the region, the read depth data can be corrected.

[0098] Figure 7 is a diagram illustrating a process of performing secondary correction for removing noise from primarily corrected read depth according to an embodiment of the present disclosure, and Figure 8 is a diagram illustrating an example of graphs (810, 820) and coefficient sets (830, 840) used in secondary correction according to an embodiment of the present disclosure. Secondary correction is a process of removing data noise, and after macroscopic variation in read depth occurring in large units in an FFPE-processed sample has been corrected through primary correction, secondary correction may correspond to a process of removing random variation in read depth occurring in regional units. Secondary correction may be performed through a decomposing step (710), a thresholding step (720) and a reconstruction step (730), and may be performed on the entire genome, or may be separately performed on a part of the genome (for example, a chromosome). 補正が行われる過程を示す図であり、図8は本開示の一実施形態による2次補正時に使用 されるグラフ(810、820)および係数セット(830、840)の例を示す図であ る。2次補正はデータのノイズを除去する過程であり、1次補正を通じてFFPE処理サ ンプルで大きい単位で現れる巨視的なリードデプスの変化量が補正された後、地域単位で 発生するランダムなリードデプスの変化量を除去する過程に該当し得る。2次補正は、分 解(decomposing)ステップ(710)、スレッショルディング(thres holding)ステップ(720)および再構築(reconstruction)ス テップ(730)を通じて行われ得、全ゲノムに対して実行されるか、またはゲノムの一 部(例えば、染色体)に対してそれぞれ行われ得る。

[0099] In the decomposing step (710), a level L for calculating a coefficient set may be set. The maximum value Lmax of L may be determined as log2(N) according to the length of input data or the number N of secondary bins, and level may be a natural number between 1 and Lmax. 最大値Lmaxは入力データの長さまたは2次ビンの数Nに応じてlog2(N)で決定 され得、levelは1ないしLmaxの間の自然数であり得る。

[0100] In the decomposing step (710), primarily corrected data may be processed in the form of a one-dimensional vector. For example, primarily corrected read depths for a plurality of bins may be expressed as a one-dimensional vector and processed. 得る。例えば、複数のビンに対する1次補正されたリードデプスが1次元ベクトルとして 表現され、処理され得る。

[0101] In one embodiment, if the data length N is not a multiple of 2, the data length N is greater than N. Zero padding may be added to the data so that it becomes a small multiple of 2. For example, 3, For read depth data for each of the 000 bins, there are 1096 zero paddy With the addition of , Lmax can be determined to be log2(4096) = 12. Through this process, the data is sequentially downsampled to 1 / 2 in the decomposition step (710) (do (wnsampling) is possible.

[0102] Then, wavelets are applied to the data at each level i=1, ..., Lmax. By applying the function (wavelet function (810) in Figure 8), detailed wavelets are obtained. The detailed wavelet coefficient (Wi) is a factor including the details of the wavelet coefficient. Several sets (the first coefficient set (830) in Figure 8) and approximate wavelet coefficients (appr Coefficient set including oximate wavelet coefficient (Vi) (Figure) The second coefficient set (840) for 8 can be determined.

[0103] The process by which the detailed wavelet coefficients and approximate wavelet coefficients are determined is given by the following equation. It is represented by 3.

number

[0104] Here, n represents the data index, which is an integer value between 0 and N-1. It can be held. When i=0, y0 (V0[n] in Figure 7) is the input for the decomposition step (710). With data (for example, first-order corrected read depth data for each of multiple bins) It's possible. f[n] is Haar, Daubechies, Coiflet, Best- Localized Daubechies or Least Asymmetric It can be selected and determined from various wavelets, such as wavelets. In this state, f[n] is a step function with discrete values. Most suitable for applying to state data (copy number variation (or read depth) profiles). A suitable Haar wavelet function (810) can be selected from Figure 8. The lett function (810) is a high-pass filter according to formula 4 below. Filter (HPF) h[n] and Low-Pass Filter It can be transformed and expressed as er,LPF)g[n]. That is, f[n] in equation 3 is H When using an ipass filter, the value can be h[n], and when using a lowpass filter, it can be g[n].

number

[0105] Here, 1x(n) is the indicator function, which has a value of 1 only at n=x and 0 in the remaining region. This corresponds to a function. A high-pass filter has a specific cutoff frequency (cut off freq). Frequencies above the specified frequency are allowed to pass through, while frequencies below that frequency are attenuated. By doing so, local data features are captured, while conversely, a low-pass filter, By attenuating frequencies above a specific cutoff frequency and allowing frequencies below it to pass through, It can capture macroscopic data features.

