KR20220029001A - Method of enhancing the proportion of the unique DNA fragment used for NGS analysis of cfDNA to detect low frequency variant - Google Patents

Abstract

In the present application, disclosed is a method for increasing the proportion of a unique fragment used in next generation sequencing (NGS) analysis, in detecting a low frequency variant in cell free DNA (cfDNA) using the NGS analysis. The method according to the present application includes dividing a sample into two or more aliquots so that the proportion of collision with the same sequence, by chance, is equal to or less than a certain level, and lowering a DNA concentration per aliquot to create an NGS library, wherein NGS data generated from each aliquot are combined in an analysis step, to be analyzed, so that most unique DNA fragments can be distinguished. Therefore, it is possible to detect a low frequency genetic variant in ctDNA present in a very small amount, for example, in cfDNA than in general NGS.

Description

{Method of enhancing the proportion of the unique DNA fragment used for NGS analysis of cfDNA to detect low frequency variant}

The present application relates to a technology for analyzing genetic variation of cfDNA using NGS (Next Generation Sequencing).

Most of the cell-free DNA (cfDNA) in the blood is DNA released from haematopoietic cells in healthy people. However, in cancer patients, cfDNA contains circulating tumor DNA (ctDNA) released into the blood from cells destroyed by cancer cell death. This ctDNA contains genetic mutations related to specific cancers, and through monitoring of these genetic mutations, early detection of cancer before lesion occurs, analysis of responses to specific cancer treatments, discovery of mechanisms for generating resistance to anticancer drugs, and residual cancer It is possible to confirm the existence of

One of the methods for detecting such ctDNA is droplet digital PCR (ddPCR), which can test up to 0.001% of ctDNA. There are many cancer-causing DNA markers, but ddPCR has a limited test range.

Another method is the Targeted next-generation sequencing (NGS) method (Corcoran, R. B., & Chabner, B. A. (2018). New England Journal of Medicine, 379(18), 1754-1765). This technology has the advantage of obtaining a genetic profile for a tumor due to its ability to examine the entire exons or specific markers of multiple tumor-associated genes at once.

However, this NGS method mainly uses cfDNA, but there is a problem that the amount of ctDNA contained therein is very limited. ctDNA contains only <0.1-10% of cfDNA. Furthermore, in order to obtain a statistically significant result in NGS sequencing, a minimum of 10 X read depth is required in consideration of the error. Considering that it is a genome equivalent, a minimum of 6 ng of DNA is required. During NGS analysis, the amount of information is lost at each experimental stage, and the final amount of DNA information that can be obtained (conversion rate) is 30%. Therefore, 20 ng of DNA is required to obtain the amount of information corresponding to 6 ng of ctDNA in NGS analysis, but the available DNA for NGS testing in clinical practice is very limited.

In addition, when the molecular barcode method applied to the latest NGS analysis is applied, a decrease in the available data ratio among the produced data due to an increase in the dimer formed by the adapter; In the process of cutting the cancer cell's genomic DNA, data loss due to the occurrence of a case where the same region was accidentally cut and cannot be distinguished from the one that was amplified by PCR through NGS, and the same sequence among billions of reads derived from different cells during the cutting process In the process of mapping a read sequence to a reference genome to remove You need a lot of DNA. This makes it difficult to detect cancer-related genetic mutations present in ctDNA.

Therefore, it is necessary to develop a method capable of increasing the amount of available ctDNA molecular information that can be used for analysis in the NGS test using a limited amount of cfDNA to detect genetic mutations that are present in cfDNA infrequently.

An object of the present application is to provide a method capable of increasing the amount of available ctDNA fragment information that can be used for analysis in an NGS test using a limited amount of cfDNA to detect low-frequency gene mutations.

In one aspect, the present application provides a method for improving the ratio of unique fragments used in the NGS analysis in the detection of low-frequency mutations in cfDNA (cell free DNA) using the NGS analysis.

In one embodiment, the method comprises the steps of (a) providing a specific amount of a cfDNA sample; (b) calculating the total number of genome fragments corresponding to a length of 51 to 330 bp among the genome fragments included in the specific amount of cfDNA, and the number of unique fragments, wherein the number of unique fragments is the total A value obtained by subtracting the sum of collision counts of each genome fragment corresponding to the length of 51 to 330 bp from the number of genome fragments, and the sum of collision counts is calculated from [Equation 1] disclosed herein, (c) the total number of genome fragments calculating the ratio occupied by the number of unique fragments, and dividing the specific amount of the cfDNA sample into a plurality of aliquots to increase the ratio of the unique fragments; and (d) preparing a library by tagging adapters each having a different index for each of the plurality of aliquots, performing NGS analysis, and then integrating the NGS results of each aliquot.

In one embodiment, the detection of low-frequency mutations means detection of mutations in ctDNA included in cfDNA.

In one embodiment, it is intended to detect a genetic mutation in the ctDNA of cancer cells present at a low frequency of less than about 1% among DNA genome fragments included in cfDNA.

