RU2717809C1 - Computer-implemented integral method for assessing quality of target sequencing results - Google Patents

Abstract

FIELD: biotechnology.

SUBSTANCE: invention relates to the field of biotechnology. Disclosed is a computer-implemented integral method for assessing the quality of target sequencing results. Method involves obtaining a user sample, targeted sample sequencing, obtaining sequencing characteristics of sequencing characteristics and formation of integral characteristic of sequencing by using machine learning algorithm.

EFFECT: invention provides higher speed of assessing the quality of the results of targeted sequencing.

6 cl, 7 dwg

Description

FIELD OF TECHNOLOGY

[001] This technical solution, in General, relates to the field of computer technology, and in particular, to methods and systems for assessing the quality of the results of targeted sequencing in the field of bioinformatics.

BACKGROUND

[002] Currently, a new generation sequencing technology (NGS) is widely used in clinical practice. However, until now, the cost of one study using NGS technology has remained quite high, which limits the widespread use of this method. One of the factors affecting the cost is the choice of coating during sequencing, that is, the number of times each nucleotide has been sequenced, and therefore the problem of assessing the quality of sequencing results is urgent.

[003] A prior art scientific article, "A quality control tool for high throughput sequence data" (publ. January 2014, author: S. Andrews), in which, when evaluating the quality of sequencing, rely on the quality assessment of many factors, is known. For example, such as the reading quality of each reading nucleotide, the percentage of occurrence of a particular nucleotide in each specific reading position, GC content, distribution of k-measures, content of overrepresented sequences, content of adapter sequences, coverage of various parts of the reference sequence, etc. These factors are considered and taken into account independently of each other.

[004] However, in the prior art there are no methods for integrally evaluating the quality of the results of targeted sequencing. In routine practice, evaluating the quality of sequencing on the basis of disparate metrics that evaluate different characteristics, such as the percentage of sequencing errors, the percentage of short fragments, the quality of the preparation of the sequencing library, and the presence of impurities in the sample, has insufficient scalability, since with increasing the number of samples, the time to evaluate The quality of each of them can be unacceptably large.

SUMMARY OF THE INVENTION

[005] This technical solution is aimed at eliminating the disadvantages inherent in solutions known from the prior art.

[006] The technical task or problem to be solved in this technical solution is to provide an assessment of the quality of the results of targeted sequencing.

[007] The technical result manifested in solving the above technical problem is to increase the speed of evaluating the quality of the results of targeted sequencing.

[008] An additional technical result is to improve the accuracy of evaluating the quality of the results of targeted sequencing.

[009] The specified technical result is achieved through the use of machine learning in the formation of an integral quality assessment by combining a large number of characteristics of the sequencing results.

[0010] In general, the specified technical result is achieved by implementing an integral method for evaluating the quality of the results of targeted sequencing in which at least one sample of at least one user is obtained; performing targeted sequencing of at least one sample obtained in the previous step; derive sequencing characteristics from sequencing results; form an integral characteristic of sequencing through the use of machine learning algorithm.

[0011] In some embodiments, the sequencing characteristic is the average coverage of the sample and / or the number of amplicons coated on average more than the average coverage of the sample, and / or the percentage of nucleotide pairs with quality values greater than a predetermined, and / or the percentage of readings that have been trimmed.

[0012] In some embodiments of the invention, gradient boosting of decision trees is used to form the integral characteristic.

[0013] In some embodiments of the invention, a quality measure representing a number from the range from 0 to 1 is used in the formation of the integral characteristics of sequencing.

BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The features and advantages of this technical solution will become apparent from the following detailed description and the accompanying drawings, in which:

[0015] In FIG. 1 shows an example implementation of a method for evaluating the quality of targeted sequencing results.

[0016] FIG. 2 shows the organization of the training sample in the cross-validation.

[0017] In FIG. Figure 3 shows the selection of parameters using cross-validation.

[0018] In FIG. Figure 4 shows the process of training a model using a training set.

[0019] FIG. 5 shows an example of calculating the average coverage.

[0020] In FIG. Figure 6 shows an example of an example of assessing the presence of formalin artifacts in a sample.

[0021] In FIG. 7 shows an embodiment of a system for evaluating the quality of targeted sequencing results.

DETAILED DESCRIPTION OF THE INVENTION

[0022] This technical solution can be implemented on a computer or other data processing device in the form of an automated system or computer-readable medium containing instructions for performing the above method.