[0106] Again in equation 3, the input part yi is V0[n] in Figure 7 or the low-pass filter g Vi[n] that has passed through (n) and been downsampled to 1 / 2 (where i=1,2, ...,L-1). yi*f can be Vi[n] that has passed through a low-pass filter and been down sampled to 1 / 2 (where i=1,2,...,L-1), and the result of convolution of a low pass filter or a high-pass filter may be a function that is downsampled to 1 / 2 after down sampling. Thereafter, the output value yi+1[n], when passing through a low-pass filter, corresponds to the approximation wavele t coefficients Vi in FIG. 7 (where i=1,2,...,L), and when passing through a high-pass filter , corresponds to the detailed wavelet coefficients Wi in FIG. 7 (where i=1,2,...,L) .

[0107] In a thresholding step (720), the detailed wavelet coefficients determined in the decompositio n step (710) can be thresholded .

[0108] In one embodiment, the detailed wavelet coefficients may be subjected to sof t-thresholding. For example, detailed wavelet coefficients whose absolute value is equal to or less than a threshold are replaced with 0, and even if the absolute value exceeds the threshold, the closer it is to the threshold, the more the detailed wavele t coefficients can be greatly attenuated. That is, even if the absolute value of the detailed wavelet coefficient exceeds the threshold , the closer it is to the threshold, the more it contains a certain amount of noise, so soft thresholding can remove at least part of such noise. During soft thresholding , according to the soft thresholding function (820) in FIG. 8, Wi satisfying |Wi|<threshold ​This region is replaced with 0, and the remaining region can be linearly scaled.

[0109] Additionally or alternatively, detailed wavelet coefficients are used for hierarchical thresholding (h (Lateral thresholding) is possible. That is, level i = 1, ..., the higher Lmax, the lower the threshold value λ may become. Through such a configuration, The higher the level of detail in the wavelet coefficients, including macroscopic morphological information of the signal, the closer they are to 0. The degree of substitution or attenuation is low, and the macroscopic copy number variation profile is maximized. It can be preserved. That is, high-frequency noise is concentrated through hierarchical thresholding. It can be removed by a threshold λ with respect to level i, as shown in equation 5 below, where λ is the value of 1-α*i. Weighting can be assigned to these factors.

number

[0110] Here, N is the data length or the total number of secondary bins, and α is a predetermined arbitrary constant. And that value can be changed later. For example, the result of read depth correction, the noise is sufficiently removed. If it is determined that the noise has not been removed, the threshold is increased by decreasing the alpha value. Further removal is possible. σ is the mean absolute deviation (Equation 6 below). It can be estimated using deviation (MAD).

number

[0111] Subsequently, in the rebuilding step (730), the disassembly step The steps (710) can be implemented in reverse order. For example, the thresholded detailed way The Brett coefficients and their corresponding approximate wavelet coefficients are arranged alternately. Upsampling is performed, and g[-n] and h[-n] are applied to each level. Each convolution can be performed sequentially (see Equation 4 for functions g and h).

[0112] Figure 9 shows a first heatmap (9) associated with an FF processing sample according to one embodiment of the present disclosure. 10) Figure showing an example of the second heatmap (920) associated with the FFPE-treated sample. The indices shown along the vertical axis in Figure 9 were collected from different subjects. The samples shown were collected from the same subject, and samples assigned the same number were collected from the same subject. It could be simple. For example, the "FF1" and second heats of the first heatmap (910) Map (920) "FFPE1" was collected from the same subject but processed using different methods. It could be a sample that allows for this.

[0113] In the second heatmap (920), a striped pattern can be observed, but in the first heatmap ( In 910), it can be confirmed that no striped pattern exists. Second heatmap (920 The striped pattern of the sample was similar to a common degree at each position of multiple FFPE-treated samples. As the depth changes (i.e., increases or decreases), FFP This shows the specific noise that occurs in E-processed samples.