In one embodiment, in step (d), the sample is divided into two or more aliquots so that the ratio of the native fragments is at least about 93% or more, or the ratio of events (expressed as collisions) having identical sequences is equal to or less than a certain amount of DNA per aliquot The concentration is lowered to make an NGS library, and the NGS data generated from each aliquot are aggregated and analyzed in the analysis step.

In one embodiment, the specific amount of cfDNA is 20 ng, in which case the cfDNA divides the cfDNA into 4 aliquots such that the ratio of the native fragments is about 93.9%.

In another aspect, the present application provides a library preparation method for improving the ratio of the native fragments used in the NGS analysis in the detection of low-frequency mutations in cfDNA using the NGS analysis.

In one embodiment, the method comprises the steps of (a) providing a specific amount of a cfDNA sample;

(b) calculating the number of total genome fragments corresponding to a length of 51 bp to 330 bp among the genome fragments included in the specific amount of cfDNA, and the number of unique fragments, wherein the number of unique fragments is the total It is a value obtained by subtracting the sum of collision counts of each genome fragment corresponding to a length of 51 to 330 bp from the number of genome fragments, and the collision count is calculated from Equation 1 disclosed herein, (c) from step (b) calculating the ratio of the number of unique fragments to the total number of genome fragments, and dividing the specific amount of the cfDNA sample into a plurality of aliquots to increase the ratio of the unique fragments; and (d) preparing a library for NGS by tagging adapters each having a different index for each of the plurality of aliquots.

In one embodiment, it is intended to detect a genetic mutation in the ctDNA of cancer cells present at a low frequency of less than about 1% among DNA genome fragments included in cfDNA.

In one embodiment, in step (d), the sample is divided into two or more aliquots so that the ratio of the native fragments is at least about 93% or more, or the ratio of events (expressed as collisions) having identical sequences is equal to or less than a certain amount of DNA per aliquot The concentration is lowered to make an NGS library, and the NGS data generated from each aliquot are aggregated and analyzed in the analysis step.

In one embodiment, the specific amount of cfDNA is 20 ng, in which case the cfDNA divides the cfDNA into 4 aliquots such that the ratio of the native fragments is about 93.9%.

The method according to the present application improves the read depth by increasing the amount of available ctDNA information in a situation where the amount of cfDNA that can be obtained from blood is limited in actual medical settings, thereby improving the performance of ctDNA testing using next-generation sequencing (NGS). .

In library preparation according to the general protocol used in the NGS method, fragments with the same sequence are generated due to accidental cuts at the same site although they are derived from different cells in the process of cutting the genomic DNA of cancer cells, and it is difficult to distinguish them from those amplified by PCR. , the data in the NGS analysis process is lost. The shorter the length of the DNA fragment and the higher the DNA concentration, the higher the probability that the same DNA fragment will occur by chance. In the case of ctDNA, the average length is shorter than that of normal DNA, so this event has a higher probability. This makes it particularly difficult to detect genetic mutations found in ctDNA present in small amounts in cfDNA. However, in the method according to the present application, the sample is divided into two or more aliquots so that the ratio of events (expressed as collisions) with the same sequence is less than a certain level by chance, the DNA concentration per aliquot is lowered to create an NGS library, and the NGS data generated from each aliquot is analyzed By combining analysis in steps, most unique DNA fragments can be distinguished, so it is possible to detect low-frequency gene mutations in ctDNA, for example, present in a very small amount in cfDNA than in general NGS.

1 schematically shows a method according to an embodiment of the present application.
Figure 2a is a graph showing the number of fragments (fragments count) according to the length of the genomic DNA fragment present in the actual cfDNA, shows a distribution having two peaks. The peaks are identified at 166 bp and 315 bp, respectively. The distribution of DNA fragments can be regarded as the probability of the appearance of DNA fragments by calculating the proportion in the total.
Figure 2b shows that when there are 6,600 DNA fragments at a specific loci (locus) of the genome (expressed as 6,600 X depth), the number of DNA fragments of a specific length is calculated from the probability of the DNA fragment length calculated in Figure 2a, and the corresponding It is a graph that calculates possible collisions in length. As in Fig. 2a, fragments are most distributed at 166 bp in length, and the proportion of DNA collision fragments count at this length is very high as 40.3%, and most of the cumulative collision fraction occurs in the length between 100 and 200 bp, which is typical of NGS. Represents data that is not used in the analysis and is discarded.
Figure 2c is a graph showing the ratio of the native fragment according to the amount of the starting DNA. At 20 ng, 21.4% of the fragments were classified as duplicates, which cannot be used in normal NGS analysis.
Figure 3 shows that after dividing 20ng of the starting DNA into 2, 4 and 8 aliquots according to the method according to an embodiment of the present application, it is saturated at 4 as showing an increase in FMD (Fragment Mean Depth) according to the number of aliquots .
4 is a graph comparing the FMD value by the method according to the present application and the FMD value of the existing analysis (all duplicates removed). It indicates that the FMD value in the analysis using the aliquot will be higher than in the conventional analysis.
5 shows the amount of DNA amplified in the pre-PCR step, which is the library preparation step, for each aliquot after dividing 20 ng of the starting DNA into 2, 3 and 8 aliquots according to the method according to an embodiment of the present application, the number of aliquots As α increases, the amount of DNA that can be finally obtained increases, but it was found to be saturated when divided into 4 aliquots.
6A and 6B show the VAF (Variant allele frequency) (less than 1%) calling result of the mutation before (a) / after (b) error correction, and the mutation (a) before error correction is majority in VAF 1% false positive (FP) mutation is detected, but after consensus DNA generation and error correction using aliquot information (b), only true positive (TP) remains and all FPs are removed.