[0023] The technical solution can be implemented in the form of a distributed computer system, the components of which are cloud or local servers.

[0024] In this decision, a system is understood to mean a computer system or an automated system (AS), a computer (electronic computer), CNC (numerical control), PLC (programmable logic controller), a computerized control system, and any other devices capable of performing a given, clearly defined sequence of computational operations (actions, instructions).

[0025] An instruction processing device is understood to mean an electronic unit or an integrated circuit (microprocessor) executing machine instructions (programs).

[0026] An instruction processing device reads and executes machine instructions (programs) from one or more data storage devices. Storage devices may include, but are not limited to, hard disks (HDDs), flash memory, ROM (read only memory), solid state drives (SSDs), optical drives, and cloud storage.

[0027] A program is a sequence of instructions for execution by a computer control device or an instruction processing device.

[0028] DNA sequencing - determining the sequence of nucleotides in a DNA molecule. This can mean both amlicon sequencing (reading the sequences of isolated DNA fragments obtained by PCR reaction - such as the 16S rRNA gene or its fragments), and genome-wide sequencing (reading the sequences of all DNA present in the sample).

[0029] Readings (sometimes reads from the English. "Reads") are data representing the nucleotide sequences of DNA fragments obtained using a DNA sequencer.

[0030] FASTA is a format for recording DNA sequences.

[0031] FASTQ is a DNA sequence recording format in which the hardware reading quality of each position is recorded.

[0032] Reading mapping is a bioinformatic method for analyzing the results of a new generation sequencing, which consists in determining positions in the reference base of genomes or genes, from which each specific short reading could most likely be obtained.

[0033] DNA sequencing creates a set of readings. The reading length of modern sequencers ranges from several hundred to several thousand nucleotides.

[0034] Amplikon (Engl. "Amplicon") - the nucleotide sequence of DNA, multiplied (amplified) using PCR.

[0035] The decision tree is a binary classification algorithm based on machine learning with a teacher. In other words, before making the classification, a tree is constructed, which is a structure in which the conditions for signs in the internal nodes and the classes are in the leaves.

[0036] A training sample (English “training sample”) is a sample by which the dependence model is tuned (parameter optimization).

[0037] Cross-validation, sometimes referred to as cross-validation, is a model validation technique for verifying how successfully used statistical analysis in a model is capable of working on an independent data set.

[0038] The method for evaluating the quality of the results of targeted sequencing implemented by the processor shown in FIG. 1 in a block diagram may include the following steps.

[0039] Step 110: at least one sample of at least one user is obtained.

[0040] In this technical solution, the user can be understood as a patient.

[0041] At least one user DNA sample is pre-prepared for sequencing. The sample should be a biological material from which it is possible to isolate DNA. As such biomaterial for the analysis of mutations associated with cancer, various types of biological material containing DNA can be used, for example: tumor tissue samples (biopsy or surgical material) in the form of a frozen preparation or paraffinized blocks, circulating DNA from blood plasma or other human fluid media, as well as whole blood or a cheek swab, as a source of normal tissue for analysis of hereditary mutations.

[0042] Different types of samples have different requirements.

[0043] In some embodiments, the compliance of the biomaterial with the requirements is determined by the processor based on previously entered parameters of the obtained sample and the criteria for suitability of the biomaterial. Below are the main criteria for the suitability of biomaterial. For example, for the sample “Paraffinized tumor samples fixed in formalin (paraffin blocks)”, the indicators should be as follows. The volume of fabric in the block is at least 10 mm3, and the surface area of the fabric in the block is at least 5 mm2. The minimum tumor content along the accompanying sections from the upper and lower surfaces of the analyzed fragment, the percentage of area of the malignant tissue should be at least 10% (according to cell nuclei). Samples obtained after acid decalcification of bone tissue are unsuitable for analysis due to DNA degradation. The use of unbuffered formalin during fixation or too long fixation of tissue leads to DNA damage and may render the sample unsuitable for analysis.

[0044] For example, for a Liquid Biopsy sample (circulating DNA analysis), the requirements are as follows:

[0045] A blood sample is collected in a tube with K3EDTA (5-10 ml);

[0046] For 10 minutes, centrifuged to precipitate the cell pellet and separate the plasma;

[0047] 1 ml plasma is taken into SafeLock type microtubes (1.5 ml);

[0048] Centrifugation at a speed of 12-20 thousand rpm;

[0049] It is immediately frozen at -80 ° C or frozen in liquid nitrogen.