[0114] Figure 10 shows an FFPE-treated sample and associated lead data according to one embodiment of the present disclosure. Example of a distribution graph (1010, 1020) of read depth ratios for each sample before and after PUSS correction. This is a figure. The first graph (1010) shows the FFPE-treated sample and the associated The second graph (10) shows the mean distribution of the sample-specific read depth ratio before correcting for read depth. 20) The present disclosure complements the FFPE-treated sample and associated read depth. The mean distribution of read depth ratios per sample after correction is shown in the graph (1010, 10 The index shown along the vertical axis in 20) indicates the specific location on the genome's base sequence. show.

[0115] In the first graph (1010), the reads are determined by the sample processing method (FFPE, FF). It can be seen that the depth ratios follow different distributions. On the other hand, in the second graph (1020) This is because the read depth associated with the FFPE-treated sample is corrected, It can be confirmed that the read-to-depth ratios for different pull processing methods follow similar distributions. When determining the copy number variation profile (or read depth), FFPE-treated samples Even if used, the read depth correction performed according to this disclosure is similar to that which uses FF processed samples. It can be seen that similar data can be obtained.

[0116] Figure 11 shows an example of the distribution of read depth by chromosome according to one embodiment of the present disclosure. The graphs (1110, 1120, 1130, 1140) are displayed along their respective horizontal axes. The index indicates the autosomal chromosome number and sex chromosome, and the index is displayed along the vertical axis. The 'x' indicates the lead depth ratio. That is, the graph (1110, 1120, 1130, 11 40) Each black dot is located at a specific position in the base sequence of a normal cell sample. The read depth obtained as a result and the sequencing results of the abnormal cell sample This may represent the ratio between the read depth and the given lead depth, expressed in units of bins.

[0117] The first graph (1110) is a graph related to the FF processing sample, the second graph (1120) ) is a graph related to FFPE-treated samples from the same subject, and the third graph (1130) is the The data in graph (1120) is the graph associated with the data first amended by this disclosure. The fourth graph (1140) shows the data from the second graph (1120) as amended by this disclosure. This corresponds to the graph associated with the second-corrected data.

[0118] The distributions in the first graph (1110) and the second graph (1120) are slightly different, and the second graph ( It can be seen that noise occurred at a high frequency in graph 1120). On the other hand, in the third graph (113 In (0), it can be seen that the data has been corrected to be similar to the distribution in the first graph (1110). Furthermore, in the fourth graph (1140), noise is added from the data in the third graph (1130). By removing it, its distribution becomes very similar to the distribution in the first graph (1110). This can be confirmed.

[0119] Figure 12 shows the copy number variation profile obtained by read depth correction according to one embodiment of the present disclosure. SNR (signal-to-noise ratio) is used to evaluate the degree of qualitative improvement of the signal. This is a graph (1200). The SNR value is calculated for each sample (shown as a dot). Therefore, the SNR graph (1200) shown in Figure 12 shows the read depth for each sample. The first box plot (1210) shows the SNR value calculated before correction, and The second box plot shows the SNR value calculated after correcting the read depth of the FFPE-treated sample. (1220) is included. The SNR value can be calculated by the following formula 7.

number

[0120] Here, DFF is an interpretation of the read depth of the FF-processed samples as a random variable. DFFPE interprets the read depth of FFPE-processed samples as a random variable. be.

[0121] If the correction is performed only on the read depth of the FFPE-treated sample according to this disclosure, D FF remains the same before and after read depth correction, while DFFPE changes before and after read depth correction. This is possible. As a result, the SNR value in the first box plot (1210) corrects for read depth. Furthermore, the second box plot (1220) shows an overall increase. That is, Read It can be confirmed that depth correction improves the quality of the copy number variation profile.

[0122] Figure 13 shows a method for determining a copy number variation profile according to one embodiment of the present disclosure (1300). Flowchart. Method (1300) involves at least one processor (for example) If, the user terminal's processor, the information processing system's processor, or the copy number variation profile This can be done by a processor in a file determination device, etc. Method (1300) is a process The system obtains whole-genome analysis results associated with target samples taken from subjects. It can be initiated by (S1310). In one embodiment, the target sample is F FPE (Formalin-Fixed, Paraffin-Embedded) processing It could be a sample.