In the NGS analysis using cfDNA, the present application is derived from different cells in the cfDNA pool of a specific sample, but the same site is accidentally cut and cannot be distinguished from that amplified by PCR. When a library is prepared by dividing a specific amount of cfDNA into a plurality of aliquots in a manner that minimizes the collision count of

As used herein, “cfDNA (cell-free DNA, cfDNA)” includes genomic fragments of various lengths present in blood, but the chromatin portion that is not protected by histone proteins is mainly truncated and shows the mode at a length of 166 bp. cfDNA is mostly DNA released from haematopoietic cells in healthy people, and includes ctDNA derived from cancer cells in cancer patients as described below due to cancer cell death. cfDNA can be extracted from blood, and a reagent/kit for extracting it is commercially available, and the method is known (Clara Perez-Barrios et al. Traansl Lung Cancer Res 5 (2016).

As used herein, “circulating tumor DNA (ctDNA)” is a genomic fragment derived from cancer cells, included in cfDNA. It contains only <0.1-10% of total cfDNA. Due to the rapid self-replication of cancer cells, ctDNA has fewer histone-protected sites and consequently is shorter than healthy cell-derived cfDNA . al . Sci Transl Med 10, (2018).) short. These ctDNAs contain genetic mutations related to specific cancers, and through monitoring of these genetic mutations using blood, early detection of cancer before lesion occurs, response analysis to specific cancer treatment methods, discovery of mechanisms for generating resistance to anticancer drugs, It can be usefully used to confirm the presence of residual cancer. In one embodiment according to the present application, it is intended to detect a genetic mutation in ctDNA derived from cancer cells that is present at a low frequency of less than about 1% among DNA genome fragments included in cfDNA.

As used herein, “NGS (Next Generation Sequencing)” refers to one of genome sequencing technologies, which can analyze nucleotide sequences at high speed by processing DNA fragments derived from the genome in parallel. To this end, library preparation, including the process of adding and amplifying indexes, molecular barcodes, etc. to fragments, alignment of the calculated raw data, and error handling and derivation of nucleotide sequences through mapping to reference nucleotide sequences, etc. A data analysis process is required. Next-generation sequencing can be used as a variety of analysis platforms depending on the purpose. For example, analysis platforms for next-generation sequencing include Illumina NextSeq, Illumina NovaSeq, ThermoFisher Ion Proton, Pacific Biosciences Sequel II, BGI MGI, etc., and library preparation kits and methods used for each platform are available from the platform manufacturer. can be obtained

Such NGS has the following essential problems due to its characteristics. First, as an error handling method, the error varies depending on the experimental method and the NGS platform. For example, Illumina Inc.'s equipment has an average error rate of 0.1 to 1% per nucleotide. In general, since cfDNA testing aims at a limit of detection (LOD) of less than 1% of the AF (Allele Frequency), traditional NGS experiments cannot distinguish true mutations from errors. On the other hand, the NGS experiment always includes a DNA amplification step using PCR no matter which method is used. However, in DNA amplification, the amplification efficiency of each DNA fragment is different due to various factors such as DNA GC content and DNA length, so it is not possible to obtain a result in which all fragments are amplified to a uniform degree. Therefore, the analysis step compensates for this effect by removing duplicates (including both PCR duplicates and collisions with duplicates amplified from one DNA). In this case, the picard tool is usually used, leaving the same read as the reference genome (referred to as the read sequence read by NGS), and excluding the remaining duplicates. If a random nucleotide error occurs during duplicates, the sequence appears as different reads, although duplicated from one DNA. In traditional NGS, these reads are largely ignored and only the read closest to the reference genome is used for analysis. However, ctDNA is DNA derived from different cells by chance due to its short length, but it often has the exact same sequence. Therefore, the mutated DNA may be ignored. Later, in the ctDNA test using molecular barcoding, barcode sequence or UMI (unique molecular identifier) technology was developed to overcome the above problem. With this technology, reads from different cells can be distinguished using barcode sequences, and PCR errors and true mutations can be distinguished among these reads. This process is generally called error correction. In this case, the sequencing error is calculated as a Q value for each base. In general, the error rates at the beginning and end of the entire Illumina sequencing step are greater. Since the molecular barcode sequence is read at the beginning of sequencing, it is bound to be relatively more susceptible to errors. Due to the problem of the barcode sequence being read incorrectly, the original purpose of distinguishing unique molecules is greatly undermined. To overcome this problem, there have been many attempts such as adjusting the length of the barcode sequence and combining the sequences precisely (Smith et al., Somervuo et al, Genome Res. 27, 491-499 (2017); Somervuo, P) et al . BMC Bioinformatics 19, 257 (2018)). Also, the conversion rate is lowered by the adapter dimer with barcode. Usually, the length of ctDNA is shorter than that of cfDNA and longer than that of adapter dimer. When removing the adapter dimer, the DNA length is used. Because the DNA length is longer due to the barcode sequence, it is more difficult to distinguish from ctDNA, and when the adapter dimer is removed, more ctDNA is lost. As a result, the conversion rate of ctDNA itself is lower than that of total cfDNA.

