Method and device for determining fetal concentration

CN118969092APending Publication Date: 2024-11-15GENEMIND BIOSCIENCES CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411332382.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2022-12-30
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The data characteristics of the third-generation sequencing technology are different from the second-generation sequencing technology. The existing fetal concentration assessment methods are not suitable for the third-generation sequencing platform, especially when the gender is a female fetus, it is impossible to effectively evaluate fetal concentration.

Method used

By obtaining the sequencing sequence of the cfDNA sample to be tested, comparing it with the reference genome, determining the aligned logarithm of the segment interval in which each sequencing sequence falls, calculating the central position of the nucleosome, and aligning and processing the relevant aligned logarithm. Combining the concentration of the first fetus and the logarithm after addition treatment, the fetus concentration is determined.

Benefits of technology

It has achieved accurate assessment of fetal concentration on various sequencing platforms, which is suitable for men and women, which has improved the accuracy and universality of the assessment, and made up for the shortcomings of the third-generation sequencing platform for fetal concentration assessment of female fetal concentration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969092A_ABST
    Figure CN118969092A_ABST
Patent Text Reader

Abstract

The invention discloses a fetal concentration determination method and device, and the method comprises the steps: obtaining a sequencing sequence of a to-be-detected cfDNA sample; comparing the sequencing sequence with a reference genome to obtain a comparison result; and determining the fetal concentration based on the comparison result. The fetal concentration can be determined based on the comparison result of the sequencing sequence and the reference genome, the application of each sequencing platform including a single molecule sequencing platform can be met, and the accuracy of fetal concentration determination is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of bioinformatics, and more specifically to a method and device for determining fetal concentration. Background Art

[0002] In non-invasive prenatal genetic testing (NIPT), the fetal concentration refers to the proportion of fetal deoxyribonucleic acid (DNA) in the peripheral blood free DNA (cfDNA) of pregnant women. Fetal concentration is an important parameter in NIPT.

[0003] For the second-generation sequencing technology, there are currently methods for assessing fetal concentration, for example, using the distribution characteristics of cfDNA length, using methylation characteristics, using SNP (single nucleotide polymorphism) characteristics, using the XY chromosome content of male fetuses, etc. The third-generation sequencing technology is a single-molecule sequencing technology. Its sequencing principle is different from that of the second-generation sequencing technology. The data characteristics generated by sequencing are quite different from those of the second-generation sequencing technology. Therefore, the fetal concentration assessment method used for the second-generation sequencing is not applicable to the third-generation sequencing platform. Summary of the invention

[0004] The sequencing principle based on the third-generation sequencing technology is different from that of the second-generation sequencing technology. The data characteristics generated by the third-generation sequencing technology are also different from those of the sequencing data obtained by the second-generation sequencing technology. These differences make the two also present certain differences in application. For example, the third-generation sequencing platform cannot measure the length of the inserted sequence, so it is impossible to determine the fetal concentration by the distribution characteristics of the cfDNA length; for example, the third-generation sequencing platform cannot measure methylation information, so it is not suitable to determine the fetal concentration by extracting methylation characteristics; for example, the difference in the sequencing principle of the third-generation sequencing platform leads to the error characteristics generated in its sequencing process being different from the error characteristics generated by the second-generation sequencing, resulting in the SNP method applicable to the second-generation sequencing not being applicable to the determination of the fetal concentration of the third-generation sequencing platform. The determination of fetal concentration using XY chromosome content can only be based on the premise that the fetus is a male fetus, and it is also not suitable for the determination of fetal concentration for female fetuses. Therefore, based on the above situation, it is necessary to develop a fetal concentration assessment method that can take into account various sequencing platforms and can be applied to fetuses of different genders.

[0005] In view of this, this application provides the following technical solutions:

[0006] A method for determining fetal concentration, comprising:

[0007] Obtaining the sequencing sequence of the cfDNA sample to be tested;

[0008] Segmenting the reference genome to obtain a plurality of segmented intervals, and determining a first alignment number of each of the sequencing sequences falling into each of the segmented intervals;

[0009] determining a first fetal concentration based on the first comparison log;

[0010] Determine the nucleosome center position based on the second alignment number that each base site of the reference genome is the start position of the sequencing sequence alignment;

[0011] Based on the nucleosome center position, align and sum the second alignment numbers of corresponding positions in all nucleosome regions to obtain the summed second alignment numbers;

[0012] Determining a comparison result based on the first fetal concentration and the second comparison logarithm after the summation process;

[0013] Based on the comparison results, the fetal concentration is determined.

[0014] A device for determining fetal concentration, comprising:

[0015] A second acquisition unit is used to obtain a sequencing sequence of the cfDNA sample to be tested;

[0016] A second determination unit is used to segment the reference genome to obtain multiple segment intervals, and determine a first alignment number of each sequencing sequence falling into each segment interval;

[0017] a third determining unit, configured to determine a first fetal concentration based on the first comparison logarithm;

[0018] A fourth determination unit, configured to determine the nucleosome center position based on a second alignment number at which each base site of the reference genome is the start position of the sequencing sequence alignment;

[0019] A processing unit, configured to align and sum the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position to obtain a second alignment number after the summation;

[0020] a fifth determining unit, configured to determine a comparison result based on the first fetal concentration and the second comparison logarithm after the summation process;

[0021] A sixth determination unit is used to determine the fetal concentration based on the comparison result.

[0022] The present application discloses a method and device for determining fetal concentration, which can determine fetal concentration based on the comparison results of sequencing sequences and reference genomes, and can meet the application of various sequencing platforms including single-molecule sequencing platforms, thereby improving the accuracy of fetal concentration determination. In addition, the method for determining fetal concentration provided by the present application does not rely on the identification and confirmation of gender-dependent features, and therefore can be applied to the determination of fetal concentrations of both male and female fetuses, making up for the deficiency that the third-generation sequencing platform cannot effectively evaluate the fetal concentration of female fetuses. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0024] Figure 1 A schematic flow chart of a method for determining fetal concentration provided in an embodiment of the present application;

[0025] Figure 2 A schematic diagram of the relationship between a fetal concentration calculated by an Enet model and a reference fetal concentration provided in an embodiment of the present application;

[0026] Figure 3 A schematic diagram of the relationship between a fetal concentration calculated by a WRSC model and a reference fetal concentration provided in an embodiment of the present application;

[0027] Figure 4 A schematic diagram of the relationship between the fetal concentration calculated by the gmFF_V7 original parameters and the reference fetal concentration provided in an embodiment of the present application;

[0028] Figure 5 A schematic diagram of the relationship between the fetal concentration calculated after linear correction of gmFF_V7 and the reference fetal concentration provided in an embodiment of the present application;

[0029] Figure 6 A schematic diagram of the distribution of a set of data provided in an embodiment of the present application after alignment and summation according to nucleosome regions;

[0030] Figure 7 A schematic diagram of a model prediction effect after randomly splitting a test set and a training set provided in an embodiment of the present application;

[0031] Figure 8 A schematic flow chart of another method for determining fetal concentration provided in an embodiment of the present application;

[0032] Fig. 9A schematic diagram of the structure of a device for determining fetal concentration provided in an embodiment of the present application;

[0033] Fig.10 A schematic diagram of the structure of another device for determining fetal concentration provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The technical solution of the present application will be clearly and completely described below in conjunction with the embodiments of the present application and the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0035] The terms "first" and "second" in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. For example, without departing from the scope of the embodiments of this application, the expressions "first comparison number" and "second comparison number" in this application can be interchanged.

[0036] The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may include steps or units that are not listed.

[0037] The term "sequencing" can also be called "nucleic acid sequencing" or "gene sequencing", that is, the three can be expressed interchangeably, referring to the determination of base types and arrangement orders in nucleic acid sequences; including synthesis sequencing (sequencing by synthesis, SBS) and / or ligation sequencing (sequencing by ligation, SBL), including DNA sequencing and / or RNA sequencing, including long fragment sequencing and / or short fragment sequencing. The so-called long fragments and short fragments are relative, such as nucleic acid molecules longer than 1Kb, 2Kb, 5Kb or 10Kb can be called long fragments, and those shorter than 1Kb or 800bp can be called short fragments.

[0038] The term "single-molecule sequencing technology" refers to sequencing technology based on a single copy of a base signal acquisition unit, that is, a technology that can directly read the sequence of nucleic acid molecules without using a specific amplification technology to amplify the signal of the nucleic acid molecule to be tested. The term "single-molecule sequencing platform" refers to a sequencing platform based on single-molecule sequencing technology, including various sequencers based on single-molecule sequencing technology, including but not limited to Genocare sequencing platform of Genmab Biotechnology.

[0039] A method for determining fetal concentration is provided in an embodiment of the present application. The method can be applied to various sequencing platforms to evaluate fetal concentration. In particular, through this method, a fetal concentration evaluation method can be established based on a single molecule sequencing platform, especially a fetal concentration evaluation method suitable for female fetuses.

[0040] See also Figure 1 , is a flow chart of a method for determining fetal concentration provided in an embodiment of the present application, the method may include the following steps:

[0041] S101. Obtain the sequencing sequence of the cfDNA sample to be tested.

[0042] Among them, cfDNA refers to highly fragmented DNA that exists in the human blood circulation and is free from cells. In the present application embodiment, the cfDNA sample to be tested refers to the cfDNA sample whose fetal concentration needs to be determined. Exemplarily, the cfDNA sample can be peripheral blood of a pregnant woman containing free deoxyribonucleic acid (or free DNA).

[0043] In the embodiment of the present application, the cfDNA sample to be tested refers to a sample containing cfDNA whose fetal concentration is to be determined, and the sequencing sequence of the cfDNA sample to be tested refers to the sequence obtained by sequencing the cfDNA in the sample to be tested, that is, the sequencing sequence of the fragmented DNA in the sample to be tested. It should be understood that the sequencing sequence of the cfDNA sample to be tested has no requirements for the sequencing instrument and sequencing platform, that is, the sequencing sequence of the cfDNA sample to be tested can be derived from the sequencing data obtained by any sequencing platform. In some embodiments, the sequencing sequence of the cfDNA sample to be tested can be obtained by a single molecule sequencing platform (Genocare sequencing platform of Zhenmai Bio, nanopore sequencing platform, etc.), or by a second sequencing platform (such as the illumina platform of the second generation sequencing, the BGI platform). Exemplarily, the sequencing sequence of the cfDNA sample to be tested is the data obtained by DNA sequencing of the cfDNA sample to be tested based on the single molecule sequencing platform. Thus, the sequencing sequence of the cfDNA sample to be tested derived from the single molecule sequencing platform can be realized by the method provided in the embodiment of the present application, and the analysis of the fetal concentration of the cfDNA sample to be tested is realized.

[0044] S102. Compare the sequencing sequence with the reference genome to obtain an alignment result.

[0045] In this step, the reference genome refers to the human reference genome, illustratively, such as the human reference genome hg19, but not limited thereto.

[0046] S103. Determine the fetal concentration based on the comparison results.

[0047] In order to accurately obtain the final fetal concentration in the embodiment of the present application, the sequencing sequence is compared with the reference genome, and the obtained comparison result can be the comparison probability of the cfDNA of the mother and the infant at different position intervals on the reference genome, can be the comparison result corresponding to the distribution difference of the two ends of the cfDNA of the mother and the infant on the nucleosome, can be the comparison probability of the cfDNA of the mother and the infant at different position intervals on the reference genome and the comparison result corresponding to the distribution difference of the two ends of the cfDNA of the mother and the infant on the nucleosome, and can also be a newly formed comparison result after processing the above two comparison results.

[0048] The method for determining the fetal concentration provided in the embodiments of the present application is described in detail below.

[0049] The fetal concentration is determined by using the difference in the probability of alignment of maternal and fetal cfDNA at different positions on the reference genome.

[0050] Among the free DNA (cfDNA) obtained from the maternal peripheral blood, part comes from the mother and part comes from the fetus. When the DNA from the mother and fetus are compared with the reference genome, there will be certain differences in the coverage depth in different regions of the reference genome. Therefore, the fetal concentration can be determined based on the comparison results corresponding to the differences.

[0051] In one embodiment of the present application, the fetal concentration is determined by using the difference in the comparison probability of maternal and infant cfDNA at different position intervals on the reference genome. Specifically, the method for determining the fetal concentration includes:

[0052] S111. Obtain the sequencing sequence of the cfDNA sample to be tested.

[0053] In this step, the cfDNA sample to be tested is the cfDNA sample whose fetal concentration needs to be determined. The method provided in the implementation of this application does not have clear requirements on the source of the cfDNA sample to be tested and the characteristic parameters of the cfDNA sample to be tested.

[0054] S112. Compare the sequencing sequence with the reference genome to obtain an alignment result.

[0055] This step includes:

[0056] S1121. Segment the reference genome to obtain multiple segmentation intervals.

[0057] S1122. Determine a first alignment number for each sequencing sequence falling into each of the segmented intervals, so that the first alignment number is determined as the alignment result.

[0058] It should be noted that the "first" in the first alignment number is only used to distinguish it from the alignment numbers that appear later, and there is no order relationship between the "first" and the "second". This method segments the reference genome into segments of fixed lengths, and then counts the number of sequencing sequences that fall into each segmented interval, that is, calculates how many sequencing sequences fall into the corresponding segmented intervals, and the number of sequencing sequences that fall into each segmented interval is the first alignment number, which is also the alignment result in this scenario. Correspondingly, in some scenarios, the number of sequencing sequences that fall into each segmented interval can also be referred to as the "coverage number", which is referred to as the alignment number in the subsequent embodiments of this application for the convenience of description, and is distinguished from other types of alignment results, referred to as the "first alignment number".

[0059] It should be understood that the "first comparison number" described in the embodiment of the present application can be the absolute number of sequencing sequences falling into the corresponding segmented interval, or the median of sequencing sequences falling into the corresponding segmented interval. In addition, in the statistical process, the number of sequencing sequences falling into each segmented interval is counted, and a certain error tolerance ratio is allowed, that is, sequencing sequences with a preset number of base errors within a certain length range are allowed as sequencing sequences for comparison.

[0060] In order to obtain accurate comparison results, the reference genome can be segmented according to fixed lengths, such as determining the segmentation length of the reference genome based on the characteristics of the sequencing sequence.

[0061] There are some special chromosomes in the reference genome, and these chromosomes may have certain biases or other particularities in the process of determining the fetal concentration by comparing the reference genome. In one implementation scenario, the bias brought by the special chromosomes (hereinafter referred to as specific chromosomes) in the reference genome is reduced by removing the segmented intervals brought by these chromosomes, thereby reducing the impact of their particularity on the test results. In some embodiments, the specific chromosomes include at least one of the X chromosome, the Y chromosome, the mitochondrial chromosome, the chromosome 13, the chromosome 18 and the chromosome 21. Among them, by removing the segmented intervals corresponding to the X chromosome and the Y chromosome, the difference between the above chromosomes of male and female fetuses is avoided, and the impact on the results of the subsequent concentration quantitative model is avoided; chromosomes 13, chromosomes 18 and chromosomes 21 are removed because these chromosomes themselves have a large probability of duplication and deletion. One of the screening goals of NIPT is to screen for duplications and deletions of chromosomes 13, 18, and 21. When chromosomes 13, 18, and 21 are used as reference chromosomes, deviations may occur when normalizing the data, such as affecting the normalized logarithm, which in turn affects the calculation results and the accuracy of the NIPT screening results.