[0123] Subsequently, the processor uses the obtained whole-genome analysis results to analyze the genome (genom e) For each of the predetermined bins above, the target sample and related The connected read depth can be calculated (S1320).

[0124] Subsequently, the processor can correct the read depth associated with the target sample (S 1330).

[0125] In one embodiment, the processor uses multiple data collected from multiple subjects different from the subject. Includes read depth data for each of the multiple bins associated with each of the samples. Using the dataset, a first-order correction is performed on the read depth associated with the target sample. It is possible. In this case, the dataset consists of normal cells collected from each of multiple subjects. Sample and associated data, or abnormal cell samples taken from each of multiple subjects The data may include at least one of the data related to the subject, and may be collected from each of multiple subjects. Normal cell samples and abnormal cell samples taken from each of multiple subjects. Each of them is an FFPE-treated sample and an FF (Fresh Frozen) sample. It may contain a sample.

[0126] In one embodiment, the dataset corrects the read depth of the target sample. For use in this purpose, the correction direction of the lead depth for each of the multiple bins and related information Further information may be included. The correction direction of the lead depth for each of the multiple bins and related information. The report states that for each of the multiple bins, all FFPE-processed samples in the dataset The distribution of read depth associated with each body, and for each of the multiple bins within the dataset This can be determined based on the distribution of the entire FF-processed sample and associated read depth.

[0127] In one embodiment, the dataset corrects the read depth of the target sample. For use in this context, the correction size of the read depth for each of the multiple bins and related It may contain further information. The correction size and related information are provided for each of the multiple bins. The average of all FFPE-processed samples in the set and their associated read depths, and For each of the multiple bins, the entire FF-processed sample in the dataset and related This can be determined based on the average read depth.

[0128] The processor uses information related to the correction direction and information related to the correction size to determine the correction direction. A first-order correction can be performed by correcting the read depth associated with the acquired sample.

[0129] The processor then performs a second-order correction to remove noise from the first-order corrected read depth. Obtain. In one embodiment, the processor obtains a predetermined wavelet (wavel Using et) and first-order corrected read depth, multiple wavelet coefficients (wav Determine the elet coefficients and select multiple wavelet coefficients for threshold Thresholding multiple thresholded waves The second-order corrected read depth can be determined based on the Brett coefficient. The resulting wavelet may be a Haar wavelet.

[0130] The processor determines that the absolute value of a particular wavelet coefficient among several wavelet coefficients If it is below a predetermined threshold, a specific wavelet coefficient is replaced with 0, and a specific wavelet coefficient is replaced with 0. If the absolute value of the beblet coefficient is greater than or equal to the threshold, the smaller the difference between the absolute value and the threshold, the better. By attenuating (attenuating) specific wavelet coefficients to a high degree, Therefore, multiple wavelet coefficients can be thresholded. The higher the number of wavelet coefficients, the lower the threshold can become.

[0131] Subsequently, the processor uses the corrected read depth to interact with the target sample. Consecutive copy number variation profiles (p The profile can be determined (S1340). In one embodiment, the target sample This includes normal and abnormal cell samples taken from subjects, and a processor This involves read depth associated with normal cell samples and read depth associated with abnormal cell samples. S, and the allele frequency associated with the target sample. The copy number variation profile can be determined based on ncy.

[0132] The flowchart shown in Figure 13 and the above description are merely examples, and in some embodiments... It can be implemented differently. For example, one or more steps may be omitted, or the order of each step may be changed. The following changes, or one or more steps are repeated, or one or more steps are performed multiple times It can be repeated multiple times.

[0133] The above method is for a computer to execute on a computer on a computer-readable storage medium. It may be provided as a saved computer program. The medium is executable on a computer. You can either keep the program running or save it temporarily for execution or download. It may be such. Also, the medium may be in the form of one or more hardware components combined. It may be various recording or storage means, but is directly connected to a computer system. It is not limited to a specific medium, but can exist distributed across a network. Examples of mediums include... This includes hard disks, floppy disks, magnetic media such as magnetic tape, and CD-R Optical recording media such as OM and DVD, floppy disks Magneto-optical medium such as disk And program instructions are stored in ROM, RAM, flash memory, etc. It is possible that it is composed of such a system. Also, as an example of other media, it can be used to distribute applications. App stores and other sites, servers, etc. that supply or distribute various software. Recording media or storage media managed by them can also be mentioned.