in other words, The sequence may appear identical in the NGS result because the genomic DNA released from the cell is derived from different cells during the cleavage process, but the same site is accidentally cut. On the other hand, since DNA is amplified by PCR during library preparation during the NGS experiment and finally appears as an overlapping sequence, even if the sequence is coincidentally identical, it is removed as a redundant sequence in the general analysis process. In particular, ctDNA contains a specific gene mutation in cancer cells, so it is important to detect it. However, ctDNA is contained in a very small amount in cfDNA, and there is a problem in that information is lost in the process of removing the redundant sequence and thus cannot be detected.

Due to these problems of NGS, in cfDNA analysis of blood using NGS, information on ctDNA that can be used for analysis is further limited. Also, as mentioned above, the amount of cfDNA contained in blood and the amount of ctDNA contained therein are very limited, and the amount of blood that can be collected in clinical practice is also very limited, so increasing the amount of DNA itself is limited. For example, in a normal person, the average amount of DNA in the blood is about 4.4 ng/ml (Raymond, CK, Hernandez, J., Karr, R., Hill, K. & Li, M. Collection of cell-free DNA for genomic analysis of solid tumors in a clinical laboratory setting. PLoS One 12, (2017).). Among these, since DNA must be left for other tests (eg, ddPCR: 25ng, Real-Time PCR: 1~100ng), the amount of DNA available for NGS test in actual clinical practice is inevitably more limited.

The method according to the present application minimizes sequences lost due to error handling in the NGS analysis process to detect cancer-related low-frequency mutations, for example, less than 1% mutations, present in ctDNA contained in very small amounts in cfDNA. can

Accordingly, in one aspect, the present application relates to a method of improving the ratio of unique fragments used in NGS analysis in the detection of low-frequency mutations in cfDNA (cell free DNA) using NGS (Next Generation Sequencing) analysis.

In one embodiment, the method comprises the steps of (a) providing a specific amount of a cfDNA sample; (b) calculating the total number of genome fragments corresponding to a length of 51 to 330 bp among the genome fragments included in the specific amount of cfDNA, and the number of unique fragments, wherein the number of unique fragments is the total It is a value obtained by subtracting the sum of collision counts of each genome fragment corresponding to the length of 51 to 330 bp from the number of genome fragments, and the sum of collision counts is calculated from the following equation,

.

.

In the above formula, q(k-1;d): Probability of a number equal to k among n numbers in the range of [1,d], k: a specific number, d: a range of numbers, n: the number of numbers.

(c) calculating the ratio of the number of unique fragments to the total number of genome fragments, and dividing the specific amount of the cfDNA sample into a plurality of aliquots to increase the ratio of the unique fragments; and (d) for each of the plurality of aliquots, tagging an adapter each having a different index to prepare a library, performing NGS (Next Generation Sequencing) analysis, and then integrating the NGS results of each aliquot. .

Referring to FIG. 1 , the method according to the present application is characterized in that a sample is divided into two or more aliquots to make an NGS library. Since NGS data generated from each aliquots can be combined and analyzed in the analysis stage to obtain sequence information from a larger amount of ctDNA than general NGS, it enables accurate detection of gene mutations with low VAF (Variant Allele Frequency) of ctDNA. .

Available ctDNA data herein refers to NGS data that can be used to detect mutations in DNA derived from mutated tumor cells in a state in which DNAs derived from normal cells and tumor cells are mixed and cannot be distinguished from each other.

A specific amount of DNA in the method according to the present application means the amount of cfDNA extracted from a blood sample, usually obtainable from blood taken from a patient, and generally allocated for NGS analysis. For example, in a normal person, the average amount of DNA in the blood is about 4.4 ng/ml (Raymond, CK, Hernandez, J., Karr, R., Hill, K. & Li, M. Collection of cell-free DNA for genomic analysis of solid tumors in a clinical laboratory setting. PLoS One 12, (2017).). In general, if about 5 ml of blood is collected from one patient for NGS analysis, about 20 ng of cfDNA can be obtained, but the specific amount may vary depending on the carcinoma or the progress of the cancer.