[0062] The embodiment of the present application can adjust the range of specific chromosomes according to the expected needs of fetal concentration detection, such as eliminating X chromosomes and Y chromosomes, or eliminating chromosomes 13, 18 and 21, or eliminating mitochondrial chromosomes. Of course, multiple situations in the above situations can be eliminated at the same time, such as removing all X chromosomes, Y chromosomes, mitochondrial chromosomes, chromosomes 13, chromosomes 18 and chromosomes 21 or their corresponding segmented fragments, so that the segmented intervals obtained after the reference genome is segmented do not contain the segmented intervals corresponding to the above chromosomes.

[0063] The present application can remove specific chromosomes into segmented intervals in a variety of ways. The segmented intervals obtained after segmentation can be screened based on specific chromosomes, or these specific chromosomes can be removed before segmentation. Specifically, the corresponding processing method can be selected based on actual application requirements, and the present application does not limit this.

[0064] In one embodiment, the reference genome is segmented to obtain multiple segmentation intervals, including: segmenting the reference genome based on a preset segmentation length to obtain multiple initial segmentation intervals; removing the segmentation interval corresponding to a specific chromosome in the initial segmentation interval to obtain multiple segmentation intervals. That is, after the reference genome is segmented into multiple segmentation intervals, the segmentation interval corresponding to the specific chromosome in the reference genome is removed and does not participate in the calculation. It should be understood that the "segmentation interval corresponding to a specific chromosome" refers to the multiple segmentation intervals formed after the specific chromosome is segmented according to the preset segmentation length.

[0065] In another embodiment, segmenting the reference genome to obtain multiple segment intervals includes: removing a specific chromosome from the reference genome to obtain a reference genome after the removal; and segmenting the reference genome after the removal based on a preset segmentation length to obtain multiple segment intervals. That is, before segmenting the reference genome, the specific chromosome is first removed.

[0066] The preset segmentation length can be adjusted according to the amount of data to be processed and the expected relative accuracy, for example, the reference genome can be segmented according to the length of each 50k (5000) bases. In order to make the obtained first alignment number more accurate, the first alignment number can also be corrected to obtain a corrected first alignment number, so that the corrected first alignment number is determined as the alignment result.

[0067] In some embodiments, the correction includes GC correction. GC correction is an operation in the bioinformatics analysis process. The reason for GC correction is that the gene sequencer and the corresponding sequencing process will cause a certain GC preference. For example, some sequencers have a higher probability of measuring high GC sequences, while some sequencers have a higher probability of measuring low GC sequences. In this embodiment, the processing method for determining the fetal concentration is based on the comparison probability of different regions on the reference genome, and the GC preference of the sequencer will affect the judgment of the actual comparison probability of different regions. In order to improve the accuracy, GC correction is required.

[0068] At this time, the first alignment number is corrected to obtain the corrected first alignment number, including: based on the first alignment number of each segmented interval and the average first alignment number of all segmented intervals, the first alignment number of each segmented interval is normalized to obtain the normalized first alignment number of each segmented interval; the normalized first alignment number of each segmented interval is GC corrected to obtain the GC corrected first alignment number. The average first alignment number of all segmented intervals refers to the ratio of the sum of the first alignment numbers of each segmented interval to the total number of segmented intervals.

[0069] After obtaining the first comparison numbers of all segmented intervals, in order to facilitate calculation and processing, the first comparison numbers of all segmented intervals will be normalized. The corresponding formula is:

[0070] Normalized first alignment number of each segment interval=first alignment number of each segment interval / average first alignment number of all segment intervals.

[0071] After obtaining the first alignment number of each segmented interval, GC correction will be performed to obtain the first alignment number after GC correction.

[0072] In one embodiment, performing GC correction on the normalized first alignment number of each segment interval to obtain the GC-corrected first alignment number includes:

[0073] (1) Generate a first relationship curve based on the GC content and the normalized first alignment number corresponding to each segment interval.

[0074] In this embodiment, the GC content can be represented by a variety of parameters, such as the GC content or the median GC content of each segmented interval, or other parameters that can provide feedback on the GC content of the segmented interval.

[0075] In an embodiment of the present application, GC filtering can be performed on the segmented intervals forming the first relationship curve, such as filtering the segmented intervals according to GC content to filter out some abnormal or statistically insignificant segments, thereby optimizing the first relationship curve and reducing the occupancy of data processing resources during data processing.

[0076] In one embodiment, GC filtering is achieved in the following manner: before the step of generating a first relationship curve based on the GC content and the normalized first logarithm corresponding to each segmented interval, the following processing is performed: the segmented intervals are filtered based on the GC content, and the segmented intervals whose GC content does not meet the preset requirements are removed, so that the first relationship curve is generated based on the GC content of each segmented interval after GC filtering and the normalized first logarithm of each segmented interval.

[0077] In another embodiment, GC filtering is achieved in the following manner: in the process of performing GC correction on the normalized first logarithm based on the first relationship curve, the segmented intervals are filtered based on the GC content, and the normalized first logarithm after GC filtration is GC corrected based on the first relationship curve to obtain the GC-corrected first logarithm.

[0078] In the above embodiment, as an example, based on the GC content and the normalized first alignment number corresponding to each segmented interval, a first relationship curve is generated, which is achieved in the following manner: the x-axis represents the GC content of each segmented interval (i.e., the proportion of GC bases corresponding to the segmented interval on the reference genome), and the y-axis represents the normalized first alignment number corresponding to each segmented interval. Integrate the data of each segmented interval and the GC content within the interval, that is, generate a scatter plot based on the GC content of each segmented interval and the corresponding normalized first alignment number. Smooth the scattered data in the scatter plot to obtain a smooth curve. Among them, smoothing the scattered data in the scatter plot can be performed using a LOWESS algorithm, for example, to obtain a smooth curve, that is, to obtain the first relationship curve, such as y. s =f(x s ).

[0079] Specifically, as an example, the first relationship curve can be generated in the following manner: divide the horizontal axis into different segments, then calculate the median GC content of the scattered points in the same segment, use this median GC content to represent the value corresponding to this segment, and then draw all these GC content medians on the graph for smoothing to obtain the first relationship curve. In another example, smoothing can be performed directly based on all the data, that is, all data points are taken into consideration and smoothed together to obtain a smooth curve, so that the Spearman correlation coefficient obtained for the test set is higher. It should be noted that this smoothing method is also applicable to the smoothing and correction of the comparison probability described later.

[0080] (2) Performing GC correction on the normalized first alignment number based on the first curve to obtain the GC-corrected first alignment number.

[0081] In some embodiments, GC correction is performed based on a subtraction relationship or a division relationship between the normalized first alignment number and the GC content of the segmented interval corresponding to the normalized first alignment number to determine the GC-corrected first alignment number.

[0082] In one example, the GC correction method is "subtraction", and the corresponding correction formula is: GC =yf(x); In another example, the GC correction method is "division", and the corresponding correction formula is: GC =y / f(x).

[0083] In the above two correction formulas, x is the GC content of the segmented interval, y is the normalized first logarithm corresponding to the segmented interval, and y GC is the normalized first alignment number after GC correction.

[0084] In some embodiments, the correction further includes alignment probability correction to improve the accuracy of the obtained first alignment number. That is, in some embodiments, the correction includes both GC correction and alignment probability correction.

[0085] In one implementation of the embodiments of the present application, the first alignment number is corrected to obtain the corrected first alignment number, and the method further includes: truncating the reference genome based on a specific sliding window length, aligning the truncated sequence with the reference genome, and counting the first alignment number in each segmented interval of the reference genome, and determining the first alignment number as the sliding window alignment number; based on the sliding window alignment number, determining the normalized sliding window alignment number of each segmented interval; performing alignment probability correction based on the normalized sliding window alignment number of each segmented interval and the normalized first alignment number, and determining the first alignment number after the alignment probability correction.

[0086] The embodiment of the present application adopts the "matching probability correction" method to correct the first comparison number, based on the following considerations: if different regions of the reference sequence corresponding to the reference genome are cut off according to a specific length and used as a sequencing sequence, and then compared back to the reference genome, the number of times different segmented intervals are compared will be different, because for the sequences in some regions, there are similar sequences in many places on the reference sequence of the reference genome, so these regions are more easily compared. For example, the human genome reference sequence hg19 is subjected to a sliding window to cut the sequence according to a step length of 2 bases and a window of 37 bases, and these sequences are compared with the human reference genome hg19, and then the number of comparisons on different segmented regions of the human reference genome is counted, and this comparison number can be called the "sliding window comparison number". Specifically, the specific sliding window length can be considered based on the actual application scenario, such as based on the characteristics of the reference genome adopted, or the characteristics of the sample to be tested. After determining the specific sliding window length, the reference genome is intercepted, the intercepted sequence is compared with the reference genome, and the first alignment number on each segmented interval on the reference genome is counted, and the first alignment number at this time is determined as the sliding window alignment number. After obtaining the sliding window alignment number, in order to facilitate calculation and processing, the normalized sliding window alignment number of each segmented interval is determined based on the sliding window alignment number; the normalized sliding window alignment number is corrected for alignment probability. The method for alignment probability correction can refer to the GC correction method mentioned above.

[0087] In one embodiment, performing alignment probability correction based on the normalized sliding window alignment number and the normalized first alignment number of each segmented interval, and determining the first alignment number after alignment probability correction, comprises:

[0088] (1) Based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval, a second relationship curve is generated.

[0089] Among them, the normalized sliding window ratio logarithm of each segmented interval can be represented by multiple parameters, such as the normalized sliding window ratio logarithm of each segmented interval or the median of the sliding window ratio logarithm, or other parameters that can feedback the normalized sliding window ratio logarithm of each segmented interval.

[0090] In an embodiment of the present application, the normalized sliding window ratio logarithm forming the second relationship curve can be filtered to filter out some abnormal or statistically insignificant segments, thereby optimizing the second relationship curve and improving the processing accuracy when performing related data processing based on the second relationship curve.

[0091] In one embodiment, the normalized sliding window logarithm that forms the second relationship curve is filtered, and this is achieved in the following manner: before the step of generating the second relationship curve based on the normalized sliding window logarithm of each segmented interval and the normalized first logarithm corresponding to each segmented interval, the following processing is performed: the segmented intervals are filtered based on the normalized sliding window logarithm to obtain segmented intervals whose normalized sliding window logarithm is not less than the first target threshold, so that the second relationship curve is generated in the segmented intervals whose normalized sliding window logarithm is not less than the first target threshold.

[0092] In one embodiment, the normalized sliding window logarithm that forms the second relationship curve is filtered in the following manner: based on the second relationship curve, the normalized first logarithm is corrected for comparison probability, and in the process of obtaining the first logarithm after the comparison probability correction, the segmented interval is filtered based on the normalized sliding window logarithm, and the second relationship curve interval in which the normalized sliding window logarithm in the second relationship curve is not less than the first target threshold is retained to obtain the second relationship curve after comparison probability filtering; based on the second relationship curve after comparison probability filtering, the sliding window normalized logarithm corresponding to each segmented interval is corrected for comparison probability to obtain the first logarithm after comparison probability correction.

[0093] In the above embodiment, based on repeated verification, the first target threshold may be selected as 0.8, that is, the segmented intervals whose normalized sliding window ratio logarithm is less than 0.8 are removed.

[0094] In the above embodiment, as an example, based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval, a second relationship curve is generated, which is achieved in the following manner: the x-axis represents the normalized sliding window comparison number of each segmented interval, the y-axis represents the normalized first comparison number corresponding to each segmented interval, and the data of each segmented interval and its corresponding GC content are integrated, that is, according to the GC content of each segmented interval and the corresponding normalized first comparison number, a scatter plot is generated; the scattered data in the scatter plot is smoothed to obtain a smooth curve. Exemplarily, the smoothing of the scattered data in the scatter plot can be achieved using the LOWESS algorithm.

[0095] (2) Performing alignment probability correction on the normalized first alignment number based on the second relationship curve to obtain the first alignment number after alignment probability correction.

[0096] In some embodiments, the alignment probability is corrected according to a subtraction relationship or a division relationship between the normalized first alignment number and the sliding window alignment number of the segmented interval corresponding to the normalized first alignment number to determine the first alignment number after the alignment probability correction.

[0097] In one example, the method of comparison probability correction is "subtraction", and the corresponding correction formula is:比对概率 =yf(x); In another example, the method of comparison probability correction is "division", and the corresponding correction formula is: y 比对概率 =y / f(x).

[0098] In the above two correction formulas, x is the normalized sliding window comparison number of each segment interval, and y corresponds to the normalized first comparison number of each segment interval. 比对概率 is the normalized first pairwise comparison after correction for the comparison probability.

[0099] S113. Determine the fetal concentration based on the comparison results.

[0100] In the embodiment of the present application, the fetal concentration can be determined in combination with a machine learning method of a neural network. In one embodiment of the present application, the fetal concentration is determined based on the comparison results, including:

[0101] Obtaining first training sample data, wherein each sample in the first training sample data is annotated with a first characteristic value and a first target value, the first characteristic value is the sample logarithm, and the first target value is the actual fetal concentration of the sample;

[0102] Performing machine learning modeling on the first training sample data based on the specific model result to obtain a first fetal concentration quantitative model;

[0103] The comparison result is input into the first fetal concentration quantitative model to obtain the first fetal concentration, and the first fetal concentration is determined as the fetal concentration.

[0104] In some embodiments, performing machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model includes:

[0105] Dividing the first training sample data into a training set and a test set; performing machine learning modeling based on the training set to obtain an initial model;

[0106] The test set is processed based on the initial model to obtain an estimated fetal concentration for each test sample in the test set;

[0107] Obtaining a comparison result based on comparing the estimated fetal concentration with the actual fetal concentration of each test sample in the test set;

[0108] The model parameters of the initial model are adjusted based on the comparison results to obtain the first fetal concentration quantitative model.

[0109] Among them, the first training sample data can be data in a database of cfDNA in the peripheral blood of pregnant women with known fetal concentrations obtained by various sequencing platforms including single-molecule sequencing platforms. For example, a certain number of samples with known fetal concentrations are used to segment the reference genome chromosome reference sequence (for example, it can be segmented by 5k bases, 50k bases, 100k bases, 300k bases or 800k bases, etc.), and the number of alignments of the comparison data of each sample in different segmentation intervals is counted respectively, and the alignment numbers are normalized, GC and alignment probability are corrected using the LOWESS method, and the final normalized and corrected alignment numbers in different segmentation intervals are used as feature values, and the known actual fetal concentration corresponding to each sample is used as the target value, and a specific model is used for learning and modeling, wherein the specific model can be a classification model, a convolutional neural network model, etc. The embodiment of the present application does not limit the specific model structure, as long as the training sample data can be learned and the fetal concentration can be predicted. The process of model learning and training is an iterative process, so the training samples can be divided into training sets and test sets. The accuracy of the model currently trained is tested by the test set. If the accuracy is lower than the set accuracy threshold, the model parameters in the model structure of the current model are adjusted by the test set until the error between the fetal concentration output by the model and the actual fetal concentration meets the corresponding error range.