[0050] A user sample is preliminarily obtained.

[0051] The above primary data is obtained by using a sampling kit, which may include a sample container having a process reagent component and configured to receive a sample from a collection site by a user, which may be remote. Additionally or alternatively, a sampling kit can be provided directly through a sample collection device installed in a room or on the street, which is intended to facilitate receiving a sample from a user. In other embodiments, a sampling kit may be delivered at a clinic or other medical facility to medical laboratory equipment, and previously delivered to the user, for example, by courier. However, providing the kit (s) for user sampling to the system may additionally or alternatively be carried out in any other suitable way, for example, frozen in a sterile container.

[0052] Step 120: perform targeted sequencing of at least one sample obtained in the previous step by means of a sequencing platform;

[0053] Obtaining high-performance targeted sequencing data involves several steps, which may differ depending on the particular set of methods used, the selected sequencing technology, and the type of biological material. Genomic DNA is isolated by appropriate sets of reagents depending on the type of sample: using the sorbent method or the method of heterophasic extraction. The concentration of the isolated double-stranded DNA should be at least 1 ng / μl. DNA can be stored at 37 ° C for 10-20 minutes, at 2-8 ° C - for 12-24 hours, at -20 ° C - for a long time.

[0054] When analyzing the spectrophotometric purity of a DNA preparation, the ratio of the absorption of the solution at wavelengths of 260 and 280 nm is determined in the range of 1.8-2.0.

[0055] For the preparation of DNA libraries receive biomaterial in a volume equivalent to at least 50 ng of DNA. At the next stage, manipulations are performed in accordance with the manufacturer’s instructions for targeted enrichment kits (if applicable) and a library preparation kit. The present invention is applicable to various types of DNA libraries, including those obtained by targeted enrichment by multiplex PCR or hybridization with specific RNA / DNA probes.

[0056] Target sequencing is performed in a mode providing an average coverage of target regions of at least 200x. The invention was tested on the lllumina and Torrent sequencing platforms, but given the unified sequencing data format, it can be applied to other platforms and high-performance sequencing technologies.

[0057] However, the sequencing results have a uniform format and the invention can be applied to them.

[0058] Step 130: obtain sequencing characteristics from the sequencing results.

[0059] In this step, by using the processor, the following steps are performed sequentially:

1) Mapping readings to the reference genome of H. sapiens;

2) Count the number of readings covering each position of the reference;

3) Determine the set of mutations;

4) Determine the mutation spectrum for the sample.

[0060] This technical solution uses an integral characteristic that allows you to quantify the quality of a sample based on a set of characteristics of sequencing data.

[0061] In some embodiments of the invention, the average coating of the sample is used for an overall integrated characteristic. This parameter characterizes how fully the genomic sequence has been read. The greater its value, the greater the likelihood that all target regions have been read and all genetic information contained in a biological sample is presented more fully. The average coverage of a sample is defined as the number of reads multiplied by their length and divided by the size of the target read nucleotide sequence. For example, 100 million readings of 100 bp were read. It was read completely by the 4 billion bp human genome. As a result, the average coverage is defined as 100 * 10 ^ 6 * 100/4 * 10 ^ 9 = 2.5.

[0062] In some embodiments, the number of amplicons coated on average with more than the average coating of the sample is used. This parameter characterizes how many target sequences covered are larger than the average coverage value. Thus, it is possible to understand how many amplicons are overrepresented in the sample and to understand how the target sequences are unevenly represented.

[0063] Also used is the percentage of nucleotide pairs (abbreviated bp) with quality values greater than Q30 (for the Illuminumin sequencer) and Q20 (for the lonTorrent sequencer). Quality values show how well a particular nucleotide is read in sequences obtained using a sequencer. The threshold in terms of the quality of separating nucleotides read with high accuracy is specific for different sequencing technologies and counting the percentage of nucleotides read with quality values above these thresholds, we can understand how successful the launch is.

[0064] In another embodiment, the percentage of readings that have been trimmed is used, which is understood as the process of removing regions containing nucleotides with low quality values from the reading sequence. Thus, it is possible to estimate the percentage of saved readings after the work of the algorithm for deleting the regions of readings read with low quality. And the percentage of such readings is large enough, it can be judged that the sequencing was successful.