As a next step, the method according to the present disclosure includes obtaining the number of unique fragments among the fragments by using the number of genomic fragments contained in a specific amount of cfDNA and a collision count.

In the present application, the number of total genome fragments corresponding to a length of 51 to 330 bp among the genome fragments included in a specific amount of cfDNA and the number of unique fragments are calculated. The cfDNA genome fragment has a distribution with two peaks with a length of 5 bp to 991 bp. The peaks, respectively, have a mode at 166 bp and 315 bp, with the first peak at 166 bp having a ratio 16 times greater than the second peak at 315 bp. Among cfDNA fragments, information is lost due to excessive cleavage of less than 50 bp, or fragments exceeding 330 bp are not useful for testing because most are not ctDNA. Accordingly, most of the ctDNA information is contained in a fragment of 51 to 330 bp, and in one embodiment according to the present application, a fragment of 51 to 330 bp in length is used for analysis in cfDNA including genomic fragments of various lengths.

The number of genomic fragments contained in cfDNA is a concept corresponding to the number of molecules of the fragment, and in humans, 1 ng DNA is usually 330 Genome Equivalents (The amount of DNA present in all genes in one cell. This number depends on the size of the genome of a specific organism and is calculated by converting genome base pairs into ug DNA). In one embodiment according to the present application, the length of the genomic fragment of cfDNA is about 51 bp to 330 bp, for example, the number of fragments having a length of 51 - 330 bp in 20 ng of cfDNA is 6,173 with reference to Tables 1 and 2.

In the method according to the present application, the number of fragments with identical sequences is calculated by chance in a specific amount of cfDNA containing genomic fragments of various lengths using collision count, and when the number of fragments with identical sequences is subtracted from the total number of fragments, the unique fragment becomes the number of A case in which DNA sequences are coincidentally identical in a specific DNA sample containing genomic fragments of various lengths is called collision, and the number of DNA fragments with identical sequences can be counted according to the collision counting method.

The Collison count is based on the Birthday paradox, and is the probability of the same individual by chance in a group of a certain size. d), which can be expressed as Equation 1 as follows (Might, Matt. "Collision hash collisions with the birthday paradox". Matt Might's blog. Retrieved 17 July 2015).

[Equation 1]

In the above formula, q(k-1;d): Probability that there is a number equal to k among n numbers in the range of [1,d], k: a specific number, d: a range of numbers, and n: the number of numbers.

For example, it is known that there are 6,600 Genome Equivalents in 20 ng of cfDNA. In this case, based on a specific loci (locus), there are DNA fragments derived from 6,600 different genomes (expressed in 6,600 X depth). ), this fragment exists in various lengths with the same distribution as the probability calculated in FIG. 2A, and the number is the same as the distribution of 2b and the values shown in Table 1. However, assuming that the length reflecting the information of the ctDNA including the actual low-frequency sequence mutation is in the range of 51 to 330 bp, the total of the corresponding genomic DNA fragments is 6,173 calculated in Table 2. And a different collision count value is calculated for each fragment of each length. For example, in the case of a 166bp fragment, according to Table 1, the length distribution probability is 0.02874, which corresponds to 189.687 out of 6,600, and the collision counts at this time are as follows. It is calculated to be 76.4513.

In this way, according to Equation 1, the collision counts of each fragment with a length of 51 to 330 bp are calculated and summed to 1,319, which corresponds to 21.4% of 6,173 (see Table 2).

As mentioned earlier, cfDNA is DNA derived from different cells due to the nature of it, but it often happens that they have exactly the same sequence by chance, and traditional NGS methods cannot distinguish between PCR duplicates or different cells. Therefore, the mutated DNA may be ignored, resulting in loss of available ctDNA information.

Therefore, in the present application, the NGS test is performed by calculating the probability of occurrence of the same DNA by chance, and dividing the sample in a method that minimizes this probability.

For example, when calculated theoretically, as mentioned above, there are 6,600 Genome Equivalents in about 20 ng DNA, and 6,173 fragments having a length of 51 to 330 bp that are analytically significant. Among them, according to the collision count, 1,319 copies (21.4%) found that the DNA sequences were coincidentally identical to each other, making them indistinguishable from PCR duplicates, and these data were not used in NGS analysis and were discarded. Therefore, even DNA information from different cells is discarded in the step of removing PCR duplicates during the analysis process. Due to this, the number of different unique DNA fragments among actual DNA is 4,854 (6,173-1,319), and the ratio of unique fragments is 0.786 (4,854/6,173). That is, only 78.6% of the fragments are used for analysis.

Because unique DNA fragments are from different cells and it is important for infrequent mutation detection that sequences are not accidentally removed as they are considered identical, it is advantageous to increase the ratio of unique fragments as much as possible. As mentioned above, a case in which DNA sequences are coincidentally identical in a specific DNA sample is called collision, and the number of DNA fragments with identical sequences among DNAs of a specific length can be counted according to the collision counting method.

