A truly unbiased in vitro assay (Abnoba-Seq) to reveal the off-target activity of one or more target-specific programmable nucleases in cells

Abstract

The present invention relates to an in vitro method for detecting non-specific sites of one or more target-specific programmable nucleases in a genome, which allows for the detection of non-specific sites in an input sample of less than 1 μg of genomic DNA. The present invention also relates to an in vitro method for detecting non-specific sites of one or more target-specific programmable nucleases in a genome, using an in vitro method for detecting non-specific sites of one or more target-specific programmable nucleases in a genome.

Description

TECHNICAL FIELD OF THEINVENTION

[0001] The present invention relates to an in vitro method for detecting one or more non-specific sites of a target-specific programmable nuclease in a genome, which allows the detection of non-specific sites in an input sample of less than 1 μg of genomic DNA. The present invention also relates to a method for detecting non-specific sites in vivo using an in vitro method for detecting one or more non-specific sites of a target-specific programmable nuclease in a genome. Background

[0002] Genome editing by clustered regularly interspaced short palindromic repeats (CRISPR) / Cas-type nucleases, as well as transcription activator-like effector nucleases (TALENs) and zinc finger nucleases (ZFNs), are already being used therapeutically in clinical studies. These customized target-specific programmable nucleases can be targeted to specific regions in the genome, where they introduce double-stranded DNA (dsDNA) breaks that trigger cellular DNA repair mechanisms. The introduced dsDNA breaks are repaired by either nonhomologous end joining (NHEJ) or homology-directed repair (HDR), the two main cellular repair pathways (Hustedt and Durocher, 2016, Nat Cell Biol 19(1):1-9).

[0003] While target-specific programmable nucleases such as ZFNs, TALENs, and meganucleases rely on protein-DNA recognition, the specificity of CRISPR / Cas RNA-guided nucleases (CRISPR / Cas RGNs) relies on RNA-DNA recognition. In CRISPR / Cas RGNs, guide RNAs (gRNAs) hybridize to a sequence (target sequence) of about 20 nucleotides in length next to a protospacer adjacent motif (PAM), allowing the Cas endonuclease to cleave the target site (Wiedenheft et al., 2012, Nature 482, 331-38; Term et al., 2011, Curr Opin Microbiol 14,321-327). By changing the nucleotide sequence at the 5' end of the gRNA, the site targeted by CRISPR / Cas RGN is altered. In its active form, CRISPR / Cas RGN is a ribonucleoprotein (RNP) in which Cas endonuclease is bound to guide RNA (gRNA). Commonly used Cas endonucleases include Cas9, Cpf1, and their variants. For example, different Cas9 endonucleases from Streptococcus pyogenes, Neisseria meningitidis, Staphylococcus aureus, Streptococcus thermophilus, and Treponema denticola have been used to develop genome editing tools.

[0004] In addition to the actual target sequence (specific site), target-specific programmable nucleases can be activated at sites in the genome that differ to some extent from the actual target sequence, the so-called non-specific site, thereby producing dsDNA cleavage mutations at the non-specific site that differs from the specific site by up to a few nucleotides. For example, it has been shown that CRISPR-Cas9 nucleases can tolerate up to six mismatched base pairs between the gRNA and the target DNA, especially in the PAM-distal region of the target site (e.g., Tsai et al., 2015, Nat Biotechnol 33(2):187-197).

[0005] Off-target activity in human cells can have potentially harmful consequences. The production of dsDNA breaks at non-specific sites in the genome can produce insertions and deletions (indels) as well as translocations that need to be carefully monitored. For example, if the non-specific sites are located within coding regions or regulatory elements, in the worst case, they can induce oncogenic degeneration of cells. Therefore, the off-target activity of target-specific programmable nucleases must be measured and, if necessary, the specificity of the target-specific programmable nucleases used must be optimized before the genome editing tool can be used in a clinical setting. Therefore, in the context of therapeutic use of target-specific programmable nucleases, understanding and prevention of off-target effects are important.

[0006] Importantly, individual differences in the genome, such as single nucleotide polymorphisms (SNPs), as well as cell type-specific chromatin states, can alter the on- and off-target profile of any target-specific programmable nuclease (Lessard et al., 2017, Proc Natl Acad Sci USA 114((52):E11257-E11266; Verkuijl and Rots, 2019, Curr Opin Biotechnol 55:68-73.). This further highlights the need to identify potential non-target sites in a cell type-specific and individualized manner before applying a particular target-specific programmable nuclease therapeutically.

[0007] Various methods have been developed to measure the off-target activity of target-specific programmable nucleases, each with its advantages and disadvantages (summarized in Kim et al., 2019, Annu Rev Biochem 88:191-220). On the one hand, potential non-specific sites can be predicted in silico using bioinformatic algorithms based on the sequence similarity of such potential non-specific sites to the actual target sequence and then investigated by target amplicon sequencing, for example, using next-generation sequencing (NGS). The impact of already a single mismatched base pair compared to the on-target sequence is not necessarily predictable. Also, the genomic or epigenomic context, or both, may also affect the cleavage frequency. For this reason, it is very challenging to develop algorithms that can identify potential non-specific sites. Such in silico-based methods are fast and relatively inexpensive.

[0008] However, these in silico methods tend to have low true positive rates (low sensitivity) and will not detect off-target activity at sites missed by such algorithms.