[0065] You can also use the assessment of the presence of formalin artifacts from the mutation spectra and the total number of mutations found, which is carried out as follows. The number of all found single nucleotide substitutions is calculated and the distribution of the number of substitutions that are different in their types in the sample is constructed. If the number of single nucleotide substitutions C> T and G> A significantly prevails over the others, it is believed that formalin artifacts are present in the test sample, as shown in FIG. 6. In other words, if the ratio ([C> T] + [G> A]) / ([C> A] + [C> G] + [G> C] + [G> T] + [T> A] + [T> C] + [T> G] + [A> C] + [A> G] + [A> T]) is greater than 2 (in the presence of formalin artifacts, the number of substitutions of type C in T and G in A in biological samples, in connection with which it is empirically determined that the number of substitutions of this type grows by an order of magnitude compared to the total number of other substitutions), it is determined that formalin artifacts are present in the sample, where [C> T] is the number of substitutions C in T, [G> A] is the number of substitutions of G in A, etc.

[0066] You must understand that the above set of sequencing characteristics is not limited to use, and additional sequencing characteristics can be added, and therefore it will be possible to take into account the features of new sequencing technologies that have appeared in the prior art.

[0067] Step 140: an integral sequencing characteristic is generated based on the sequencing characteristics obtained in the previous step by using a machine learning algorithm on the processor.

[0068] The integral characteristic can be obtained using the machine learning algorithm XGBoost (gradient boosting). As alternative algorithms for machine learning, random trees, the support vector method, neural networks, etc. can be used. An exemplary implementation option is shown on gradient boosting of decision trees.

[0069] The decision tree itself is constructed from a training set. A training sample is a set of objects for which it is precisely known to which class this or that object belongs. Boosting is one of the methods of constructing a composition of simple classifiers, with each subsequent classifier under construction trying to compensate for the shortcomings of the previous composition of algorithms. In the case of XGBoost, a decision tree is used as a simple classification algorithm, with a rather small depth.

[0070] To train the model, namely, the composition of decision trees, a training set is used that has a minimum of 2 * N samples with satisfactory quality and N samples with unsatisfactory quality, where N is the number of sequencing characteristics. These restrictions on the number of samples are related to the fact that if the number of samples is less than the number of characteristics, then the constructed ensemble of decision trees can simply remember the training set and will obviously be retrained, that is, it cannot adequately classify objects that are not included in the training set. Moreover, it has been empirically found that in order to obtain a satisfactory quality of classification, a training sample of samples that are satisfactory in quality is required to at least double the number of attributes in terms of volume.

[0071] One of the features may be the average coverage of the sample, which is estimated as the ratio of the total coverage of the sample to the length of the genomic sequence of the sample. Suppose there is a reference nucleotide sequence, which consists of 10 pairs of nucleotides. Coating values for each position of 10 pairs can be as follows: 1 - coating 1, 2 -1, 3 - 3, 4 - 3, 5 - 3, 6 - 3, 7 - 3, 8 - 3, 9 - 3, 10 - 3. Thus, the total coverage will be 26.

[0072] Another indication is the number of amplicons coated on average with more than the average coating of the sample. This feature is evaluated as follows:

[0073] a) For each amplicon, its average coverage is estimated;

[0074] b) The average amplicon coating is compared with the average coating of the sample. Then, the number of amplicons is determined for which the average coating is larger than the average coating of the sample.

[0075] Another sign is the percentage of bp with quality values greater than Q30 (for the Illumin sequencer) and Q20 (for the lonTorrent sequencer). The value of quality is determined by the proprietary sequencer software. This attribute for each read position of the sample indicates the probability of correct reading in the -10log logarithmic scale as the probability of error. Q30 is the probability of error 10 ^ -3, and Q20 is the probability of error 10 ^ -2.

[0076] Another indication for the training sample is the percentage of readings that have been trimmed (the process of removing regions containing nucleotides with low quality values from the reading sequence).

[0077] Also, an assessment of the presence of formalin artifacts from the mutation spectra and the total number of mutations found is used as a feature.