[0110] When the number of samples used for machine learning is small, but the number of eigenvalues ​​is large (e.g., about 30,000 to 600,000 features, depending on the size of the chromosome reference sequence interval segment), the effect of overfitting cannot be avoided in both dimensionality reduction and learning. In the final fetal concentration quantitative model established, the predicted effect of the test set fetal concentration will deviate from the actual fetal concentration. For example, in the fetal concentration quantitative model established using 615 samples with known fetal concentrations, the model corresponds to the test set R 2 Only about 0.3, the Pearson correlation coefficient between true value and predicted value is about 0.5.But, when adopting the aforesaid method provided by the present application embodiment to normalize the comparison logarithm, and correct (GC correction, or GC and comparison probability correction), the correlation coefficient between the predicted value of test set and the true value improves.Compare respectively directly with the normalized segmentation interval comparison logarithm modeling, and use normalization and according to the comparison logarithm of the segmentation interval after GC and comparison probability correction to model, can obtain the modeling effect after normalization, GC correction and comparison probability correction better, that is, the correlation coefficient between the predicted value of test set and the true value is higher.

[0111] In another embodiment, determining the fetal concentration based on the comparison results comprises:

[0112] Inputting the comparison result into the first preset model to obtain an initial fetal concentration;

[0113] The initial fetal concentration is corrected according to the second preset model to obtain the fetal concentration.

[0114] Among them, the first preset model is a linear relationship pattern established based on the comparison results in the second sample data and the initial fetal concentration corresponding to the comparison results. The second sample data is the comparison results obtained by comparing the sequencing sequences in the cfDNA sample with the reference genome, and the fetal concentration corresponding to the comparison results; the second preset model is a model obtained by processing the first preset model based on the constants determined by phenomenon fitting.

[0115] Exemplarily, the first preset model can be in, is the comparison result, FF1 is the initial fetal concentration. The second preset model may be FF2=FF1·c+d, c and d are constants determined by linear fitting, and FF2 is the fetal concentration.

[0116] Exemplarily, the model that uses the difference in the probability of alignment of samples in different intervals of the reference genome reference sequence to establish a model and estimate the fetal concentration is SeqFF. Two machine learning models are used in the SeqFF method, one is the weighted ranked selection criterion (WRSC) and the other is the elastic network (Enet). The final output of the fetal concentration result is the average of the two model predictions. When the SeqFF model is tested using data from sequencing platforms, especially single-molecule sequencing platforms, the WRSC model performs better than Enet, that is, the linear correlation coefficient between the fetal concentration estimated by WRSC and the actual fetal concentration is higher than that of Enet or SeqFF, so the model parameters involved in the WRSC algorithm are given priority. For details, please refer to Figure 2 and Figure 3 , Figure 2 A schematic diagram of the relationship between the fetal concentration calculated by the Enet model and the reference fetal concentration provided in an embodiment of the present application, Figure 3 A schematic diagram of the relationship between the fetal concentration calculated by the WRSC model and the reference fetal concentration provided in an embodiment of the present application. As can be seen from the figure, the positive correlation between the fetal concentration calculated by the WRSC model and the reference fetal concentration is better than the positive correlation between the fetal concentration calculated by the Enet model and the reference fetal concentration.

[0117] For example, in the WRSC model, the chromosome reference sequence is divided according to a specific number of bases, such as 50k base intervals. Through WRSC algorithm training, the linear relationship between the coverage density and fetal concentration in different reference sequence intervals is found, as shown in the following formula. In the following formula, FF is the fetal concentration, is the coverage density of the reference sequence at different intervals, For is a constant vector of equal dimension, B is the parameter matrix used for linear transformation, and c and d are constants.

[0118]

[0119] Since SeqFF is a model based on the training of second-generation sequencing data, its sequencing principle is quite different from single-molecule sequencing. Therefore, if the above-mentioned model parameters of SeqFF are directly used to calculate the fetal concentration of data obtained by other sequencing platforms including single-molecule sequencing platforms, the calculated fetal concentration is quite different from the actual fetal concentration. The 615 NIPT clinical samples with known fetal concentrations in the above embodiment are still used as an example for analysis and testing. The original comparison file data (such as sam files) of other sequencing platforms including single-molecule sequencing platforms are directly processed using the algorithm and parameters of SeqFF, and the Pearson correlation coefficient between the obtained intermediate parameters and the actual fetal concentration is 0.714.

[0120] The calculated fetal concentration value calculated by the SeqFF model has a good linear relationship with the actual fetal concentration value, so the above formula can be used The model parameters contained in this part calculate the intermediate parameters of fetal concentration, and then use linear regression to fit the values ​​of c and d by themselves, so as to obtain the final model parameters suitable for other sequencing platforms including the current single-molecule sequencing platform.

[0121] The 615 NIPT clinical samples with known fetal concentrations in the above embodiment are still used as an example for analysis and testing. Instead, only the first alignment number of the sequencing sequence falling into each segmented interval is used for analysis; the regions with abnormal alignment probabilities in X, Y, M, 13, 18, 21 in the reference genome and on mitochondrial chromosomes are excluded for analysis; the correction method (GC correction, or GC and alignment probability correction) is optimized, etc. The data was re-analyzed using the optimized parameters, and the Pearson correlation coefficient between the intermediate parameters and the fetal concentration was 0.755. The test set (30%) and training set (70%) were randomly split 100 times to establish a linear model and determine the above-mentioned constant parameters c and d. The average test set prediction R obtained from 100 modelings 2 The value is 0.561.

[0122] It should be noted that when the linear model is actually established, the split of the test set and the training set is to evaluate the credibility of the model established by the training set that we currently randomly split in limited data. Therefore, the test set (30%) and the training set (70%) were randomly split 100 times, and the average test set prediction R2 value obtained from 100 modelings was 0.561.

[0123] Table 1 below shows 9 examples of determining fetal concentration based on the difference in the probability of alignment of maternal and fetal cfDNA at different position intervals on the reference genome.

[0124] In Table 1, Example 1 is the condition used by the original seqFFWRSC method. The "Bin" in "Bin screening" in the table refers to the segmented interval on the genome reference sequence, "Mappability correction" refers to "correction of alignment probability", "Mappability" refers to the normalized alignment number of the above sliding window method, and "groupby" refers to a GC correction method in the seqFF method. Groupby is equivalent to segmenting the GC content into intervals when doing lowess smoothing. Each segment is represented by a GC content value with three decimal places, and the median of the normalized statistics corresponding to all segmented intervals within this GC content range is used to refer to the normalized statistics under the GC content. As mentioned in the table, we mark this processing method as "median". Correspondingly, there is also a normalization method marked as "all" in the table. "All" means that the GC content of all segmented intervals is taken into account. The amount and its corresponding normalized alignment number are taken into account for Lowess smoothing, and the GC content is not grouped into intervals, that is, the data of all segmented intervals are used for smoothing. "Median" refers to the median, "Frac" refers to the parameter used for Lowess smoothing, "all mapped" refers to all aligned sequences, "unique" refers to the only aligned sequence (that is, the sequence that is only aligned to one chromosome), "XYM" or "XYM chromosome" refers to the X chromosome, Y chromosome and mitochondrial chromosome, and "flag filtering" refers to the original paper of the seqFF method, which did some filtering for different segmented intervals and provided filtering files. "SeqFF binflag filtering" is to filter according to the tags marked in the original document. The solution we finally selected did not filter according to the tags in this original document.

[0125] Table 1

[0126]

[0127]

[0128] Peel's correlation coefficient assesses the degree of linear correlation between the values ​​calculated directly by the algorithm and the reference fetal concentration. Figure 4 Although the two have a high degree of linear correlation, their absolute values ​​are very different. Therefore, a linear model is needed to correct the calculation results of the algorithm so that it can accurately estimate the fetal concentration. The formula is used for fitting correction. In this formula, This whole block is the numerical result calculated by gmFF_V7. c and d are two parameters fitted in the linear simulation, which are used as constants in the subsequent model.

[0129] Table 1 evaluates the advantages and disadvantages of different treatment parameters and conditions through the Pearson correlation coefficient between the predicted fetal concentration and the reference fetal concentration (fetal concentration provided by other platforms).

[0130] As can be seen from Table 1, compared with the algorithm of Example 1 (Pearson correlation coefficient is 0.71425), the Pearson correlation coefficients of the other algorithms are all greater than 0.74, and the Pearson correlation coefficients of the algorithms of Examples 6, 8 and 9 are greater than 0.75. Compared with other algorithms, the algorithm of Example 8 can obtain the highest Pearson correlation coefficient, which is 0.75471.

[0131] The relationship between the fetal concentration calculated in Example 8 and the reference fetal concentration is shown in the figure below: Figure 5 As shown. Figure 5 It can be seen that there is a high positive correlation between the calculated fetal concentration and the reference fetal concentration.

[0132] The fetal concentration can be calculated using the difference in the distribution of the two ends of the maternal and fetal cfDNA on the nucleosome.

[0133] Nucleosomes are the basic units that make up human chromatin. When DNA is cut by enzymes, it is more difficult to cut inside the nucleosome than outside the connection between nucleosomes. Therefore, DNA is more likely to be cut at the connection between nucleosomes during digestion. Due to the differences in the structural characteristics of nucleosomes in the DNA of mothers and fetuses during apoptosis, the binding strength of nucleosomes also differs. Therefore, the probability of DNA of mothers and fetuses being cut inside nucleosomes and at the connection between nucleosomes during digestion will also be different. Specifically, the probability of fetal DNA being cut inside nucleosomes is higher than that of maternal DNA being cut inside nucleosomes. Therefore, by evaluating the positional distribution of the endpoints of cfDNA in the cfDNA sample to be tested relative to the nucleosomes, information on fetal concentration can be obtained.

[0134] In another embodiment of the present application, a method for determining fetal concentration using the difference in the distribution of the two ends of maternal and infant cfDNA on nucleosomes is provided. Specifically, the method for determining fetal concentration includes:

[0135] S121. Obtain the sequencing sequence of the cfDNA sample to be tested.

[0136] In this step, the cfDNA sample to be tested is the cfDNA sample whose fetal concentration needs to be determined. The method provided in the implementation of this application does not have clear requirements on the source of the cfDNA sample to be tested and the characteristic parameters of the cfDNA sample to be tested.

[0137] S122. Compare the sequencing sequence with the reference genome to obtain an alignment result.

[0138] This step includes:

[0139] S1221. Determine the count of the base sites at which each base site of the reference genome is the starting position of the sequencing sequence alignment as a second alignment number, and calculate the nucleosome center score corresponding to each base site of the reference genome based on the second alignment number;

[0140] S1222. Determine the nucleosome center position based on the nucleosome center score and the center score screening threshold;

[0141] S1223. Align and sum the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position to obtain the summed second alignment number;

[0142] S1224. Perform dimensionality reduction processing on the second comparison number after the addition processing to obtain a normalized second comparison number after dimensionality reduction, and determine the normalized second comparison number after dimensionality reduction as the comparison result.

[0143] It should be noted that the second alignment number in this embodiment is to distinguish it from the first alignment number in the above embodiment, and there is no order between the two. The count of the base sites of each base site of the reference genome that are the starting position of the sequencing sequence alignment is determined as the second alignment number, and correspondingly, the second alignment number can also be called "count", that is, the alignment starting positions of different sequences are counted, and then the number of times different sites on the reference genome are exactly the starting positions of the sequence alignment is obtained.

[0144] Since the probability of the internal region of the nucleosome being the cut of cfDNA (that is, the starting end of the sequencing sequence) is lower than the probability of the nucleosome junction region being the cut of cfDNA, it is necessary to find and confirm the regional range of the nucleosome.

[0145] In step S1221, the counting of the base sites of the comparison starting position of the sequencing sequence in the reference genome is referred to as the comparison starting position of all sequencing sequences, and then the counting of each site on the reference genome as the comparison starting position of these sequences is counted. The present application embodiment determines the count as the second comparison number. It should be understood that the second comparison number can be the absolute number of times that each site on the reference genome is used as the comparison starting position of these sequences, or it can be the median number of times that each site on the reference genome is used as the comparison starting position of these sequences. In addition, in the statistical process, each site on the reference genome is counted as the comparison starting position of these sequences, allowing a certain fault tolerance ratio to occur, that is, allowing a sequencing sequence with a preset number of base errors within a certain length range, as the sequencing sequence on the comparison.

[0146] After determining the count as the second alignment number, the nucleosome center score corresponding to each base site of the reference genome is calculated based on the second alignment number. The "nucleosome center score" can be understood as the probability score of each site being the nucleosome center position. The "nucleosome center position" refers to the center of the nucleosome region of this site.

[0147] In some embodiments, calculating the nucleosome center score corresponding to each base position of the reference genome includes:

[0148] (1) calculating a first average count of a first specific number of base ranges corresponding to each base site on the left and right of the reference genome;

[0149] (2) calculating a second average count of a second specific number of bases corresponding to the left and right sides of each base site of the reference genome;

[0150] (3) Determine the nucleosome center score corresponding to each site based on the first average count and the second average count.

[0151] The first specific number and the second specific number are determined based on the number of bases in the nucleosome region and the number of bases in the nucleosome junction region. With x representing a base site, the first specific number = (the number of bases in the nucleosome region - 1) / 2, and the second specific number = (the number of bases in the nucleosome region - 1) / 2 + the number of bases in the nucleosome junction region / 2.

[0152] For example, according to literature reports, the length of the nucleosome region is 147 bases, and the connecting regions at both ends of the nucleosome are about 20 bases each. In this case, the first specific number is (147-1) / 2=73, and the second specific number is (147-1) / 2+20=93. That is, assuming that x is the center position of the nucleosome, the range of 73 bases to the left and right of this position is the nucleosome region, and the range of the 74th to 93rd bases corresponding to the left and right of this position is the connecting region of the nucleosome.

[0153] The above formula calculates the ratio of the sum of the average counts of the 74-93 base ranges corresponding to the position to be investigated to the average counts of the 73 base ranges to the left and right of the position to be investigated. Obviously, for a specific nucleosome region, the closer the position to be investigated is to the center of the nucleosome, the larger the value.

[0154] Further, the nucleosome center score corresponding to each site is determined according to the first average count and the second average count. The calculation formula of the nucleosome center score can be expressed as:

[0155]

[0156] In the formula, x is the base site, [x-93,x-74] represents the interval from 93 nucleotides on one side of x to 74 nucleotides on the same side, [x+93,x+74] represents the interval from 93 nucleotides on the other side of x to 74 nucleotides on the same side, and [x-73,x+73] represents the interval from 73 nucleotides on the other side of x to 73 nucleotides on the other side of x.

[0157] In some embodiments, considering that the sequencing error rate of a single-molecule sequencer tends to be higher than that of the second-generation sequencing technology, when performing probability analysis and comparison of each base in the nucleosome region as the center position of the nucleosome, the accuracy and efficiency of determining the center position of the nucleosome can be improved by truncating regions with relatively high error rates at both ends of the sequence.