[0134] The methods, operations, or techniques described herein can be embodied by various means. For example, these The technique involves hardware, firmware, software, or a combination of these. This can be realized. Various exemplary logic blocks and modules described in conjunction with the disclosure of this application. The steps, circuits, and algorithms are electronic hardware, computer software, etc. A typical engineer would understand that this can be realized through A, or a combination of both. To clearly explain this mutual substitution between hardware and software, various Exemplary components, blocks, modules, circuits, and steps are shown in their functional view. From that point of view, the above explanation was generally given. Whether such a function is embodied in hardware, Whether it is embodied as software or not depends on the specific application and system. It depends on the design requirements imposed on the entire system. A typical engineer would consider each specific application. While various methods may be used to realize the functions described, such realization may fall within the scope of this disclosure. It should not be interpreted as something that causes deviation.

[0135] Processing units used to execute techniques in hardware implementation. This refers to one or more ASICs, DSPs, or digital signal processing devices (digital sig nal processing devices (DSPDs), programmable logic devices Programmable logic devices (PLDs), fields Programmable gate array (field programmable gate array) Rays (FPGA), processor, controller, microcontroller, microp Rosser, electronic device, and other electronic devices designed to perform the functions described herein. It can be embodied in a unit, a computer, or a combination thereof.

[0136] Therefore, various exemplary logic blocks, modules, etc., described in conjunction with this disclosure, And the circuitry is a general-purpose processor, DSP, ASIC, FPGA or other programmable logic. Devices, discrete gates and transistor logic, discrete hardware components, or this To be embodied or accomplished in any combination of those designed to perform the functions described in the wish. This can be done. A general-purpose processor can be a microprocessor, but as an alternative, a processor The system can use any conventional processor, controller, microcontroller, or stateman. It can be thin. A processor is also a combination of computing devices, for example DSP and microprocessor, multiple microprocessors, one working in conjunction with the DSP core This can be embodied as the above microprocessors, or any other combination of configurations.

[0137] In firmware and / or software implementations, the technique is random access Random access memory (RAM), read-only memory (r ead-only memory (ROM), non-volatile RAM (non-volatile e random access memory; NVRAM), PROM (progr. ammable read-only memory), EPROM (erasable programmable read-only memory), EEPROM (e (Electrically erasable PROM), flash memory, compact Compact discs (CDs), magnetic or optical data storage devices It can be embodied as instructions stored on a computer-readable medium such as a chair. An instruction may be executable by one or more processors, and the processor (single or multiple) The number of devices may be made to perform certain aspects of the functions described herein.

[0138] The embodiments described above are currently being developed on one or more standalone computer systems. Although described as utilizing aspects of the subject matter shown, this disclosure is not limited to that, It can work with any computing environment, such as networks and distributed computing environments. It can be realized in this way. Furthermore, aspects of the subject matter in this disclosure include multiple processing chips. This can be embodied in devices, and storage can be similarly affected across multiple devices. These devices may include PCs, network servers, and portable devices.

[0139] Although this disclosure has been described in relation to some embodiments, the inventions of this disclosure belong to Various modifications and Modifications may be made. Furthermore, such modifications and changes are subject to the patents attached herein. It should be considered to fall within the scope of the request.

Claims

1. A method for determining copy number variation profiles by read depth correction, which is performed by at least one processor, The steps include obtaining whole-genome analysis results associated with target samples taken from subjects, Based on the obtained whole-genome analysis results, the step of calculating the read depth associated with the target sample for each of the predetermined bins on the genome, The steps include correcting the read depth associated with the target sample, Using the corrected read depth, the copy number variant (Cop) associated with the target sample is used. This includes the step of determining the y Number Variation profile, The step of correcting the read depth associated with the target sample includes the step of performing a primary correction of the read depth associated with the target sample based on the correction information, The correction information is determined for each of the plurality of bins based on a comparison between the read depth of a sample taken from a plurality of subjects different from the subject and treated with FFPE (Formalin-Fixed, Paraffin-Embedded) and the read depth of a sample taken from the plurality of subjects and treated with FF (Fresh Frozen), and includes the correction direction and correction size of the read depth for each of the plurality of bins, a method for determining copy number variation profiles.