Herein, in order to calculate the probability of occurrence for each length of the cfDNA fragment, the distribution of fragment length was obtained by examining cfDNA from real blood (see FIG. 2a ). And based on this probability, when there is 6,600 X depth at a specific loci, the number of DNA fragments for each length was calculated (see FIG. 2b, Table 1). Table 1 below shows the results of collision counting calculations for fragments with a length of 51 bp to 330 bp, except for too short or long DNA sequences. According to Table 1, when 20ng is used, for example, the collision count ratio is as high as 40.3% at a length of 166bp, which is data that is not used in the existing NGS analysis and is discarded.

[Table 1-1]

[Table 1-2]

[Table 1-3]

[Table 1-4]

[Table 1-5]

[Table 1-6]

[Table 1-7]

[Table 1-8]

[Table 1-9]

On the other hand, when a smaller amount of cfDNA was used, it was found that the collision count was lowered to 6.06% at 166bp, that is, the ratio of the collision count increased as the amount of input cfDNA used in the NGS experiment increased (Table 2). Reference). The smaller the amount of starting DNA, the greater the percentage of data available for analysis. Therefore, it was discovered that more DNA information can be used for analysis when a specific amount of DNA is divided by a certain number and tested with a smaller starting DNA amount.

[Table 2]

Accordingly, in the method according to the present application, the cfDNA sample is divided into a plurality of aliquots to increase the ratio of native fragments, and NGS analysis is performed for each aliquot. For example, referring to Table 1, when the starting DNA is 20 ng, 1,319 (21.4%) fragments are determined to be identical sequences by chance and cannot be used for analysis. In this case, 380 (=95*4) fragments (6.1%) may be identical by chance. As a result, if the NGS library is prepared by dividing into four, 93.9% of the DNA fragment can be used for analysis, and 15.2% more DNA fragment sequences can be used for analysis compared to 78.6% if not divided.

In the method according to the present application, by appropriately dividing and analyzing the starting cfDNA sample so that DNA having the same sequence is minimized by chance in one starting cfDNA sample, low-frequency mutations derived from the unique fragment can be found just by distinguishing each aliquot without a molecular barcode. can

In one embodiment, a library is prepared by appropriately dividing a specific amount of starting DNA so that the ratio of native fragments is at least 93%.

Considering that the amount of cfDNA that can be obtained by taking blood from a patient in clinical practice is 20 ng, in one embodiment, the specific amount of cfDNA is 20 ng, and the cfDNA is divided into 4 aliquots so that the ratio of native fragments is 93.9% do.

In the next step of the method according to the present application, in order to distinguish each divided plural aliquots, a library is prepared by tagging an adapter including a different index for each aliquot, and NGS analysis is performed, and then the NGS results of each aliquot are obtained. Including the step of integrating. In the present application, the index that distinguishes each of the aliquots is referred to as a tube barcode. Selection of adapters including different indices and library preparation including tagging methods may differ depending on the specific platform employed, and those skilled in the art may select an appropriate one with reference to the description of Examples and the like herein.

In one embodiment, the sequence is performed by a multiplex sequencing method on an NGS platform such as Illumina.

In the next step of the method according to the present application, the NGS data of each aliquot is independently mapped to the reference genome and error correction is performed. As a result of this step, one bam file is created for each aliquot. These bam files are combined into one bam file. Thereafter, gene mutations are detected from the integrated bam file using a general mutation analysis program (Mutect2, Varscan, Vardict, Strelka2, etc.).

In another aspect, the present application relates to a library preparation method for improving the ratio of native fragments used for NGS analysis in detecting low-frequency mutations in cfDNA using NGS analysis.

In one embodiment, the method comprises the steps of (a) providing a specific amount of a cfDNA sample; (b) calculating the number of total genome fragments corresponding to a length of 51 bp to 330 bp among the genome fragments included in the specific amount of cfDNA, and the number of unique fragments, wherein the number of unique fragments is the total number of It is a value obtained by subtracting the sum of collision counts of each genome fragment corresponding to the 51-330 bp length from the number of genome fragments, and the collision count sum is calculated from Equation 1, (c) the total genome fragments from step (b) calculating the ratio of the number of the unique fragments to the number, and dividing the specific amount of the cfDNA sample into a plurality of aliquots to increase the ratio of the unique fragments; and (d) preparing a library for NGS by tagging adapters each having a different index for each of the plurality of aliquots.

Each step included in the method may refer to the aforementioned bar.

Hereinafter, examples are presented to help the understanding of the present invention. However, the following examples are only provided for easier understanding of the present invention, and the present invention is not limited to the following examples.

Example

Example 1. Sample Segmentation NGS Analysis Using Different Indices

What was found herein was experimentally verified as follows.