[0158] In some embodiments, the "nucleosome center score" and the subsequent "nucleosome ratio" are calculated after shortening the left and right ends of the nucleosome region by n bases. n is a natural number less than or equal to 5, and exemplary, n is 1, 2, 3, 4 or 5.

[0159] At this time, the calculation formula of the nucleosome center score can be expressed as:

[0160]

[0161] In the formula, x is the base site, [x-93,x-74-n] represents the interval from 93 nucleotides on one side of x to 74-n nucleotides on the same side, [x+93,x+74-n] represents the interval from 93 nucleotides on the other side of x to 74-n nucleotides on the same side, and [x-73-n,x+73-n] represents the interval from 73-n nucleotides on the other side of x to 73-n nucleotides on the other side of x.

[0162] For example, when n=5, the calculation formula of the nucleosome center score can be expressed as:

[0163]

[0164] In the formula, x is the base site, [x-93,x-69] represents the interval from 93 nucleotides on one side of x to 69 nucleotides on the same side, [x+93,x+69] represents the interval from 93 nucleotides on the other side of x to 69 nucleotides on the same side, and [x-68,x+68] represents the interval from 68 nucleotides on the other side of x to 68 nucleotides on the other side of x.

[0165] In step S1223, based on the nucleosome center score and the center score screening threshold, the nucleosome center score is determined, including:

[0166] (1) based on the maximum value of the nucleosome center score, determining the position of the maximum value on the reference genome, and determining the position as the first position;

[0167] (2) clearing the nucleosome center scores of the bases of the specific data on both sides of the first position, and determining the maximum value again based on the remaining nucleosome center scores after clearing, and performing determination of the position of the maximum value on the reference genome until the nucleosome center score is less than a second target threshold, and determining the screened position as a candidate nucleosome center position;

[0168] (3) Based on the base center score screening threshold and the nucleosome center score corresponding to the candidate nucleosome center position, the nucleosome center position is determined.

[0169] Among them, the first position is based on the position of the maximum value of the nucleosome center score on the reference genome, that is, find the maximum value of the nucleosome center score calculated above, record their positions on the reference genome, and clear the nucleosome center score of the bases of the specific data on both sides of the position (such as 147 bases on both sides), because the position found is considered to be the nucleosome center position, and there can be no other nucleosome center position within at least 147 bases on both sides of the nucleosome center position. Then find the position corresponding to the maximum value of the current nucleosome center score again, and continue to execute the step of clearing the score of the next year in the nucleosomes on both sides. This cycle stops until the maximum nucleosome center score value is lower than the second target threshold, and the position screened out based on the execution process is the nucleosome center position. In theory, the average technology of the nucleosome connection area should be higher than the nucleosome area. Based on the above formula, the nucleosome center score should be greater than 2. In order to facilitate processing and improve accuracy, in one embodiment, the value of the second target threshold can be set to 2.2. That is, when screening the nucleosome center region, the cycle continues until the value of the maximum nucleosome center score is lower than 2.2 and then the iterative processing is stopped.

[0170] After all candidate nucleosome center positions are obtained, some positions with abnormally high "nucleosome center scores" among the candidate nucleosome center positions may be affected by various position factors such as alignment probability preference, and their corresponding second alignment numbers are unreliable. Therefore, some of these candidate nucleosome center positions are used as the final selected nucleosome center positions to screen the nucleosome center scores.

[0171] Table 2 below provides several Score screening conditions and the Pearson correlation coefficients of fetal concentration and nucleosome ratio under the corresponding conditions.

[0172] Table 2

[0173]

[0174] In Table 2, "score" refers to "nucleosome center score", "bp" refers to the number of bases, "linker" refers to the nucleosome linker region, and "nucleosome ratio" refers to "nucleosome ratio". The Pearson correlation coefficient between the "nucleosome ratio" and the reference fetal concentration was used to preliminarily screen the nucleosome center position and the calculation method of the nucleosome ratio. [10,) indicates a range greater than or equal to 10, and the right parentheses are open intervals.

[0175] As can be seen from Table 2, compared with not processing the nucleosome region, shortening the left and right ends of the nucleosome region by 5 bases each and then calculating the "nucleosome center score" and the subsequent "nucleosome ratio", the Pearson correlation coefficient between the fetal concentration and the nucleosome ratio is higher.

[0176] In step S1223, the second alignment numbers of corresponding positions in all nucleosome regions are aligned and summed based on the nucleosome center position to obtain the summed second alignment numbers.

[0177] In some embodiments, when aligning and superimposing according to the nucleosome region, the range of 200 bases to the left and right of the "nucleosome center position" is taken for alignment and summation, that is, a vector of 401 elements is obtained, and the 201st element is the nucleosome center position. Such an operation has no effect on the calculation of the "nucleosome ratio", but because the machine learning algorithm will be introduced to establish the model later, the information in a larger range around the nucleosome region is extracted to improve the potential information content and prediction ability of the model.

[0178] After determining the nucleosome center position, for the sample data to be analyzed, the sequence start position count corresponding to each position on the reference genome can be calculated according to the above method, that is, the second alignment number is obtained, and then based on the processing process of the nucleosome center position, the second alignment numbers of 147 bases to the left and right of the nucleosome center position are taken out. This operation is performed for all nucleosome center positions. After the operation is completed, all the taken out data are added up, that is, the second alignment numbers of the corresponding positions in all nucleosome regions are aligned and added based on the nucleosome center position, wherein the corresponding positions in the nucleosome region include: a range of a specific number of bases to the left and right with the nucleosome center position as the center. After repeated verification, the processing effect of the range of 200 bases to the left and right is the best. That is, a vector of length 295 is removed each time, the 148th element of the vector is the second alignment number of the nucleosome center position, and the other elements of the vector are the second alignment numbers corresponding to the positions to the left and right of the nucleosome center position. These vectors are added up, and the sum result can be seen in Figure 6 As shown, it shows the distribution obtained after a set of data is aligned and summed according to the nucleosome region. Figure 6 It can be seen that the statistical data of the aligned nucleosome regions obtained according to our analysis and processing method indeed show that the probability of the sequence start site falling into the nucleosome center region is much lower than the probability of falling into the nucleosome connection region, which is consistent with the biochemical principle and indirectly proves the reliability of our above analysis method.

[0179] The count mean of the 147 bases in the nucleosome region corresponding to the summed vector (that is, within the range of 73 bases on the left and right of the 148th element of the vector) is calculated, and then the ratio of the mean to the count mean of the left and right nucleosome connecting regions is calculated (for example, the range of 74-93 bases corresponding to the left and right of the center position of the nucleosome is the connecting region of the nucleosome). This ratio can be called the "nucleosome ratio", that is, the comparison result determined based on the nucleosome-related information in the embodiment of the present application. The higher the corresponding "nucleosome ratio", the higher the fetal concentration should be in theory. A linear model is established through the relationship between the nucleosome ratio and the reference fetal concentration to obtain the relationship between the comparison result and the fetal concentration, thereby achieving the purpose of determining the fetal concentration based on the comparison result.

[0180] Specifically, a linear model between the second comparison logarithm and the fetal concentration may also be established, which may be achieved by a machine learning processing method, wherein the fetal concentration is determined based on the comparison result, including:

[0181] (1) obtaining second training sample data, wherein each sample in the second training sample data is labeled with a characteristic value and a target value, the characteristic value is a normalized second logarithm after dimensionality reduction, and the target value is an actual fetal concentration;

[0182] (2) creating a second fetal concentration quantitative model based on the second training sample; inputting the comparison result into the second fetal concentration quantitative model to obtain the fetal concentration.

[0183] The process of generating the second fetal concentration quantitative model is similar to the process of generating the first fetal concentration quantitative model, except that the eigenvalues ​​learned by the two are different. Therefore, the creation process of the first fetal concentration quantitative model can be referred to and will not be described in detail here.

[0184] For example, the ratio of the average count of the nucleosome region to the average count of the nucleosome junction region is used as a parameter to establish a linear model with the known fetal concentration. Due to the low accuracy of the start site in single-molecule sequencing data, when the bases with high error rates at both ends of the nucleosome region are not cut off, the Pearson correlation coefficient is only about 0.3 (see Table 2). Therefore, in one embodiment, the above Figure 6 The counts of all 401 positions in the dataset are used as raw features, and a machine learning model is established (principal component analysis dimensionality reduction + elastic network modeling) to estimate the fetal concentration. Exemplarily, for all a certain number of data such as 615 sets of existing reference fetal concentrations, the count vectors of the alignment and summation of the nucleosome center positions are calculated. Each set of realizations corresponds to a vector containing 401 elements, and these 401 elements are the eigenvalues ​​of the original input. The test set (30%) and training set (70%) were randomly split 100 times, and the Pearson correlation coefficient between the fetal concentration estimated by the test set and the true fetal concentration was 0.58, R 2 is 0.33. Figure 7 The model prediction performance is shown after one of the training sets is randomly split into the test set. Figure 7 In the figure, the abscissa represents the fetal concentration calculated by the nucleosome method, and the ordinate represents the reference fetal concentration.

[0185] In some embodiments, the calculation accuracy of fetal concentration based on sequencing data obtained from various sequencing platforms, especially single-molecule sequencing platforms, can be improved by processing the original feature numbers.

[0186] In one embodiment, the original features are normalized. Specifically, with the middle position as the center, 200 counts (second comparison numbers) are taken on the left and right, and these counts are divided by the counts at the center position, thereby obtaining 401 normalized features. This normalization process can reduce the impact of noise on the model, thereby improving the performance of the model.

[0187] In another embodiment, new features are created based on the original features. Exemplarily, based on the 401 original features mentioned above, two new features are created, the first new feature is the ratio of the count mean of the 20 most coordinate positions to the count mean of the central 147 positions, and the second new feature is the ratio of the count mean of the 20 most right positions to the count mean of the central 147 positions. Thus, 403 features are formed.

[0188] In another embodiment, the original features can be subjected to principal component dimensionality reduction (the principal component transformation model is obtained on the training set and the same transformation is performed on the test set) to simplify the complex relationship between the variables, make the model more concise, and reduce overfitting of the model.

[0189] The above three embodiments can be used alone to process the original features, or can be used in combination to process the original features.

[0190] Exemplarily, after creating new features based on the original features to form 403 features, principal component analysis and dimensionality reduction are performed on the 403 features (the principal component transformation model is obtained on the training set and used for the same transformation on the test set), the number of principal components and the retained information are weighed, and finally the 403 features are reduced to 33 while retaining 90% of the effective information.

[0191] Machine learning was used to establish the model. To avoid overfitting of the model, the model was subjected to L1 and L2 regularization (L1 and L2 regularization is a regularization method that adds additional terms after the loss function to prevent overfitting of the model), and the best combination of L1 and L2 regularization parameters was selected through iteration. When training the model, in order to more accurately evaluate the prediction effect of the model, the test set (30%) and the training set (70%) were randomly split 100 times for model training. The Pearson correlation coefficient between the fetal concentration estimated by the final model test set and the actual fetal concentration was 0.58, R 2 is 0.33. Figure 7 The model prediction performance is shown after one of the training sets is randomly split into the test set.

[0192] Another method for determining fetal concentration is provided in another embodiment of the present application, see Figure 8 , the method may include the following steps:

[0193] S201. Obtain the sequencing sequence of the cfDNA sample to be tested.

[0194] The sequencing sequence is as described above and will not be repeated here to save space.

[0195] S202, segmenting the reference genome to obtain multiple segmentation intervals, and determining the first alignment number of each sequencing sequence falling into each segmentation interval.

[0196] In this step, in order to accurately obtain the comparison results, the reference genome can be segmented according to fixed lengths, such as determining the segmentation lengths of the reference genome according to the characteristics of the sequencing sequence.

[0197] There are some special chromosomes in the reference genome, and these chromosomes may have certain biases or other particularities in the process of determining the fetal concentration by comparing the reference genome. In one implementation scenario, the bias brought by the special chromosomes (hereinafter referred to as specific chromosomes) in the reference genome is reduced by removing the segmented intervals brought by these chromosomes, thereby reducing the impact of their particularity on the test results. In some embodiments, the specific chromosomes include at least one of the X chromosome, the Y chromosome, the mitochondrial chromosome, the chromosome 13, the chromosome 18 and the chromosome 21. Among them, by removing the segmented intervals corresponding to the X chromosome and the Y chromosome, the difference between the above chromosomes of male and female fetuses is avoided, and the impact on the results of the subsequent concentration quantitative model is avoided; chromosomes 13, chromosomes 18 and chromosomes 21 are removed because these chromosomes themselves have a large probability of duplication and deletion. One of the screening goals of NIPT is to screen for duplications and deletions of chromosomes 13, 18, and 21. When chromosomes 13, 18, and 21 are used as reference chromosomes, deviations may occur when normalizing the data, such as affecting the normalized logarithm, which in turn affects the calculation results and the accuracy of the NIPT screening results.

[0198] The embodiment of the present application can adjust the range of specific chromosomes according to the expected needs of fetal concentration detection, such as eliminating X chromosomes and Y chromosomes, or eliminating chromosomes 13, 18 and 21, or eliminating mitochondrial chromosomes. Of course, multiple situations in the above situations can be eliminated at the same time, such as removing all X chromosomes, Y chromosomes, mitochondrial chromosomes, chromosomes 13, chromosomes 18 and chromosomes 21 or their corresponding segmented fragments, so that the segmented intervals obtained after the reference genome is segmented do not contain the segmented intervals corresponding to the above chromosomes.

[0199] The present application can remove specific chromosomes into segmented intervals in a variety of ways. The segmented intervals obtained after segmentation can be screened based on specific chromosomes, or these specific chromosomes can be removed before segmentation. Specifically, the corresponding processing method can be selected based on actual application requirements, and the present application does not limit this.

[0200] In one embodiment, the reference genome is segmented to obtain multiple segmentation intervals, including: segmenting the reference genome based on a preset segmentation length to obtain multiple initial segmentation intervals; removing the segmentation interval corresponding to a specific chromosome in the initial segmentation interval to obtain multiple segmentation intervals. That is, after the reference genome is segmented into multiple segmentation intervals, the segmentation interval corresponding to the specific chromosome in the reference genome is removed and does not participate in the calculation. It should be understood that the "segmentation interval corresponding to a specific chromosome" refers to the multiple segmentation intervals formed after the specific chromosome is segmented according to the preset segmentation length.

[0201] In another embodiment, segmenting the reference genome to obtain multiple segment intervals includes: removing a specific chromosome from the reference genome to obtain a reference genome after the removal; and segmenting the reference genome after the removal based on a preset segmentation length to obtain multiple segment intervals. That is, before segmenting the reference genome, the specific chromosome is first removed.

[0202] The preset segmentation length can be adjusted according to the amount of data to be processed and the expected relative accuracy, for example, the reference genome can be segmented according to the length of each 50k (5000) bases. In order to make the obtained first alignment number more accurate, the first alignment number can also be corrected to obtain a corrected first alignment number, so that the corrected first alignment number is determined as the alignment result.