[0078] The process of selecting model training parameters is performed using cross-validation, which is used by the processor, as shown in FIG. 4. The training sample is divided, for example, into 10 parts and for each stage of training prediction is carried out in 9 parts, after which the prediction is estimated based on the remaining part. The objects of the training sample can be samples that are pre-classified, for example, by the user or other machine learning algorithm, as having a satisfactory level of quality and an unsatisfactory level of quality. During this learning process, parameters are selected for the algorithm constructing the classifier. The selection of parameters is necessary to optimize the constructed classifier, which is why cross-validation is used. Namely, the training sample is divided into a set of subsamples, a set of subsamples without one is used for training, and the excluded subsample is used to determine the quality of clustering, and so it is repeated for all subsamples as test ones, as shown in FIG. 3. A variation of the training set is made and predictions are checked, on the data set of which a quality assessment is known. For the implementation, when XGBoost is used, the following parameters are sorted out (as shown in Fig. 2):

[0079] eta (parameter XGBoost, indicating how many times it is necessary to reduce the step size after each iteration) (range of values from 0.01 to 0.4 with a step of 0.01);

[0080] min_child_weight (minimum weight of the child node) (range of values from 0.3 to 1 in steps of 0.01);

[0081] max_depth (maximum depth of formed decision trees) (from 3 to 10 in increments of 0.01); subsample (part of the training sample used in training XGBoost) (from 0.3 to 1 in steps of 0.01);

[0082] colsample_bytree (part of the training sample used in training at each XGBoost iteration) (from 0.3 to 1 in steps of 0.01);

[0083] colsample_bylevel (part of the training sample used in training for each construction of internal nodes in decision trees at each XGBoost iteration) (from 0.3 to 1 in steps of 0.01);

[0084] the scale_pos_weight parameter (a value indicating how much more valuable the samples of the positive training sample are relatively negative; in our case, the positive sample is samples of satisfactory quality and the negative sample is samples of unsatisfactory quality) should be equal to the ratio of the value of the positive sample to negative or close to given relation.

[0085] Parameters are enumerated as follows, as shown in FIG. 5.

[0086] Previously, all parameters are set to a minimum value.

[0087] Then cross-validation is carried out, where the classification quality assessment is determined. As such an estimate, the AUC value can be used - the area under the ROC curve. The ROC curve is the relationship between the sensitivity and specificity of the classification algorithm.

[0088] Next, the eta parameter increases by 0.01, after which cross-validation is again performed and the quality of classification is evaluated.

[0089] If eta is less than 0.4, then the algorithm continues with cross-validation.

[0090] If min_child_weight is less than 1, then this parameter is increased by 0.01 and the algorithm continues with cross-validation and eta is set to 0.01.

[0091] If max_depth is less than 10, then this parameter is increased by 0.01 and the algorithm continues with cross-checking, and eta is set to 0.01, min_child_weight is set to 0.3.

[0092] If the subsample is less than 1, then it increases by 0.01 and the algorithm continues with cross-validation, the eta parameter is set to 0.01, the min_child_weight parameter is set to 0.3, and the max_depth parameter is set to 3.

[0093] If colsample_bytree is less than 1, then it increases by 0.01, the algorithm continues with cross-validation, eta is set to 0.01, min_child_weight is set to 0.3, max_depth is set to 3, and subsample is set to 0.3.

[0094] If colsample_bylevel is less than 1, then it increases by 0.01, the algorithm continues with cross-validation, eta is set to 0.01, min_child_weight is set to 0.3, max_depth is set to 3, subsample is set to 0.3 and colsample_bytree is set to 0.3 .

[0095] As a result, a set of parameters is selected for which the AUC value is maximum. And this set of parameters is the result of cross-validation. Training is carried out on the entire training sample with this set of parameters.

[0096] The resulting trained XGBoost model can then be used to assess the quality of samples obtained using the same sequencers from which the training sample was obtained. In the case of changing equipment for sequencing, it is necessary to retrain the model, namely, it is necessary to form a training sample on samples obtained using new equipment and to form a new model. In some implementations, universal models are formed that are suitable for a wide range of sequencing plants, for which a sufficiently large set of samples is used (for example, at least 1 ^ -10 thousand * [number of evaluated parameters] samples in the training set) obtained using several different sequencing plants built both on the basis of one technology and several. Thus, to assess how the quality of the samples varies both within the framework of sequencers built on the basis of one technology and between different technologies, a set of samples obtained using the most diverse equipment possible is necessary. The adequacy of the current number of samples can be estimated by the average AUC parameter during cross-validation. It should be at least 0.8 (classifiers with AUC greater than 0.8 are considered good, as is known from the source of information [2]).