The cfDNA used in the experiment was purchased from SeraseqTM ctDNA mutation mix v2 and SeraseqTM cfDNA mutation mix v2 WT (SeraCare, Milford, MA) and used.

The experimental design is shown in the table below.

[Table 3]

For cfDNA, QC (quality control) was performed using TapeStation (Agilent Inc.), and 20 ng of cfDNA based on TapeStation was used. A (adenosine) was bound to both ends of cfDNA, and an adapter (Illumina, Inc) was bound by ligation. And for application of the dual index method that can prevent even a very small amount of sample swap, PCR primers (Illumina, Inc) including i7 and i5 indexes at the 5' end and 3' end, respectively, are complementary to the adapter was combined with The PCR primers include the following common sequences and index sequences (indicated by [i7] and [i5]) that can distinguish each aliquot.

5'-CAAGCAGAAGACGGCATACGAGAT [i7] GTGACTGGAGTTCAGACGTGTGCTCTTCCGATC-sT-3'

5´-AATGATACGGCGACCACCGAGATCTACAC [i5] ACACTCTTTCCCTACACGACGCTCTTCCGATC-sT-3´

[Table 4]

For the primary library, only the gene of interest was captured using a probe (Celemics, Inc. custom panel) of 106 human genes (see Table 5), and the final library was prepared according to the existing method (Kang, JK et al. Plos). one 2020.May).

[Table 5-1]

[Table 5-2]

[Table 5-3]

Then, the prepared library was sequenced with 2x150 bp paired ends using Nextseq550 Dx (Illumina, SanDiego, CA, USA) equipment, and demultiplexed using the bcl2fastq (v2.19.0.316, Illumina Inc.) program to correspond to each aliquot. A fastq file was created. For each aliquot, two fastq files are created as pairs in forward and reverse directions. For each file, using fastp (version 0.20.0, Shifu　Chen et al.), the adapter sequence read at the end of the read was removed along with the insert (DNA fragment) to create a new fastq file. Then, use FastQC (v0.11.8, Babraham　Institute) for the fastq file (trimmed fastq) with the adapter sequence removed. If the per base sequence quality, overrepresented sequences, and adapter content items are 'good', it is determined that the QC has been passed and the next step proceeded with The trimmed fastq file was mapped to the GRCh38 version reference genome using the BWA-MEM algorithm of bwa (version 0.7.17-r1188) to create a bam file. Gencore (version 0.14.0, Shifu　Chen et al.) was used to correct sequencing errors in Bam and remove PCR duplicates, and a new bam file (collapsed bam) was created as a result. The reads recorded in this collapsed bam file reflect the information of DNA fragments present in blood because errors have been corrected and PCR duplicates have been removed. By proceeding to this process, bam files were created as much as the number divided by aliqout for each sample. To proceed to the next step, each aliquot bam file was merged into a single bam file using the merge function of sambamba (version 0.7.0, Artem Tarasov et al.). And to find out how many DNA fragments are on average for the 106 captured genes, the per base depth is calculated using the depth function of sambamba (version 0.7.0, Artem Tarasov et al.), and the base of all 106 genes The average value was obtained for the fragment mean depth (FMD).

The results are shown in FIG. 3 . When the same 20ng DNA was not divided into aliquots (1 aliquot) or divided into 2, 4, or 8 pieces, the number of unique fragments that could be used for the test was different. The number of unique fragments is saturated when divided by 4 aliquots, and it is difficult to see a big effect when dividing more than this into 8 aliquots. On the other hand, as the number of aliquots increases, the experimental complexity increases and thus the possibility of causing human error increases. Therefore, if the starting DNA amount is 20 ng, the amount of ctDNA information can be optimally obtained when divided into 4 aliquots corresponding to 5 ng for each aliquot.

In addition, it was confirmed that the FMD (Fragment mean depth) value was higher than if it was not analyzed by the method according to the present application using the DNA information obtained by dividing into aliquots (refer to FIGS. 4 and 7). That is, more DNA information can be used for testing using the same starting DNA. Table 6 shows FMD as the average sequencing depth calculated when unique fragments are mapped to the NGS test region.

[Table 6]

[Table 7]

The method of creating the existing library is similar to the method described above. The only difference is that in the method according to the present application, a sample is divided into 4 aliquots and then a different index is used. However , in the existing method, since one sample is tested with one tube, a library was prepared using only one index. .

Also, referring to Table 8, at the same input DNA amount (20ng), the amount of DNA per individual aliquot decreases as the number of aliquots increases, and the DNA of each aliquot is amplified in the pre-PCR step. It is saturated when there is 5 ng of DNA in it. As a result, it was found that the total amount of pre-PCR DNA increased as the number of aliquots increased, and was saturated when the number of aliquots was increased to 4 (see FIG. 5). This is a phenomenon similar to saturation when the FMD value is 4 aliquots, and can be viewed as a feature when dividing aliquots into 4 aliquots. Additionally, since 1000 to 2000 ng of amplified DNA is used in one NGS experiment, only the amount of DNA for one analysis can be obtained if the aliquot is not divided. can be obtained, there is an additional advantage that it can be used for re-inspection in the future verification process or used for other experiments.