[0203] In some embodiments, the correction includes GC correction. GC correction is an operation in the bioinformatics analysis process. The reason for GC correction is that the gene sequencer and the corresponding sequencing process will cause a certain GC preference. For example, some sequencers have a higher probability of measuring high GC sequences, while some sequencers have a higher probability of measuring low GC sequences. In this embodiment, the processing method for determining the fetal concentration is based on the comparison probability of different regions on the reference genome, and the GC preference of the sequencer will affect the judgment of the actual comparison probability of different regions. In order to improve the accuracy, GC correction is required.

[0204] At this time, the first alignment number is corrected to obtain the corrected first alignment number, including: based on the first alignment number of each segmented interval and the average first alignment number of all segmented intervals, the first alignment number of each segmented interval is normalized to obtain the normalized first alignment number of each segmented interval; the normalized first alignment number of each segmented interval is GC corrected to obtain the GC corrected first alignment number. The average first alignment number of all segmented intervals refers to the ratio of the sum of the first alignment numbers of each segmented interval to the total number of segmented intervals.

[0205] After obtaining the first comparison numbers of all segmented intervals, in order to facilitate calculation and processing, the first comparison numbers of all segmented intervals will be normalized. The corresponding formula is:

[0206] Normalized first alignment number of each segment interval=first alignment number of each segment interval / average first alignment number of all segment intervals.

[0207] After obtaining the first alignment number of each segmented interval, GC correction will be performed to obtain the first alignment number after GC correction.

[0208] In one embodiment, performing GC correction on the normalized first alignment number of each segment interval to obtain the GC-corrected first alignment number includes:

[0209] (1) Generate a first relationship curve based on the GC content and the normalized first alignment number corresponding to each segment interval.

[0210] In this embodiment, the GC content can be represented by a variety of parameters, such as the GC content or the median GC content of each segmented interval, or other parameters that can provide feedback on the GC content of the segmented interval.

[0211] In an embodiment of the present application, GC filtering can be performed on the segmented intervals forming the first relationship curve, such as filtering the segmented intervals according to GC content to filter out some abnormal or statistically insignificant segments, thereby optimizing the first relationship curve and reducing the occupancy of data processing resources during data processing.

[0212] In one embodiment, GC filtering is achieved in the following manner: before the step of generating a first relationship curve based on the GC content and the normalized first logarithm corresponding to each segmented interval, the following processing is performed: the segmented intervals are filtered based on the GC content, and the segmented intervals whose GC content does not meet the preset requirements are removed, so that the first relationship curve is generated based on the GC content of each segmented interval after GC filtering and the normalized first logarithm of each segmented interval.

[0213] In another embodiment, GC filtering is achieved in the following manner: in the process of performing GC correction on the normalized first logarithm based on the first relationship curve, the segmented intervals are filtered based on the GC content, and the normalized first logarithm after GC filtration is GC corrected based on the first relationship curve to obtain the GC-corrected first logarithm.

[0214] In the above embodiment, as an example, based on the GC content and the normalized first alignment number corresponding to each segmented interval, a first relationship curve is generated, which is achieved in the following manner: the x-axis represents the GC content of each segmented interval (i.e., the proportion of GC bases corresponding to the segmented interval on the reference genome), and the y-axis represents the normalized first alignment number corresponding to each segmented interval. Integrate the data of each segmented interval and the GC content within the interval, that is, generate a scatter plot based on the GC content of each segmented interval and the corresponding normalized first alignment number. Smooth the scattered data in the scatter plot to obtain a smooth curve. Among them, smoothing the scattered data in the scatter plot can be performed using a LOWESS algorithm, for example, to obtain a smooth curve, that is, to obtain the first relationship curve, such as y. s =f(x s ).

[0215] Specifically, as an example, the first relationship curve can be generated in the following manner: divide the horizontal axis into different segments, then calculate the median GC content of the scattered points in the same segment, use this median GC content to represent the value corresponding to this segment, and then draw all these GC content medians on the graph for smoothing to obtain the first relationship curve. In another example, smoothing can be performed directly based on all the data, that is, all data points are taken into consideration and smoothed together to obtain a smooth curve, so that the Spearman correlation coefficient obtained for the test set is higher. It should be noted that this smoothing method is also applicable to the smoothing and correction of the comparison probability described later.

[0216] (2) Performing GC correction on the normalized first alignment number based on the first curve to obtain the GC-corrected first alignment number.

[0217] In some embodiments, GC correction is performed based on a subtraction relationship or a division relationship between the normalized first alignment number and the GC content of the segmented interval corresponding to the normalized first alignment number to determine the GC-corrected first alignment number.

[0218] In one example, the GC correction method is "subtraction", and the corresponding correction formula is: GC =yf(x); In another example, the GC correction method is "division", and the corresponding correction formula is: GC =y / f(x).

[0219] In the above two correction formulas, x is the GC content of the segmented interval, y is the normalized first logarithm corresponding to the segmented interval, and y GC is the normalized first alignment number after GC correction.

[0220] In some embodiments, the correction further includes alignment probability correction to improve the accuracy of the obtained first alignment number. That is, in some embodiments, the correction includes both GC correction and alignment probability correction.

[0221] In one implementation of the embodiments of the present application, the first alignment number is corrected to obtain the corrected first alignment number, and the method further includes: truncating the reference genome based on a specific sliding window length, aligning the truncated sequence with the reference genome, and counting the first alignment number in each segmented interval of the reference genome, and determining the first alignment number as the sliding window alignment number; based on the sliding window alignment number, determining the normalized sliding window alignment number of each segmented interval; performing alignment probability correction based on the normalized sliding window alignment number of each segmented interval and the normalized first alignment number, and determining the first alignment number after the alignment probability correction.

[0222] The embodiment of the present application adopts the "matching probability correction" method to correct the first comparison number, based on the following considerations: if different regions of the reference sequence corresponding to the reference genome are cut off according to a specific length and used as a sequencing sequence, and then compared back to the reference genome, the number of times different segmented intervals are compared will be different, because for the sequences in some regions, there are similar sequences in many places on the reference sequence of the reference genome, so these regions are more easily compared. For example, the human genome reference sequence hg19 is subjected to a sliding window to cut the sequence according to a step length of 2 bases and a window of 37 bases, and these sequences are compared with the human reference genome hg19, and then the number of comparisons on different segmented regions of the human reference genome is counted, and this comparison number can be called the "sliding window comparison number". Specifically, the specific sliding window length can be considered based on the actual application scenario, such as based on the characteristics of the reference genome adopted, or the characteristics of the sample to be tested. After determining the specific sliding window length, the reference genome is intercepted, the intercepted sequence is compared with the reference genome, and the first alignment number on each segmented interval on the reference genome is counted, and the first alignment number at this time is determined as the sliding window alignment number. After obtaining the sliding window alignment number, in order to facilitate calculation and processing, the normalized sliding window alignment number of each segmented interval is determined based on the sliding window alignment number; the normalized sliding window alignment number is corrected for alignment probability. The method for alignment probability correction can refer to the GC correction method mentioned above.

[0223] In one embodiment, performing alignment probability correction based on the normalized sliding window alignment number and the normalized first alignment number of each segmented interval, and determining the first alignment number after alignment probability correction, comprises:

[0224] (1) Based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval, a second relationship curve is generated.

[0225] Among them, the normalized sliding window ratio logarithm of each segmented interval can be represented by multiple parameters, such as the normalized sliding window ratio logarithm of each segmented interval or the median of the sliding window ratio logarithm, or other parameters that can feedback the normalized sliding window ratio logarithm of each segmented interval.

[0226] In an embodiment of the present application, the normalized sliding window ratio logarithm forming the second relationship curve can be filtered to filter out some abnormal or statistically insignificant segments, thereby optimizing the second relationship curve and improving the processing accuracy when performing related data processing based on the second relationship curve.

[0227] In one embodiment, the normalized sliding window logarithm that forms the second relationship curve is filtered, and this is achieved in the following manner: before the step of generating the second relationship curve based on the normalized sliding window logarithm of each segmented interval and the normalized first logarithm corresponding to each segmented interval, the following processing is performed: the segmented intervals are filtered based on the normalized sliding window logarithm to obtain segmented intervals whose normalized sliding window logarithm is not less than the first target threshold, so that the second relationship curve is generated in the segmented intervals whose normalized sliding window logarithm is not less than the first target threshold.

[0228] In one embodiment, the normalized sliding window logarithm that forms the second relationship curve is filtered in the following manner: based on the second relationship curve, the normalized first logarithm is corrected for comparison probability, and in the process of obtaining the first logarithm after the comparison probability correction, the segmented interval is filtered based on the normalized sliding window logarithm, and the second relationship curve interval in which the normalized sliding window logarithm in the second relationship curve is not less than the first target threshold is retained to obtain the second relationship curve after comparison probability filtering; based on the second relationship curve after comparison probability filtering, the sliding window normalized logarithm corresponding to each segmented interval is corrected for comparison probability to obtain the first logarithm after comparison probability correction.

[0229] In the above embodiment, based on repeated verification, the first target threshold may be selected as 0.8, that is, the segmented intervals whose normalized sliding window ratio logarithm is less than 0.8 are removed.

[0230] In the above embodiment, as an example, based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval, a second relationship curve is generated, which is achieved in the following manner: the x-axis represents the normalized sliding window comparison number of each segmented interval, the y-axis represents the normalized first comparison number corresponding to each segmented interval, and the data of each segmented interval and its corresponding GC content are integrated, that is, according to the GC content of each segmented interval and the corresponding normalized first comparison number, a scatter plot is generated; the scattered data in the scatter plot is smoothed to obtain a smooth curve. Exemplarily, the smoothing of the scattered data in the scatter plot can be achieved using the LOWESS algorithm.

[0231] (2) Performing alignment probability correction on the normalized first alignment number based on the second relationship curve to obtain the first alignment number after alignment probability correction.

[0232] In some embodiments, the alignment probability is corrected according to a subtraction relationship or a division relationship between the normalized first alignment number and the sliding window alignment number of the segmented interval corresponding to the normalized first alignment number to determine the first alignment number after the alignment probability correction.

[0233] In one example, the method of comparison probability correction is "subtraction", and the corresponding correction formula is: 比对概率 =yf(x); In another example, the method of comparison probability correction is "division", and the corresponding correction formula is: y 比对概率 =y / f(x).

[0234] In the above two correction formulas, x is the normalized sliding window comparison number of each segment interval, and y corresponds to the normalized first comparison number of each segment interval. 比对概率 is the normalized first pairwise comparison after correction for the comparison probability.

[0235] S203. Determine a first fetal concentration based on the first comparison logarithm.

[0236] In the embodiment of the present application, the fetal concentration can be determined in combination with a machine learning method of a neural network. In one embodiment of the present application, the fetal concentration is determined based on the comparison results, including:

[0237] Obtaining first training sample data, wherein each sample in the first training sample data is annotated with a first characteristic value and a first target value, the first characteristic value is the sample logarithm, and the first target value is the actual fetal concentration of the sample;

[0238] Performing machine learning modeling on the first training sample data based on the specific model result to obtain a first fetal concentration quantitative model;

[0239] The comparison result is input into the first fetal concentration quantitative model to obtain the first fetal concentration, and the first fetal concentration is determined as the fetal concentration.

[0240] In some embodiments, performing machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model includes:

[0241] Dividing the first training sample data into a training set and a test set; performing machine learning modeling based on the training set to obtain an initial model;

[0242] The test set is processed based on the initial model to obtain an estimated fetal concentration for each test sample in the test set;

[0243] Obtaining a comparison result based on comparing the estimated fetal concentration with the actual fetal concentration of each test sample in the test set;

[0244] The model parameters of the initial model are adjusted based on the comparison results to obtain the first fetal concentration quantitative model.

[0245] Among them, the first training sample data can be data from a database of cfDNA in the peripheral blood of pregnant women with known fetal concentrations obtained based on various sequencing platforms including single-molecule sequencing platforms. For example, a certain number of samples with known fetal concentrations are used to segment the reference genome chromosome reference sequence (for example, it can be segmented by 5k bases, 50k bases, 100k bases, 300k bases or 800k bases, etc.), and the number of alignments of the alignment data of each sample in different segmented intervals is counted respectively, and the alignment numbers are normalized, GC and alignment probability are corrected using the LOWESS method, and the final normalized and corrected alignment numbers in different segmented intervals are used as feature values, and the known actual fetal concentration corresponding to each sample is used as the target value, and a specific model is used for learning and modeling, wherein the specific model can be a classification model, a convolutional neural network model, etc. The embodiment of the present application does not limit the specific model structure, as long as the training sample data can be learned and the fetal concentration can be predicted. The process of model learning and training is an iterative process, so the training samples can be divided into training sets and test sets. The accuracy of the model currently trained is tested by the test set. If the accuracy is lower than the set accuracy threshold, the model parameters in the model structure of the current model are adjusted by the test set until the error between the fetal concentration output by the model and the actual fetal concentration meets the corresponding error range.

[0246] When the number of samples used for machine learning is small, but the number of eigenvalues ​​is large (e.g., about 30,000 to 600,000 features, depending on the size of the chromosome reference sequence interval segment), the effect of overfitting cannot be avoided in both dimensionality reduction and learning. In the final fetal concentration quantitative model established, the predicted effect of the test set fetal concentration will deviate from the actual fetal concentration. For example, in the fetal concentration quantitative model established using 615 samples with known fetal concentrations, the model corresponds to the test set R 2Only about 0.3, the Pearson correlation coefficient between true value and predicted value is about 0.5.But, when adopting the aforesaid method provided by the present application embodiment to normalize the comparison logarithm, and correct (GC correction, or GC and comparison probability correction), the correlation coefficient between the predicted value of test set and the true value improves.Compare respectively directly with the normalized segmentation interval comparison logarithm modeling, and use normalization and according to the comparison logarithm of the segmentation interval after GC and comparison probability correction to model, can obtain the modeling effect after normalization, GC correction and comparison probability correction better, that is, the correlation coefficient between the predicted value of test set and the true value is higher.

[0247] In another embodiment, determining the fetal concentration based on the comparison results comprises:

[0248] Inputting the comparison result into the first preset model to obtain an initial fetal concentration;

[0249] The initial fetal concentration is corrected according to the second preset model to obtain the fetal concentration.

[0250] Among them, the first preset model is a linear relationship pattern established based on the comparison results in the second sample data and the initial fetal concentration corresponding to the comparison results. The second sample data is the comparison results obtained by comparing the sequencing sequences in the cfDNA sample with the reference genome, and the fetal concentration corresponding to the comparison results; the second preset model is a model obtained by processing the first preset model based on the constants determined by phenomenon fitting.

[0251] Exemplarily, the first preset model can be in, is the comparison result, FF1 is the initial fetal concentration. The second preset model may be FF2=FF1·c+d, c and d are constants determined by linear fitting, and FF2 is the fetal concentration.