[0097] To conduct the analysis using a trained XGBoost model for the analyzed sample (set of readings), the attributes are determined in a manner similar to that used when working with the training sample, using the processor. The list of attributes is not specified, as it was shown above that this set is flexibly selected at the training stage. Further, on the basis of these characteristics, an assessment is made using, based on XGBoost, a sample quality classifier. Signs of the sample are fed to the input of the classifier. Based on these characteristics, it gives a value in the range from 0 to 1. If the value obtained using the classifier exceeds 0.5, then it is believed that the sample has a satisfactory level of quality. If the value is less than 0.5, then unsatisfactory. If the predicted value of XGBoost is 0.5, then the quality of the sample cannot be estimated.

[0098] As a result of applying the developed technical solution to the data obtained as a result of sequencing, a quality measure is obtained by the processor, which is a number from the range from 0 to 1.

[0099] An example implementation of the above technical solution by a processor is shown below.

[00100] The following features are used for training: average coverage,% of readings that have been trimmed.

[00101] The following objects are used for training:

[00102] Positive selection (satisfactory quality):

[00103] First object: average coverage - 500x,% of reading after trimming 90%

[00104] Second object: average coverage - 400x,% of reading after trimming 80%

[00105] Third object: average coverage - 600x,% of reading after trimming 70%

[00106] Fourth object: average coverage - 400x,% of reading after trimming 75%

[00107] Negative sampling (poor quality):

[00108] First object: average coverage - 100x,% of reading after trimming 40%

[00109] The second object: the average coverage is 400x,% of reading after trimming 30%

[00110] Next, a classifier is constructed using this training set.

[00111] The objects for which quality is evaluated are taken. Suppose these are two objects with such signs:

First object: average coverage - 400x,% of reading after trimming 78%

[00112] Second object: average coverage - 100x,% of reading after trimming 34%

[00113] These objects are fed to the input of the classifier built in paragraph 3. For the first object, we obtain a rating of 0.87. For the second 0.34. 0.87> 0.5 - the first tested object of satisfactory quality. 0.34 <0.5 - the second tested object of unsatisfactory quality.

[00114] Referring to FIG. 7, this technical solution can be implemented in the form of a computing system 700, which contains one or more of the following components:

a processing component 701 comprising at least one processor 702,

- memory 703,

- power component 704,

- component 705 multimedia,

- component 706 audio,

- an input / output (I / O) interface 707,

- sensor component 708,

- component 709 data transfer.

[00115] The processing component 701 mainly controls all operations of the system 700, for example, generates an integral sequencing characteristic by using a machine learning algorithm, and also controls the display, telephone call, data transfer, camera operation, and recording operation of a user’s mobile communication device. Processing module 701 may include one or more processors 702 that implement instructions for completing all or part of the steps of the above methods. In addition, the processing module 701 may include one or more modules for a convenient interaction process between other processing modules 701 and other modules. For example, the processing module 701 may include a multimedia module for conveniently facilitating interaction between the multimedia component 705 and the processing component 701.

[00116] The memory 703 is configured to store various types of data to support the operation of the system 700, for example, a database with user profiles. Examples of such data include instructions from any application or method, contact data, address book data, messages, images, videos, etc., all of which operate on system 700. Memory 703 may be implemented as any type of volatile a memory device, non-volatile memory device, or a combination thereof, for example, static random access memory (RAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Standing only memory (EPROM), programmable ROM (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk and the other is not limited.

[00117] Power component 704 provides electricity to various components of system 700. Power component 704 may include a power management system, one or more power sources, and other nodes for generating, controlling, and distributing power to system 400.

[00118] The multimedia component 705 includes a screen providing an output interface between a system 700 that can be installed on a mobile user and user communication device. In some implementations, the screen may be a liquid crystal display (LCD) or touch panel (SP). If the screen includes a touch panel, the screen may be implemented as a touch screen for receiving an input signal from a user. The touch panel includes one or more touch sensors in the sense of gesturing, touching and sliding the touch panel. The touch sensor can not only feel the border of the user's touch or the gesture of turning over, but also determine the duration of time and pressure associated with the operation mode of touching and sliding. In some embodiments, the media component 705 includes one front camera and / or one rear camera. When the system 700 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive multimedia data from outside. Each front camera and rear camera may be a single fixed optical system of the lens or may have a focal length or optical zoom.