[Table 8]

Example 2. Variant allele frequency (VAF) improvement of variation before and after error correction in analysis using the method according to the present application

In the case of using molecular barcode in cfDNA analysis, if PCR duplicates of DNA derived from the same cell using barcode are confirmed, the nucleotide sequences are compared with each other. Creates a single consensus DNA sequence that represents duplicates. In this way, the error rate can be reduced to 1/10000 (10e-4).

The method according to the present application assumes that DNA from different cells does not occur by chance when properly divided into aliquots, and all duplicates are assumed to be PCR duplicates, similar to when using molecular barcodes, to make consensus DNA with a majority rule. However, in fact, since there may be identical DNA by chance from different cells, in the error correction process, if there are many different bases, it is excluded from the analysis.

On the other hand, dividing DNA into aliquots itself has the effect of repeated experiments. Using this point, statistically significant discrepancies between aliquots are excluded from the analysis results. As a result, it was possible to lower the error rate by a factor of 10 to 1/100000 (10e-5). As shown in FIGS. 6A and 6B , a large number of false positive (FP) mutations were detected at 1% of the VAF before error correction ( FIG. 6A ). On the other hand, after error correction, only true positive (TP) remained and all FPs disappeared (FIG. 6b).

Although the exemplary embodiments of the present application have been described in detail above, the scope of the present application is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present application as defined in the following claims are also included in the scope of the present application. will belong to

All technical terms used in the present invention, unless otherwise defined, have the same meaning as commonly understood by one of ordinary skill in the art of the present invention. The contents of all publications herein incorporated by reference are incorporated herein by reference.

Claims (8)

In the detection of low-frequency mutations in cfDNA (cell free DNA) using NGS (Next Generation Sequencing) analysis, a method for improving the ratio of unique fragments used in the NGS analysis,
(a) providing a specific amount of a cfDNA sample;
(b) calculating the total number of genome fragments corresponding to a length of 51 to 330 bp among the genome fragments included in the specific amount of cfDNA, and the number of unique fragments, wherein the number of unique fragments is the total It is a value obtained by subtracting the sum of the collision counts of each genome fragment corresponding to the length of 51 to 330 bp from the number of genome fragments,
The collision count sum is calculated from the following equation,

.
In the above formula, q(k-1;d): Probability of a number equal to k among n numbers in the range of [1,d], k: a specific number, d: a range of numbers, n: the number of numbers.
(c) calculating the ratio of the number of unique fragments to the total number of genome fragments, and dividing the specific amount of the cfDNA sample into a plurality of aliquots to increase the ratio of the unique fragments; and
(d) preparing a library by tagging adapters each having a different index for each of the plurality of aliquots, performing NGS (Next Generation Sequencing) analysis, and then integrating the NGS results of each aliquot.
The method of claim 1,
The method of claim 1, wherein the low frequency variation is less than 1% frequency.
3. The method of claim 1 or 2,
The method, wherein in step (d), the ratio of the native fragments is at least 93%.
3. The method of claim 1 or 2,
wherein the specific amount of cfDNA is 20 ng, in which case the cfDNA is divided into 4 aliquots of the cfDNA such that the ratio of the native fragments is 93.9%.
In the detection of low-frequency mutations in cfDNA (cell free DNA) using NGS (Next Generation Sequencing) analysis, a library preparation method for improving the ratio of unique fragments used in the NGS analysis, the method comprising:
(a) providing a specific amount of a cfDNA sample;
(b) calculating the number of total genome fragments corresponding to a length of 51 bp to 330 bp among the genome fragments included in the specific amount of cfDNA, and the number of unique fragments, wherein the number of unique fragments is the total number of It is a value obtained by subtracting the sum of the collision counts of each genome fragment corresponding to the length of 51 to 330 bp from the number of genome fragments,
The collision count sum is calculated from the following equation,

.
In the above formula, q(k-1;d): Probability of a number equal to k among n numbers in the range of [1,d], k: a specific number, d: a range of numbers, n: the number of numbers.
(c) calculating a ratio of the number of unique fragments to the total number of genome fragments from step (b), and dividing the specific amount of cfDNA sample into a plurality of aliquots to increase the ratio of unique fragments; and
(d) for each of the plurality of aliquots, each tagging an adapter including a different index to prepare a library for NGS.
6. The method of claim 5,
The method of claim 1, wherein the low frequency is less than 1%.
7. The method according to claim 5 or 6,
The method, wherein in step (d), the ratio of the native fragments is at least 93%.
7. The method according to claim 5 or 6,
wherein the specific amount of cfDNA is 20 ng, in which case the cfDNA is divided into 4 aliquots of the cfDNA such that the ratio of the native fragments is 93.9%.

Images