[0252] Exemplarily, the model that uses the difference in the probability of alignment of samples in different intervals of the reference genome reference sequence to establish a model and estimate the fetal concentration is SeqFF. Two machine learning models are used in the SeqFF method, one is the weighted ranked selection criterion (WRSC) and the other is the elastic network (Enet). The final output of the fetal concentration result is the average of the two model predictions. When the SeqFF model is tested using data from sequencing platforms, especially single-molecule sequencing platforms, the WRSC model performs better than Enet, that is, the linear correlation coefficient between the fetal concentration estimated by WRSC and the actual fetal concentration is higher than that of Enet or SeqFF, so the model parameters involved in the WRSC algorithm are given priority. For details, please refer to Figure 2 and Figure 3 , Figure 2A schematic diagram of the relationship between the fetal concentration calculated by the Enet model and the reference fetal concentration provided in an embodiment of the present application, Figure 3 A schematic diagram of the relationship between the fetal concentration calculated by the WRSC model and the reference fetal concentration provided in an embodiment of the present application. As can be seen from the figure, the positive correlation between the fetal concentration calculated by the WRSC model and the reference fetal concentration is better than the positive correlation between the fetal concentration calculated by the Enet model and the reference fetal concentration.

[0253] For example, in the WRSC model, the chromosome reference sequence is divided according to a specific number of bases, such as 50k base intervals. Through WRSC algorithm training, the linear relationship between the coverage density and fetal concentration in different reference sequence intervals is found, as shown in the following formula. In the following formula, FF is the fetal concentration, is the coverage density of the reference sequence at different intervals, For is a constant vector of equal dimension, B is the parameter matrix used for linear transformation, and c and d are constants.

[0254]

[0255] Since SeqFF is a model based on the training of second-generation sequencing data, its sequencing principle is quite different from single-molecule sequencing. Therefore, if the above-mentioned model parameters of SeqFF are directly used to calculate the fetal concentration of data obtained by other sequencing platforms including single-molecule sequencing platforms, the calculated fetal concentration is quite different from the actual fetal concentration. The 615 NIPT clinical samples with known fetal concentrations in the above embodiment are still used as an example for analysis and testing. The original comparison file data (such as sam files) of other sequencing platforms including single-molecule sequencing platforms are directly processed using the algorithm and parameters of SeqFF, and the Pearson correlation coefficient between the obtained intermediate parameters and the actual fetal concentration is 0.714.

[0256] The calculated fetal concentration value calculated by the SeqFF model has a good linear relationship with the actual fetal concentration value, so the above formula can be used The model parameters contained in this part calculate the intermediate parameters of fetal concentration, and then use linear regression to fit the values ​​of c and d, so as to obtain the final model parameters suitable for various sequencing platforms including single-molecule sequencing platforms.

[0257] The 615 NIPT clinical samples with known fetal concentrations in the above embodiment are still used as an example for analysis and testing. Instead, only the first alignment number of the sequencing sequence falling into each segmented interval is used for analysis; the regions with abnormal alignment probabilities in X, Y, M, 13, 18, 21 in the reference genome and on mitochondrial chromosomes are excluded for analysis; the correction method (GC correction, or GC and alignment probability correction) is optimized, etc. The data was re-analyzed using the optimized parameters, and the Pearson correlation coefficient between the intermediate parameters and the fetal concentration was 0.755. The test set (30%) and training set (70%) were randomly split 100 times to establish a linear model and determine the above-mentioned constant parameters c and d. The average test set prediction R obtained from 100 modelings 2 The value is 0.561.

[0258] It should be noted that when the linear model is actually established, the split of the test set and the training set is to evaluate the credibility of the model established by the training set that we currently randomly split in limited data. Therefore, the test set (30%) and the training set (70%) were randomly split 100 times, and the average test set prediction R2 value obtained from 100 modelings was 0.561.

[0259] Table 1 above shows 9 examples of determining fetal concentration based on the difference in the probability of alignment of maternal and fetal cfDNA at different position intervals on the reference genome.

[0260] As can be seen from Table 1, compared with the algorithm of Example 1 (Pearson correlation coefficient is 0.71425), the Pearson correlation coefficients of the other algorithms are all greater than 0.74, and the Pearson correlation coefficients of the algorithms of Examples 6, 8 and 9 are greater than 0.75. Compared with other algorithms, the algorithm of Example 8 can obtain the highest Pearson correlation coefficient, which is 0.75471.

[0261] The relationship between the fetal concentration calculated in Example 8 and the reference fetal concentration is shown in the figure below: Figure 5 shown.

[0262] S204, determining the nucleosome center position based on the second alignment number of each base site of the reference genome being the start position of the sequencing sequence alignment.

[0263] This step includes:

[0264] Determine the count of the base sites at which each base site of the reference genome is the starting position of the sequencing sequence alignment as the second alignment number, and calculate the nucleosome center score corresponding to each base site of the reference genome based on the second alignment number;

[0265] The nucleosome center position was determined based on the nucleosome center score and the center score filtering threshold.

[0266] In some embodiments, calculating the nucleosome center score corresponding to each base position of the reference genome includes:

[0267] (1) calculating a first average count of a first specific number of base ranges corresponding to each base site on the left and right of the reference genome;

[0268] (2) calculating a second average count of a second specific number of bases corresponding to the left and right sides of each base site of the reference genome;

[0269] (3) Determine the nucleosome center score corresponding to each site based on the first average count and the second average count.

[0270] The first specific number and the second specific number are determined based on the number of bases in the nucleosome region and the number of bases in the nucleosome junction region. With x representing a base site, the first specific number = (the number of bases in the nucleosome region - 1) / 2, and the second specific number = (the number of bases in the nucleosome region - 1) / 2 + the number of bases in the nucleosome junction region / 2.

[0271] For example, according to literature reports, the length of the nucleosome region is 147 bases, and the connecting regions at both ends of the nucleosome are about 20 bases each. In this case, the first specific number is (147-1) / 2=73, and the second specific number is (147-1) / 2+20=93. That is, assuming that x is the center position of the nucleosome, the range of 73 bases to the left and right of this position is the nucleosome region, and the range of the 74th to 93rd bases corresponding to the left and right of this position is the connecting region of the nucleosome.

[0272] The above formula calculates the ratio of the sum of the average counts of the 74-93 base ranges corresponding to the position to be investigated to the average counts of the 73 base ranges to the left and right of the position to be investigated. Obviously, for a specific nucleosome region, the closer the position to be investigated is to the center of the nucleosome, the larger the value.

[0273] Further, the nucleosome center score corresponding to each site is determined according to the first average count and the second average count. The calculation formula of the nucleosome center score can be expressed as:

[0274]

[0275] In the formula, x is the base site, [x-93,x-74] represents the interval from 93 nucleotides on one side of x to 74 nucleotides on the same side, [x+93,x+74] represents the interval from 93 nucleotides on the other side of x to 74 nucleotides on the same side, and [x-73,x+73] represents the interval from 73 nucleotides on the other side of x to 73 nucleotides on the other side of x.

[0276] In some embodiments, considering that the sequencing error rate of a single-molecule sequencer tends to be higher than that of the second-generation sequencing technology, when performing probability analysis and comparison of each base in the nucleosome region as the center position of the nucleosome, the accuracy and efficiency of determining the center position of the nucleosome can be improved by truncating regions with relatively high error rates at both ends of the sequence.

[0277] In some embodiments, the "nucleosome center score" and the subsequent "nucleosome ratio" are calculated after shortening the left and right ends of the nucleosome region by n bases. n is a natural number less than or equal to 5, and exemplary, n is 1, 2, 3, 4 or 5.

[0278] At this time, the calculation formula of the nucleosome center score can be expressed as:

[0279]

[0280] In the formula, x is the base site, [x-93,x-74-n] represents the interval from 93 nucleotides on one side of x to 74-n nucleotides on the same side, [x+93,x+74-n] represents the interval from 93 nucleotides on the other side of x to 74-n nucleotides on the same side, and [x-73-n,x+73-n] represents the interval from 73-n nucleotides on the other side of x to 73-n nucleotides on the other side of x.

[0281] For example, when n=5, the calculation formula of the nucleosome center score can be expressed as:

[0282]

[0283] In the formula, x is the base site, [x-93,x-69] represents the interval from 93 nucleotides on one side of x to 69 nucleotides on the same side, [x+93,x+69] represents the interval from 93 nucleotides on the other side of x to 69 nucleotides on the same side, and [x-68,x+68] represents the interval from 68 nucleotides on the other side of x to 68 nucleotides on the other side of x.

[0284] Based on the nucleosome center score and the center score screening threshold, the nucleosome center score is determined, including:

[0285] (1) based on the maximum value of the nucleosome center score, determining the position of the maximum value on the reference genome, and determining the position as the first position;

[0286] (2) clearing the nucleosome center scores of the bases of the specific data on both sides of the first position, and determining the maximum value again based on the remaining nucleosome center scores after clearing, and performing determination of the position of the maximum value on the reference genome until the nucleosome center score is less than a second target threshold, and determining the screened position as a candidate nucleosome center position;

[0287] (3) Based on the base center score screening threshold and the nucleosome center score corresponding to the candidate nucleosome center position, the nucleosome center position is determined.

[0288] Among them, the first position is based on the position of the maximum value of the nucleosome center score on the reference genome, that is, find the maximum value of the nucleosome center score calculated above, record their positions on the reference genome, and clear the nucleosome center score of the bases of the specific data on both sides of the position (such as 147 bases on both sides), because the position found is considered to be the nucleosome center position, and there can be no other nucleosome center position within at least 147 bases on both sides of the nucleosome center position. Then find the position corresponding to the maximum value of the current nucleosome center score again, and continue to execute the step of clearing the score of the next year in the nucleosomes on both sides. This cycle stops until the maximum nucleosome center score value is lower than the second target threshold, and the position screened out based on the execution process is the nucleosome center position. In theory, the average technology of the nucleosome connection area should be higher than the nucleosome area. Based on the above formula, the nucleosome center score should be greater than 2. In order to facilitate processing and improve accuracy, in one embodiment, the value of the second target threshold can be set to 2.2. That is, when screening the nucleosome center region, the cycle continues until the value of the maximum nucleosome center score is lower than 2.2 and then the iterative processing is stopped.

[0289] After all candidate nucleosome center positions are obtained, some positions with abnormally high "nucleosome center scores" among the candidate nucleosome center positions may be affected by various position factors such as alignment probability preference, and their corresponding second alignment numbers are unreliable. Therefore, some of these candidate nucleosome center positions are used as the final selected nucleosome center positions to screen the nucleosome center scores.

[0290] Table 2 above provides several Score screening conditions and the Pearson correlation coefficients of fetal concentration and nucleosome ratio under the corresponding conditions.

[0291] S205 , aligning and summing the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position to obtain the summed second alignment numbers.

[0292] In some embodiments, when aligning and superimposing according to the nucleosome region, the range of 200 bases to the left and right of the "nucleosome center position" is taken for alignment and summation, that is, a vector of 401 elements is obtained, and the 201st element is the nucleosome center position. Such an operation has no effect on the calculation of the "nucleosome ratio", but because the machine learning algorithm will be introduced to establish the model later, the information in a larger range around the nucleosome region is extracted to improve the potential information content and prediction ability of the model.

[0293] After the nucleosome center position is determined, for the sample data to be analyzed, the sequence start position count corresponding to each position on the reference genome can be calculated according to the above method, that is, the second alignment number is obtained, and then based on the processing process of the nucleosome center position, the second alignment number of 147 bases to the left and right of the nucleosome center position is taken out. This operation is performed for all nucleosome center positions.

[0294] S206. Determine a comparison result based on the first fetal concentration and the second comparison number after the addition process.

[0295] In this embodiment, the method for processing the feature values ​​provided by the second processing mode is consistent with that when the second processing mode is used alone. The difference is that the fetal concentration predicted in the first processing mode is added as one of the features, and the principal component analysis dimensionality reduction process is performed together. During model training, in order to ensure the stability of the model and the reliability of the model evaluation results, the test set (30%) and the training set (70%) are randomly split 100 times. The elastic network model also performs L1 and L2 regularization on the model, and selects the best combination of L1 and L2 regularization parameters through iteration. The specific effects of different models are detailed in Tables 3 and 4. In the following table, r2_test_enst represents the R value of the test set when modeling with the enst model. 2 ; r2_train_enst represents the R of the training set when using the enst model for modeling 2 ; corr_test_enst represents the Pearson correlation coefficient of the test set when the enst model is used for modeling; corr_train_enst represents the Pearson correlation coefficient of the training set when the enst model is used for modeling. And so on.

[0296] It should be noted that "randomly split the test set (30%) and the training set (70%) 100 times" and calculate the R obtained by 100 splits 2 Mean, STD, Pearson correlation coefficient mean, STD, is to evaluate the credibility of the model and model effect established by our current randomly split training set in limited data. When establishing the model, only one training set and test set split and modeling were done, R 2is 0.61, and the Pearson correlation coefficient is 0.79. As can be seen from the table below, the R 2 The mean values ​​of the Pearson correlation coefficients are 0.61 and 0.78, and the STD values ​​are 0.04 and 0.03, respectively. Therefore, it is believed that the model effect (R 2 The Pearson correlation coefficient is 0.61 and 0.79).

[0297] Table 3

[0298] Mean r2_test_enst 0.6033 r2_train_enst 0.6478 corr_test_enst 0.7835 corr_train_enst 0.8063 r2_test_lasso 0.6078 r2_train_lasso 0.6565 corr_test_lasso 0.7839 corr_train_lasso 0.8111 r2_test_liner 0.6059 r2_train_liner 0.6614 corr_test_liner 0.7824 corr_train_liner 0.8132 r2_test_linerstep 0.6028 r2_train_linerstep 0.6286 corr_test_linerstep 0.7810 corr_train_linerstep 0.7927 r2_test_ridge 0.6066 r2_train_ridge 0.6604 corr_test_ridge 0.7824 corr_train_ridge 0.8132 r2_test_keras 0.5565 r2_train_keras 0.6408 corr_test_keras 0.7724 corr_train_keras 0.8209 r2_test_lgb 0.5340 r2_train_lgb 0.9127 corr_test_lgb 0.7488 corr_train_lgb 0.9724

[0299] Table 4

[0300]

[0301]

[0302] As a possible implementation of the present application, in some embodiments, based on the first fetal concentration and the second alignment number after the summation process, determining the alignment result includes: performing dimensionality reduction processing on the second alignment number after the summation process corresponding to each base site in the nucleosome region to obtain a normalized second alignment number after the dimensionality reduction; determining the alignment result based on the first fetal concentration and the normalized second alignment number after the dimensionality reduction. Alternatively, determining the initial alignment result based on the first fetal concentration and the second alignment number after the summation process; performing dimensionality reduction processing on the initial alignment result, and determining the initial alignment result after the dimensionality reduction as the alignment result.

[0303] Exemplarily, the 401 original features in the second processing mode (i.e., the 401 counts obtained by aligning and summing the nucleosome center positions) and the fetal concentration values ​​directly obtained by the WRSC method in the first determination mode are combined into 402 parameter data as the original input feature values, and the principal component analysis is used for dimensionality reduction, and the machine learning algorithm is used for modeling. The machine learning model can be a variety of models such as elastic network, Lasso, linear regression, stepwise regression, ridge regression, Keras, lightGBM, etc. Compared with other models, ridge regression has better effect, and the R of the fetal concentration estimation of the final test set is 1.34. 2 It is 0.61.