2. The method for determining a copy number variation profile according to claim 1, wherein the target sample includes a normal cell sample and an abnormal cell sample taken from the subject, and the step of determining the copy number variation profile includes determining the copy number variation profile based on read depth associated with the normal cell sample, read depth associated with the abnormal cell sample, and allele frequency associated with the target sample.

3. The method for determining a copy number variation profile according to claim 1, wherein the target sample is a sample treated with FFPE (Formalin-Fixed, Paraffin-Embedded).

4. The copy number variation profiling method according to claim 1, wherein the read depth associated with the target sample includes the read depth for each of a predetermined number of bins on the genome of the target sample, and each of the read depths of the FFPE-treated sample and the FF-treated sample includes the read depth for each of the number of bins.

5. The copy number variation profile determination method according to claim 1, wherein the FFPE-treated sample or the FF-treated sample comprises at least one of the normal cell samples or abnormal cell samples taken from each of the plurality of subjects.

6. The method for determining a copy number variation profile according to claim 1, wherein the correction direction is determined for each of the plurality of bins based on a comparison between the distribution of read depths obtained from the FFPE-treated sample and the distribution of read depths obtained from the FF-treated sample.

7. The method for determining a copy number variation profile according to claim 1, wherein the correction size is determined for each of the plurality of bins based on a comparison between the average read depth obtained from the FFPE-treated sample and the average read depth obtained from the FF-treated sample.

8. The copy number variation profile determination method according to claim 1, wherein the step of performing the primary correction includes a step of correcting the read depth associated with the target sample based on the correction direction and the correction size.

9. The method for determining a copy number variant profile according to claim 1, wherein the step of correcting the read depth associated with the target sample further includes the step of performing a secondary correction to remove noise from the primary corrected read depth.

10. The method for determining a copy number variant profile according to claim 9, wherein the step of performing a secondary correction to remove noise from the first-order corrected read depth includes the steps of determining a plurality of wavelet coefficients using a predetermined wavelet and the first-order corrected read depth, thresholding the plurality of wavelet coefficients, and determining a secondary corrected read depth based on the thresholded plurality of wavelet coefficients.

11. The copy number variation profile determination method according to claim 10, wherein the predetermined wavelets correspond to Haar wavelets.

12. The method for determining a copy number variant profile according to claim 10, wherein the step of thresholding the plurality of wavelet coefficients includes the step of replacing a specific wavelet coefficient with 0 if the absolute value of that specific wavelet coefficient is less than a predetermined threshold, and the step of attenuating the specific wavelet coefficient to a higher degree if the absolute value of that specific wavelet coefficient is greater than or equal to the threshold, the smaller the difference between the absolute value and the threshold.

13. The method for determining a copy number mutation profile according to claim 12, wherein the threshold becomes lower as the specific wavelet coefficient reaches a higher level.

14. A computer-readable recording medium for performing the method according to any one of claims 1 to 13 on a computer.

15. A computing device, The system includes a communication module, a memory, and at least one processor connected to the memory and configured to execute at least one computer-readable program contained in the memory. The at least one program obtains whole-genome analysis results associated with a target sample taken from a subject, calculates the read depth associated with the target sample for each of a predetermined number of bins on the genome based on the obtained whole-genome analysis results, corrects the read depth associated with the target sample, and uses the corrected read depth to analyze the copy number variation (Cop) associated with the target sample. Includes instructions for determining the y Number Variation profile, Correcting the read depth associated with the target sample includes performing a primary correction of the read depth associated with the target sample based on the correction information. The correction information is determined for each of the plurality of bins based on a comparison between the read depth of a sample taken from a plurality of subjects different from the subject and treated with FFPE (Formalin-Fixed, Paraffin-Embedded) and the read depth of a sample taken from the plurality of subjects and treated with FF (Fresh Frozen), and includes the correction direction and correction size of the read depth for each of the plurality of bins, a computing device.

Images