[00119] The audio component 706 is configured to output and / or input an audio signal. For example, the audio component 706 includes a single microphone (MIC), which is configured to receive an external audio signal when the system 700 is in an operation mode, for example, a call mode, a recording mode, and a speech recognition mode. The resulting audio signal may be further stored in the memory 703 or routed through the data transmission component 709. In some embodiments, the audio component 706 also includes one speaker configured to output an audio signal.

[00120] An I / O interface 707 provides an interface between the processing component 701 and any peripheral interface module. The above peripheral interface module may be a keyboard, steering wheel, button, etc. These buttons may include, but are not limited to, the start button, volume button, start button, and lock button.

[00121] The sensor component 708 contains one or more sensors and is configured to provide various aspects of assessing the state of the system 700. For example, the sensor component 708 can detect on / off states of the system 700, the relative position of components, such as a display and keypad, of one component system 700, the presence or absence of contact between the user and the system 700, as well as the orientation or acceleration / deceleration and temperature change of the system 700. The sensor component 708 contains a proximity sensor, with the ability to detect the presence of an object in the vicinity when there is no physical contact. The sensor component 708 includes an optical sensor (for example, CMOS or CCD image sensor) configured to be used in visualization of the application. In some embodiments, the sensor component 708 comprises an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[00122] The data transmission component 709 is configured to facilitate wired or wireless communication between the system 700 and other devices. System 700 may access a wireless network based on a communication standard such as WiFi, 2G, 3G, 5G, or a combination thereof. In one exemplary embodiment, the data transmission component 709 receives a broadcast signal or broadcast, related information from an external broadcast control system via a broadcast channel. In one embodiment, the data transmission component 709 comprises a near field communication (NFC) module to facilitate near field communication. For example, the NFC module may be based on radio frequency identification (RFID) technology, infrared data association technology (IrDA), ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[00123] In an exemplary embodiment, system 700 may be implemented by one or more Specialized Integrated Circuits (ICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPs), Programmable Logic Devices (PLCs), logic circuits programmed in operating conditions (ППВМ), controller, microcontroller, microprocessor or other electronic components, and can be configured to implement the album display method.

[00124] In an exemplary embodiment, the non-volatile computer-readable medium contains instructions also provided, for example, memory 703 includes instructions where instructions are executed by processor 701 of system 700 to implement the above-described methods for the automated design of a multimodal freight transportation service. For example, a non-volatile computer-readable medium may be a ROM, random access memory (RAM), compact disk, magnetic tape, floppy disks, optical storage devices and the like.

[00125] Computing system 700 may include a display interface that transmits graphics, text, and other data from a communications infrastructure (or from a frame buffer, not shown) for display on multimedia component 705. Computing system 700 further includes input devices or peripherals. Peripheral devices may include one or more devices for interacting with a user's mobile communication device, such as a keyboard, microphone, wearable device, camera, one or more sound speakers, and other sensors. Peripheral devices may be external or internal to the user's mobile communication device. The touch screen can typically display graphics and text, and also provides a user interface (for example, but not limited to a graphical user interface (GUI)) through which the user can interact with the user's mobile communication device, for example, access and interact with applications running on the device.

[00126] Elements of the claimed technical solution are in a functional relationship, and their joint use leads to the creation of a new and unique technical solution. Thus, all blocks are functionally connected.

[00127] All the blocks used in the system can be implemented using electronic components used to create digital integrated circuits, which is obvious to a person skilled in the art. Not limited to, can be used microcircuits, the logic of which is determined during manufacture, or programmable logic integrated circuits (FPGA), the logic of which is set by programming. For programming, programmers and debugging environments are used that allow you to specify the desired structure of a digital device in the form of a circuit diagram or a program in special equipment description languages: Verilog, VHDL, AHDL, etc. Alternative FPGAs can be programmable logic controllers (PLCs), base matrix crystals ( BMK) requiring a factory production process for programming; ASIC - specialized custom large integrated circuits (LSI), which are much more expensive in small-scale and single-unit production.

[00128] Typically, the FPGA chip itself consists of the following components:

- Configurable logical blocks that implement the required logical function;

- programmable electronic communications between configurable logic blocks;

- programmable I / O blocks, providing the connection of the external output of the chip with internal logic.

[00129] Also, blocks can be implemented using read-only memory devices.

[00130] Thus, the implementation of all used blocks is achieved by standard means based on the classical principles of implementation of the foundations of computer technology.