[0304] S207. Determine the fetal concentration based on the comparison results.

[0305] In another embodiment, determining the fetal concentration based on the comparison results comprises:

[0306] The second training sample data is obtained, each sample in the second training sample data is annotated with a characteristic value and a target value, the characteristic value is the normalized second comparison number after dimensionality reduction, and the target value is the actual fetal concentration; a second fetal concentration quantitative model is created based on the second training sample data; the comparison result is input into the second fetal concentration quantitative model to obtain the fetal concentration.

[0307] This embodiment is based on the method of determining the fetal concentration by using the difference in the probability of alignment of the maternal and infant cfDNA at different position intervals on the reference genome sequence in the aforementioned embodiment (referred to as the first determination mode for the convenience of subsequent description); and the method of determining the fetal concentration by using the difference in the distribution of the two ends of the maternal and infant cfDNA on the nucleosome (referred to as the second determination mode). Since these two methods use different principles to determine the fetal concentration, the parameters of the two processing methods are combined to jointly determine the fetal concentration, the effective information of the characteristic value is more, and the estimation ability of the fetal concentration should also be stronger. For example, the 401 original features in the second processing mode (i.e., the 401 counts obtained by aligning and adding the center position of the nucleosome) and the fetal concentration value directly obtained by the WRSC method in the first determination mode are combined into 402 parameter data as the original input characteristic values, and the principal component analysis is used for dimensionality reduction, and the machine learning algorithm is used for modeling, which is conducive to obtaining more accurate fetal concentration data.

[0308] Based on the above embodiment, another embodiment of the present application further provides a device for determining fetal concentration, see Fig. 9 , the device may include:

[0309] A first acquisition unit 10 is used to acquire a sequencing sequence of a cfDNA sample to be tested;

[0310] A first comparison unit 11 is used to compare the sequencing sequence with the reference genome to obtain a comparison result;

[0311] The first determination unit 12 is used to determine the fetal concentration based on the comparison result.

[0312] As a possible implementation of the present application, the first comparison unit includes:

[0313] A first segmentation subunit is used to segment the reference genome to obtain multiple segmentation intervals;

[0314] The first determination subunit is used to determine the first alignment number of each sequencing sequence falling into each segment interval, and determine the first alignment number as the alignment result.

[0315] As a possible implementation of the present application, the first segmentation subunit is specifically used for:

[0316] The reference genome is segmented based on a preset segmentation length to obtain a plurality of initial segmentation intervals;

[0317] Eliminate the segmentation interval corresponding to the specific chromosome in the initial segmentation interval to obtain multiple segmentation intervals;

[0318] or,

[0319] Eliminate specific chromosomes in the reference genome to obtain a reference genome after elimination;

[0320] The reference genome after elimination is segmented based on a preset segmentation length to obtain multiple segmentation intervals.

[0321] The specific chromosome includes at least one of an X chromosome, a Y chromosome, a mitochondrial chromosome, chromosome 13, chromosome 18 and chromosome 21.

[0322] As a possible implementation of the present application, the device further includes:

[0323] The first correction unit is used to correct the first comparison number to obtain a corrected first comparison number, so that the corrected first comparison number is determined as the comparison result.

[0324] As a possible implementation of the present application, the first correction unit includes:

[0325] A first processing subunit is used to normalize the first alignment number of each segmented interval based on the first alignment number of each segmented interval and the average first alignment number of all segmented intervals to obtain a normalized first alignment number of each segmented interval;

[0326] The GC correction subunit is used to perform GC correction on the normalized first alignment number of each segmented interval to obtain the first alignment number after GC correction.

[0327] As a possible implementation of the present application, the GC correction subunit is specifically used for:

[0328] Generating a first relationship curve based on the GC content and the normalized first alignment number corresponding to each segment interval;

[0329] Performing GC correction on the normalized first alignment number based on the first relationship curve to obtain the GC-corrected first alignment number.

[0330] Before the step of generating a first relationship curve based on the GC content and the normalized first alignment number corresponding to each segment interval, the method further includes:

[0331] Filtering the segmented intervals based on the GC content, so as to generate a first relationship curve based on the GC content of each segmented interval after GC filtering and the normalized first logarithm corresponding to each segmented interval;

[0332] or,

[0333] Performing GC correction on the normalized first alignment number based on the first relationship curve to obtain the GC-corrected first alignment number includes:

[0334] The segmented intervals are filtered based on the GC content, and the normalized first alignment number after GC filtering is GC-corrected based on the first relationship curve to obtain the GC-corrected first alignment number.

[0335] Specifically, the GC correction is to determine the first alignment number after GC correction according to the normalized first alignment number and a subtraction relationship or a division relationship of the GC content of the segmented interval corresponding to the normalized first alignment number.

[0336] In another embodiment, the first correction unit further includes:

[0337] A truncation subunit, used to truncate the reference genome based on a specific sliding window length, align the truncation sequence with the reference genome, and count the first alignment number on each segmented interval of the reference genome, and determine the first alignment number as the sliding window alignment number;

[0338] A second determining subunit is used to determine the normalized sliding window comparison number of each segment interval based on the sliding window comparison number;

[0339] The third determination subunit is used to perform comparison probability correction based on the normalized sliding window comparison number and the normalized first comparison number of each segmented interval, and determine the first comparison number after the comparison probability correction.

[0340] As a possible implementation of the present application, the third determining subunit is specifically used for:

[0341] Generate a second relationship curve based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval;

[0342] The normalized first alignment number is subjected to alignment probability correction based on the second relationship curve to obtain the first alignment number after alignment probability correction.

[0343] In one embodiment, before the step of generating the second relationship curve based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval, the step further includes:

[0344] Filtering the segmented intervals based on the normalized sliding window comparison logarithm to obtain segmented intervals where the normalized sliding window comparison logarithm is not less than the first target threshold, so that the second relationship curve is generated in the segmented intervals where the normalized sliding window comparison logarithm is not less than the first target threshold;

[0345] or,

[0346] Performing alignment probability correction on the normalized first alignment number based on the second relationship curve to obtain the alignment probability-corrected first alignment number includes:

[0347] Based on the normalized sliding window comparison logarithm, the segmented intervals are filtered, and the second relationship curve intervals in which the normalized sliding window comparison logarithm in the second relationship curve is not less than the first target threshold are retained to obtain the second relationship curve after comparison probability filtering;

[0348] Based on the second relationship curve after the comparison probability filtering, the sliding window normalized comparison number is corrected for the comparison probability to obtain a first comparison number after the comparison probability correction.

[0349] As a possible implementation of the present application, the first determining unit includes:

[0350] A first sample acquisition subunit is used to obtain first training sample data, each sample in the first training sample data is marked with a first characteristic value and a first target value, the first characteristic value is a comparison number, and the first target value is an actual fetal concentration;

[0351] A first modeling subunit is used to perform machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model;

[0352] The model processing subunit is used to input the comparison result into the first fetal concentration quantitative model to obtain the first fetal concentration, and determine the first fetal concentration as the fetal concentration.

[0353] As a possible implementation of the present application, the first modeling subunit is specifically used for:

[0354] Dividing the first training sample data into a training set and a test set;

[0355] Perform machine learning modeling based on the training set to obtain an initial model;

[0356] The test set is processed based on the initial model to obtain an estimated fetal concentration for each test sample in the test set;

[0357] Obtaining a comparison result based on comparing the estimated fetal concentration with the actual fetal concentration of each test sample in the test set;

[0358] The model parameters of the initial model are adjusted based on the comparison results to obtain the first fetal concentration quantitative model.

[0359] Specifically, the first determining unit is further used for:

[0360] Inputting the comparison result into the first preset model to obtain the initial fetal concentration, wherein the first preset model is a linear relationship model established based on the comparison result in the second sample data and the initial fetal concentration corresponding to the comparison result, and the second sample data is the comparison result obtained by comparing the sequencing sequence in the cfDNA sample with the reference genome, and the fetal concentration corresponding to the comparison result;

[0361] The initial fetal concentration is corrected according to the second preset model to obtain the fetal concentration, wherein the second preset model is a model obtained by processing the first preset model based on constants determined by linear fitting.

[0362] In another embodiment, the first comparison unit is specifically used for:

[0363] Determine the count of the base sites at which each base site of the reference genome is the starting position of the sequencing sequence alignment as the second alignment number, and calculate the nucleosome center score corresponding to each base site of the reference genome based on the second alignment number;

[0364] Determine the nucleosome center position based on the nucleosome center score and the center score screening threshold;

[0365] Based on the center position of the nucleosome, the second alignment numbers of the corresponding positions in all nucleosome regions are aligned and summed to obtain the summed second alignment number;

[0366] Performing dimensionality reduction processing on the second alignment number after the addition processing to obtain a normalized second alignment number after dimensionality reduction, and determining the normalized second alignment number after dimensionality reduction as the alignment result;

[0367] The calculation of the nucleosome center score corresponding to each base site of the reference genome includes:

[0368] Calculate a first average count of a first specific number of base ranges corresponding to each base site on the left and right of the reference genome;

[0369] Calculate a second average count of a second specific number of base ranges corresponding to the left and right sides of each base site of the reference genome;

[0370] The nucleosome center score corresponding to each site is determined according to the first average count and the second average count.

[0371] As a possible implementation of the present application, the calculation formula of the nucleosome center score is:

[0372]

[0373] In the formula, x is the base site, [x-93,x-74-n] represents the interval from 93 nucleotides on one side of x to 74-n nucleotides on the same side, [x+93,x+74-n] represents the interval from 93 nucleotides on the other side of x to 74-n nucleotides on the same side, [x-73-n,x+73-n] represents the interval from 73-n nucleotides on the other side of x to 73-n nucleotides on the other side of x, and n is a natural number less than or equal to 5.

[0374] As a possible implementation of the present application, the nucleosome center score is determined based on the nucleosome center score and the center score screening threshold, including:

[0375] Based on the maximum value of the nucleosome center score, determine the position of the maximum value on the reference genome, and determine the position as the first position;

[0376] The nucleosome center scores of the bases of the specific data on both sides of the first position are cleared, and the maximum value is determined again based on the remaining nucleosome center scores after clearing, and the position of the maximum value on the reference genome is determined until the nucleosome center score is less than the second target threshold, and the screened position is determined as a candidate nucleosome center position;

[0377] The base center score filtering threshold and the nucleosome center score corresponding to the candidate nucleosome center position are used to determine the nucleosome center position.

[0378] Specifically, in the step of aligning and summing the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position, the corresponding positions in the nucleosome region include:

[0379] The range of a specific number of bases around the center of the nucleosome.

[0380] In one implementation, the first determining unit is further configured to:

[0381] Acquire second training sample data, each sample in the second training sample data is labeled with a characteristic value and a target value, the characteristic value is a normalized second comparison number after dimensionality reduction, and the target value is an actual fetal concentration;

[0382] Creating a second fetal concentration quantitative model based on the second training sample data;

[0383] The comparison results are input into the second fetal concentration quantitative model to obtain the fetal concentration.

[0384] Based on the above embodiments, see Fig.10In another embodiment of the present application, another device for determining fetal concentration is provided, which may include:

[0385] A second acquisition unit 20 is used to obtain a sequencing sequence of the cfDNA sample to be tested;

[0386] The second determination unit 21 is used to segment the reference genome to obtain multiple segment intervals, and determine the first alignment number of each sequencing sequence falling into each segment interval;

[0387] A third determination unit 22, configured to determine a first fetal concentration based on the first comparison logarithm;

[0388] A fourth determination unit 23 is used to determine the nucleosome center position based on the second alignment number that each base site of the reference genome is the start position of the sequencing sequence alignment;

[0389] A processing unit 24 is used to align and sum the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position to obtain the summed second alignment numbers;

[0390] A fifth determining unit 25, configured to determine a comparison result based on the first fetal concentration and the second comparison number after the summation process;

[0391] The sixth determination unit 26 is used to determine the fetal concentration based on the comparison result.

[0392] In some embodiments, the second determining unit includes:

[0393] The second segmentation subunit is specifically used for:

[0394] The reference genome is segmented based on a preset segmentation length to obtain a plurality of initial segmentation intervals;

[0395] Eliminate the segmentation interval corresponding to the specific chromosome in the initial segmentation interval to obtain multiple segmentation intervals;

[0396] or,

[0397] Eliminate specific chromosomes in the reference genome to obtain a reference genome after elimination;

[0398] The reference genome after elimination is segmented based on a preset segmentation length to obtain multiple segmentation intervals.

[0399] In some embodiments, the apparatus further comprises:

[0400] The second correction unit is used to correct the first logarithm to obtain a corrected first logarithm, and determine the first fetal concentration with the corrected first logarithm.

[0401] In some embodiments, the second correction unit includes:

[0402] A second processing subunit is used to normalize the first alignment number of each segmented interval based on the first alignment number of each segmented interval and the average first alignment number of all segmented intervals to obtain a normalized first alignment number of each segmented interval;

[0403] The GC correction subunit is used to perform GC correction on the normalized first alignment number of each segmented interval to obtain the first alignment number after GC correction.

[0404] As a possible implementation of the present application, the GC correction subunit is specifically used for:

[0405] Based on the GC content and the normalized first alignment number corresponding to each segmented interval, a first relationship curve matching a specific coordinate system is generated;

[0406] Performing GC correction on the normalized first alignment number based on the first relationship curve to obtain the GC-corrected first alignment number.

[0407] In some embodiments, the second syndrome unit is further configured to:

[0408] The reference genome is intercepted based on a specific sliding window length, the intercepted sequence is compared with the reference genome, and the first alignment number on each segmented interval of the reference genome is counted, and the first alignment number is determined as the sliding window alignment number;

[0409] Based on the sliding window comparison logarithm, determine the normalized sliding window comparison logarithm of each segment interval;

[0410] Based on the normalized sliding window alignment number and the normalized first alignment number of each segmented interval, alignment probability correction is performed to determine the first alignment number after alignment probability correction.

[0411] In some embodiments, the third determining unit is specifically configured to:

[0412] Obtaining first training sample data, each sample in the first training sample data is labeled with a first characteristic value and a first target value, the first characteristic value is a logarithm, and the first target value is an actual fetal concentration;

[0413] Performing machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model;

[0414] The first comparison result is input into the first fetal concentration quantitative model to obtain the first fetal concentration.

[0415] As a possible implementation of the present application, the third determining unit is further configured to:

[0416] Inputting the first alignment number into a first preset model to obtain an initial fetal concentration, wherein the first preset model is a linear relationship model established based on a comparison result in the second sample data and an initial fetal concentration corresponding to the comparison result, and the second sample data is a comparison result obtained by comparing a sequencing sequence in the cfDNA sample with a reference genome, and a fetal concentration corresponding to the comparison result;

[0417] The initial fetal concentration is corrected according to the second preset model to obtain the first fetal concentration, wherein the second preset model is a model obtained by processing the first preset model based on constants determined by linear fitting.

[0418] In some embodiments, the fourth determining unit includes:

[0419] A score calculation subunit, used to determine the count of the base sites at which each base site of the reference genome is the starting position of the sequencing sequence alignment as a second alignment number, and calculate the nucleosome center score corresponding to each base site of the reference genome based on the second alignment number;

[0420] The center position determination subunit is used to determine the nucleosome center position based on the nucleosome center score and the center score screening threshold.

[0421] In some embodiments, the fifth determining unit is specifically configured to:

[0422] Performing dimensionality reduction processing on the summed second alignment numbers corresponding to each base site in the nucleosome region to obtain a normalized second alignment number after dimensionality reduction;

[0423] The normalized second alignment number based on the first fetal concentration and the dimension reduction was determined as the alignment result;

[0424] or,

[0425] Determining an initial comparison result based on the first fetal concentration and the summed second comparison number;

[0426] The initial comparison result is subjected to dimensionality reduction processing, and the initial comparison result after dimensionality reduction is determined as the comparison result.

[0427] It should be noted that the execution process of each unit of the fetal concentration determination device in this embodiment can be referred to in Figure 8 The method for determining the fetal concentration will not be described in detail here.

[0428] In another embodiment of the present application, a readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, each step of the method for determining the fetal concentration as described above is implemented.

[0429] In another embodiment of the present application, an electronic device is further provided, and the electronic device may include:

[0430] Memory, used to store applications and data generated by the running of applications;

[0431] A processor is used to execute an application program to implement any of the above-mentioned methods for determining fetal concentration.

[0432] It should be noted that the specific implementation of the processor in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.

[0433] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0434] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0435] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0436] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for determining fetal concentration, characterized in that: include: Obtaining the sequencing sequence of the cfDNA sample to be tested; Segmenting the reference genome to obtain a plurality of segmented intervals, and determining a first alignment number of each of the sequencing sequences falling into each of the segmented intervals; determining a first fetal concentration based on the first logarithm; Determine the nucleosome center position based on the second alignment number that each base site of the reference genome is the start position of the sequencing sequence alignment; Based on the nucleosome center position, align and sum the second alignment numbers of corresponding positions in all nucleosome regions to obtain the summed second alignment numbers; Determining a comparison result based on the first fetal concentration and the second comparison logarithm after the summation process; Based on the comparison results, the fetal concentration is determined.

2. The method according to claim 1, characterized in that The reference genome is segmented to obtain multiple segment intervals, including: The reference genome is segmented based on a preset segmentation length to obtain a plurality of initial segmentation intervals; Eliminating the segmented interval corresponding to the specific chromosome in the initial segmented interval to obtain multiple segmented intervals; or, Eliminate a specific chromosome in the reference genome to obtain a reference genome after elimination; Segmenting the reference genome after the elimination based on a preset segmentation length to obtain a plurality of segmentation intervals; Optionally, the specific chromosome includes at least one of chromosome X, chromosome Y, mitochondrial chromosome, chromosome 13, chromosome 18 and chromosome 21.

3. The method according to claim 2, characterized in that ,, before the step of determining the first fetal concentration based on the first comparison logarithm, the method further includes: Correcting the first logarithm to obtain a corrected first logarithm, and determining the corrected first logarithm as a first fetal concentration; Optionally, the correcting the first alignment number to obtain a corrected first alignment number includes: Based on the first alignment number of each segmented interval and the average first alignment number of all segmented intervals, the first alignment number of each segmented interval is normalized to obtain a normalized first alignment number of each segmented interval; Performing GC correction on the normalized first alignment number of each segmented interval to obtain a GC-corrected first alignment number; Optionally, performing GC correction on the normalized first logarithm of each segmented interval to obtain the GC-corrected first logarithm includes: Generating a first relationship curve based on the GC content and the normalized first alignment number corresponding to each segment interval; Performing GC correction on the normalized first alignment number based on the first relationship curve to obtain a GC-corrected first alignment number; Optionally, before the step of generating a first relationship curve based on the GC content and the normalized first alignment number corresponding to each segmented interval, the method further includes: Filtering the segmented intervals based on GC content, so as to generate a first relationship curve based on the GC content of each segmented interval after GC filtering and the normalized first logarithm corresponding to each segmented interval; or, The step of performing GC correction on the normalized first alignment number based on the first relationship curve to obtain the GC-corrected first alignment number includes: Filtering the segmented intervals based on GC content, and performing GC correction on the normalized first logarithm after GC filtering based on the first relationship curve to obtain a GC-corrected first logarithm; Optionally, the GC correction is to determine the first alignment number after GC correction according to a subtraction relationship or a division relationship between the normalized first alignment number and the GC content of the segmented interval corresponding to the normalized first alignment number; Optionally, the correcting the first alignment number to obtain a corrected first alignment number further includes: The reference genome is intercepted based on a specific sliding window length, the intercepted sequence is compared with the reference genome, and the first alignment number on each segmented interval of the reference genome is counted, and the first alignment number is determined as the sliding window alignment number; Based on the sliding window comparison logarithm, determining the normalized sliding window comparison logarithm of each segmented interval; Performing a comparison probability correction based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number, and determining a first comparison number after the comparison probability correction; Optionally, performing alignment probability correction based on the normalized sliding window alignment number of each segmented interval and the normalized first alignment number to determine the first alignment number after alignment probability correction includes: Generate a second relationship curve based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval; Performing alignment probability correction on the normalized first alignment number based on the second relationship curve to obtain the first alignment number after alignment probability correction; Optionally, before the step of generating the second relationship curve based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number corresponding to each segmented interval, the step further includes: Filtering the segmented intervals based on the normalized sliding window comparison logarithm to obtain segmented intervals where the normalized sliding window comparison logarithm is not less than a first target threshold, so that the second relationship curve is generated in the segmented intervals where the normalized sliding window comparison logarithm is not less than the first target threshold; or, The performing alignment probability correction on the normalized first alignment number based on the second relationship curve to obtain the alignment probability-corrected first alignment number includes: Based on the normalized sliding window comparison number, the segmented intervals are filtered, and the second relationship curve intervals in which the normalized sliding window comparison number in the second relationship curve is not less than the first target threshold are retained to obtain the second relationship curve after comparison probability filtering; The sliding window normalized alignment number is subjected to alignment probability correction based on the second relationship curve after alignment probability filtering to obtain a first alignment number after alignment probability correction.

4. The method according to claim 3, characterized in that The step of determining a first fetal concentration based on the first comparison logarithm comprises: Obtaining first training sample data, wherein each sample in the first training sample data is labeled with a first characteristic value and a first target value, the first characteristic value is a logarithm, and the first target value is an actual fetal concentration; Performing machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model; Inputting the first comparison result number into the first fetal concentration quantitative model to obtain a first fetal concentration; Optionally, performing machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model includes: Dividing the first training sample data into a training set and a test set; Perform machine learning modeling based on the training set to obtain an initial model; Processing the test set based on the initial model to obtain an estimated fetal concentration for each test sample in the test set; Obtaining a comparison result based on comparing the estimated fetal concentration with the actual fetal concentration of each test sample in the test set; Adjusting the model parameters of the initial model based on the comparison result to obtain a first fetal concentration quantitative model; Optionally, determining a first fetal concentration based on the first comparison logarithm further comprises: Inputting the first alignment number into a first preset model to obtain an initial fetal concentration, wherein the first preset model is a linear relationship model established based on a comparison result in the second sample data and an initial fetal concentration corresponding to the comparison result, and the second sample data is a comparison result obtained by comparing a sequencing sequence in the cfDNA sample with a reference genome, and a fetal concentration corresponding to the comparison result; The initial fetal concentration is corrected according to a second preset model to obtain a first fetal concentration, wherein the second preset model is a model obtained by processing the first preset model based on constants determined by linear fitting.

5. The method according to claim 1, characterized in that The method of determining the nucleosome center position based on each base site of the reference genome being a second alignment number of the sequencing sequence alignment start position comprises: Determine the count of the base sites at which each base site of the reference genome is the starting position of the sequencing sequence alignment as a second alignment number, and calculate the nucleosome center score corresponding to each base site of the reference genome based on the second alignment number; Based on the nucleosome center score and the center score screening threshold, the nucleosome center position is determined.

6. The method according to any one of claims 1 to 5, characterized in that: The step of determining a comparison result based on the first fetal concentration and the second comparison number after the summation process comprises: Performing dimensionality reduction processing on the summed second alignment numbers corresponding to each base site in the nucleosome region to obtain a normalized second alignment number after dimensionality reduction; Determine the comparison result based on the first fetal concentration and the normalized second comparison logarithm after dimension reduction; or, Determining an initial comparison result based on the first fetal concentration and the second comparison number after the summation process; The initial comparison result is subjected to dimensionality reduction processing, and the initial comparison result after dimensionality reduction is determined as the comparison result.

7. The method according to claim 6, characterized in that The calculating the nucleosome center score corresponding to each base site of the reference genome comprises: Calculate a first average count of a first specific number of base ranges corresponding to each base site on the left and right of the reference genome; Calculate a second average count of a second specific number of base ranges corresponding to each base site on the left and right of the reference genome; Determining a nucleosome center score corresponding to each site according to the first average count and the second average count; Optionally, the calculation formula for the nucleosome center score is: Wherein, x is the base site, [x-93,x-74-n] represents the interval from 93 nucleotides on one side of x to 74-n nucleotides on the same side, [x+93,x+74-n] represents the interval from 93 nucleotides on the other side of x to 74-n nucleotides on the same side, [x-73-n,x+73-n] represents the interval from 73-n nucleotides on the other side of x to 73-n nucleotides on the other side of x, and n is a natural number less than or equal to 5; Optionally, determining the nucleosome center score based on the nucleosome center score and the center score screening threshold comprises: Based on the maximum value of the nucleosome center score, determining the position of the maximum value on the reference genome, and determining the position as the first position; clearing the nucleosome center scores of the bases of the specific data on both sides of the first position, and determining the maximum value again based on the remaining nucleosome center scores after clearing, and determining the position of the maximum value on the reference genome until the nucleosome center score is less than a second target threshold, and determining the screened position as a candidate nucleosome center position; The base center score screening threshold and the nucleosome center score corresponding to the candidate nucleosome center position are used to determine the nucleosome center position; Optionally, in the step of aligning and summing the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position, the corresponding positions in the nucleosome region include: The range of each specific number of bases around the center of the nucleosome; Optionally, determining the fetal concentration based on the comparison result comprises: Acquire second training sample data, wherein each sample in the second training sample data is labeled with a characteristic value and a target value, the characteristic value is a normalized second logarithm after dimensionality reduction, and the target value is an actual fetal concentration; Creating a second fetal concentration quantitative model based on the second training sample data; The comparison result is input into the second fetal concentration quantitative model to obtain the fetal concentration.

8. A device for determining fetal concentration, characterized in that: include: A second acquisition unit is used to obtain a sequencing sequence of the cfDNA sample to be tested; A second determination unit is used to segment the reference genome to obtain multiple segment intervals, and determine a first alignment number of each sequencing sequence falling into each segment interval; a third determining unit, configured to determine a first fetal concentration based on the first comparison logarithm; A fourth determination unit, configured to determine the nucleosome center position based on a second alignment number at which each base site of the reference genome is the start position of the sequencing sequence alignment; A processing unit, configured to align and sum the second alignment numbers of corresponding positions in all nucleosome regions based on the nucleosome center position to obtain a second alignment number after the summation; a fifth determining unit, configured to determine a comparison result based on the first fetal concentration and the second comparison logarithm after the summation process; A sixth determination unit is used to determine the fetal concentration based on the comparison result.

9. The device according to claim 8, characterized in that The second determining unit includes: A second segmentation subunit, wherein the second segmentation subunit is used for: The reference genome is segmented based on a preset segmentation length to obtain a plurality of initial segmentation intervals; Eliminating the segmented interval corresponding to the specific chromosome in the initial segmented interval to obtain multiple segmented intervals; or, Eliminate a specific chromosome in the reference genome to obtain a reference genome after elimination; The reference genome after the elimination is segmented based on a preset segmentation length to obtain a plurality of segmentation intervals.

10. The device according to claim 8, characterized in that The device also includes: a second correction unit, configured to correct the first logarithm to obtain a corrected first logarithm, and determine a first fetal concentration using the corrected first logarithm; Optionally, the second correction unit includes: A second processing subunit is used to normalize the first alignment number of each segmented interval based on the first alignment number of each segmented interval and the average first alignment number of all segmented intervals to obtain a normalized first alignment number of each segmented interval; A GC correction subunit, used for performing GC correction on the normalized first alignment number of each segmented interval to obtain a GC-corrected first alignment number; Optionally, the GC correction subunit is used for: Based on the GC content and the normalized first alignment number corresponding to each segmented interval, a first relationship curve matching a specific coordinate system is generated; Performing GC correction on the normalized first alignment number based on the first relationship curve to obtain a GC-corrected first alignment number; Optionally, the second syndrome unit is further used for: The reference genome is intercepted based on a specific sliding window length, the intercepted sequence is compared with the reference genome, and the first alignment number on each segmented interval of the reference genome is counted, and the first alignment number is determined as the sliding window alignment number; Based on the sliding window comparison logarithm, determining the normalized sliding window comparison logarithm of each segmented interval; Performing a comparison probability correction based on the normalized sliding window comparison number of each segmented interval and the normalized first comparison number, and determining a first comparison number after the comparison probability correction; Optionally, the third determining unit is used to: Obtaining first training sample data, wherein each sample in the first training sample data is labeled with a first characteristic value and a first target value, the first characteristic value is a logarithm, and the first target value is an actual fetal concentration; Performing machine learning modeling on the first training sample data based on a specific model structure to obtain a first fetal concentration quantitative model; Inputting the first comparison result into the first fetal concentration quantitative model to obtain a first fetal concentration; Optionally, the third determining unit is further configured to: Inputting the first alignment number into a first preset model to obtain an initial fetal concentration, wherein the first preset model is a linear relationship model established based on a comparison result in the second sample data and an initial fetal concentration corresponding to the comparison result, and the second sample data is a comparison result obtained by comparing a sequencing sequence in the cfDNA sample with a reference genome, and a fetal concentration corresponding to the comparison result; Correcting the initial fetal concentration according to a second preset model to obtain a first fetal concentration, wherein the second preset model is a model obtained by processing the first preset model based on a constant determined by linear fitting; Optionally, the fourth determining unit includes: A score calculation subunit, used to determine the count of the base sites of each base site of the reference genome that are the starting position of the sequencing sequence alignment as a second alignment number, and calculate the nucleosome center score corresponding to each base site of the reference genome based on the second alignment number; A center position determination subunit, used to determine the center position of the nucleosome based on the nucleosome center score and the center score screening threshold; Optionally, the fifth determining unit is used to: Performing dimensionality reduction processing on the summed second alignment numbers corresponding to each base site in the nucleosome region to obtain a normalized second alignment number after dimensionality reduction; Determine the comparison result based on the first fetal concentration and the normalized second comparison logarithm after dimension reduction; or, Determining an initial comparison result based on the first fetal concentration and the second comparison number after the summation process; The initial comparison result is subjected to dimensionality reduction processing, and the initial comparison result after dimensionality reduction is determined as the comparison result.