[00131] As one skilled in the art will understand, aspects of the present technical solution may be implemented as a system, method, or computer program product. Accordingly, various aspects of the present technical solution can be implemented solely as hardware, as software (including application software and so on) or as an embodiment combining software and hardware aspects, which in general can be referred to as a “module” , “System” or “architecture”. In addition, aspects of the present technical solution may take the form of a computer program product implemented on one or more computer-readable media having a computer-readable program code that is implemented on them.

[00132] Any combination of one or more computer-readable media may also be used. The computer-readable storage medium may be, without limitation, an electronic, magnetic, optical, electromagnetic, infrared or semiconductor system, apparatus, device, or any suitable combination thereof. More specifically, examples (non-exhaustive list) of computer-readable storage media include: electrical connection using one or more wires, a portable computer diskette; hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), fiber optic connection, read-only memory on a compact disc (CD-ROM), optical storage device, magnetic storage device or any combination of the above. In the context of the present description, a computer-readable storage medium may be any flexible data medium that can contain or store a program for use by or in connection with the system, device, apparatus.

[00133] The program code embedded in a computer-readable medium can be transmitted using any medium, including, without limitation, wireless, wired, fiber optic, infrared, and any other suitable network or combination of the above.

[00134] The computer program code for performing operations for the steps of the present technical solution can be written in any programming language or combinations of programming languages, including an object-oriented programming language, such as Java, Smalltalk, C ++ and so on, and conventional procedural programming languages , for example, the programming language "C" or similar programming languages. The program code can be executed on a user's computer completely, partially, or as a separate software package, partially on a user's computer and partially on a remote computer, or completely on a remote computer. In the latter case, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN), wide area network (WAN) or a connection to an external computer (for example, via the Internet using Internet providers).

[00135] Aspects of the present technical solution have been described in detail with reference to flowcharts, schematic diagrams and / or diagrams of methods, devices (systems) and computer program products in accordance with embodiments of the present technical solution. It should be borne in mind that each block from the block diagram and / or diagrams, as well as combinations of blocks from the block diagram and / or diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a general purpose computer processor, special purpose computer, or other data processing device to create a procedure, so that instructions executed by a computer processor or other programmable data processing device provide means for implementing the functions / actions specified in block or blocks of a flowchart and / or diagram.

[00136] These computer program instructions can also be stored on a computer-readable medium that can control a computer other than a programmable data processing device or other devices that operate in a specific way, such that the instructions stored on the computer-readable medium create a device including instructions that carry out the functions / actions indicated in the block diagram and / or block.

USED INFORMATION SOURCES

1. Andrews S., FastQC A. A quality control tool for high throughput sequence data. 2010 // Google Scholar. - 2015.

2. Mieczyslaw A. Klopotek, Slawomir T. Wierzchon, Krzysztof Trojanowski Intelligent Information Processing and Web Mining: Proceedings of the International IIS: IIPWM'05 Conference held in Gdansk, Poland, June 13-16, 2005 page 276.

Claims (14)

1. Computer-implemented integrated method for assessing the quality of the results of targeted sequencing, implemented using a processor, including the following steps: • receive data on at least one sample of at least one user; • perform targeted sequencing of the data of at least one sample obtained in the previous step through the sequencing platform; • derive sequencing characteristics from the results of sequencing, wherein: map readings to the reference genome; counting the number of readings covering each reference position; determine the set of mutations; determine the mutation spectrum for the sample from which the data were obtained; • form an integral sequencing characteristic based on the sequencing characteristics obtained in the previous step by using the machine learning algorithm on the processor. 2. The method according to p. 1, characterized in that the sequencing characteristic is the average coverage of the sample and / or the number of amplicons coated on average more than the average coverage of the sample, and / or the percentage of nucleotide pairs with quality values greater than a predetermined, and / or percentage of readings that have been trimmed. 3. The method according to claim 1, characterized in that when forming the integral characteristic, gradient boosting of decision trees is used. 4. The method according to p. 1, characterized in that when forming the integral characteristics of sequencing, a quality measure is used, which is a number from the range from 0 to 1. 5. The method according to p. 1, characterized in that when forming the integral characteristics using random trees or the method of support vectors, or artificial neural networks. 6. The method according to p. 3, characterized in that during the training of decision trees, the training parameters of the model are selected using cross-validation.

Images