[0009] Alternatively, off-target activity can be measured experimentally. This means that the detection is not triggered by computer-aided prediction. Experimental methods are more laborious and expensive. They also suffer from technical and bioinformatics limitations that must be overcome. Different experimental approaches for measuring specificity are known, for example GUIDE-Seq, DISCOVER-Seq, HTGTS, TEG-Seq, Digenome-Seq, CIRCLE-Seq, and SITE-Seq. These can be broadly divided into cell-based methods and in vitro methods.

[0010] The cell-based assays / methods GUIDE-Seq (Tsai et al., 2015, Nat Biotechnol 33(2):187-197), TEG-Seq (US 2018 / 0016572 A1), DISCOVER-Seq (Wienert et al., 2019, Science 364(6437):286-289), and HTGTS (WO 2016 / 081798 A1) label non-target sites in living cells by utilizing cellular repair mechanisms triggered by target-specific programmable nuclease activity. GUIDE-Seq is a cell-based method that introduces short double-stranded oligodesoxynucleotides (dsODNs) into cells in addition to target-specific programmable nucleases. When the nucleases cleave the genome, the short dsODNs are incorporated into DNA at the site of the double-strand break and subsequently act as starting points for high-throughput sequencing. The method works well, but only for specific cell lines that can be distinguished from cells derived from individuals that differ at key sites in the genome.

[0011] Target Enriched GUIDE-Seq ("TEG-Seq 1"; US 2018 / 0016572 A1) is another cell-based method for detecting on-target and off-target DNA double-strand breaks produced by target-specific programmable nucleases in cell lines. In TEG-Seq, like GUIDE-Seq, the first step of the method is performed intracellularly and involves the co-transfection of CRISPR-Cas and double-stranded DNA "tags" into cell lines. It utilizes cellular DNA repair mechanisms to insert the genomic DNA-cleaved "tags." TEG-Seq is an optimized version of GUIDE-Seq that is adapted for the Ion Torrent family of NGS instruments to enrich for relevant amplicons prior to sequencing, allowing multiplexing of experiments by barcoding, thereby reducing non-specific sequence reads.

[0012] Importantly, TEG-Seq, like GUIDE-Seq, relies on the efficiency of co-transfection of CRISPR-Cas and double-stranded DNA "tags" in cell lines, and on the efficiency of cellular DNA repair machinery to insert the "tags" into cleaved genomic DNA. These are inherent sources of variability (e.g., transfection efficiency and efficiency of cellular DNA repair machinery in the selected cell type / cell line). Also, like GUIDE-Seq, observed toxicity of double-stranded DNA "tags" in primary cells means that TEG-Seq cannot be easily or at all transferred to certain cells (e.g., primary cells from patients). Furthermore, both TEG-Seq and GUIDE-Seq rely on identifying possible off-target sequences and genomic locations based on similarity to guide and PAM sequences to minimize false positives, and therefore may miss nonspecific sites with the least homology to the target site.

[0013] DISCOVER-Seq (Wienert et al., 2019, Science 364(6437):286-289) is another cell-based method based on the ChIP-seq method. When genetic material is cleaved by a target-specific programmable nuclease, this triggers the cellular repair machinery and DNA repair proteins such as MRE11 are recruited to the DNA double-strand breaks. The DNA is then cross-linked with the proteins bound to it. The resulting "chromatin DNA" is fragmented and an antibody against MRE11 is used to precipitate all DNA fragments with bound MRE11 protein. Finally, NGS of the isolated DNA fragments allows the identification of sites with off-target activity. The method measures actual dsDNA breaks in cells, but also those that are not induced by the target-specific programmable nuclease. Also, the method can only detect DNA double-strand breaks that have not yet been repaired at the time of cell fixation. Thus, as can be expected, false positives and a small number of true positive results limit the applicability of the method.

[0014] HTGTS, or high-throughput, genome-wide translocation sequencing (WO 2016 / 081798 A1), is another cell-based assay. HTGTS is based on the detection of translocations induced by the expression of target-specific programmable nucleases that generate on-target and off-target DNA double-strand breaks (DSBs) in cell lines. HTGTS is based on the observation that a DSB, for example at a specific site in a cell, can be ligated to another concomitantly induced DSB, for example at a non-specific site. If the two DSBs are located on different chromosomes, the result is a chromosomal translocation. To identify such events, the genomic DNA of cells transfected with target-specific programmable nucleases is extracted, fragmented, and adapters are ligated to the DNA fragments produced. The translocations are then amplified by linear amplification-mediated (LAM)-PCR using a bait primer binding close to the target site. After purification of the single-stranded PCR products, adapters are added by ligation. Nested PCR with adaptor and locus specific primers amplifies the nucleic acid sequence surrounding the bait-pre junction, and the products of the secondary PCR are sequenced and aligned against a reference sequence to identify the chromosomal location of the DNA double strand break and the translocation.

[0015] In general, these cell-based methods work very well in certain cell lines. However, GUIDE-Seq and TEG-Seq cannot be easily or at all transferred to certain types of cells (e.g., primary cells from patients) due to, for example, the observed toxicity of dsODN in primary cells. On the other hand, DISCOVER-Seq and HTGTS require a lot of input material, as they show both low specificity and sensitivity. For example, DISCOVER-Seq requires 2-10 million cells as experimental input (Wienert et al., 2019, Science 364(6437):286-289), while HTGTS requires 20-100 μg of genomic input (equivalent to 3-15 million cells) (WO 2016 / 081798 A1). Thus, these methods are not well suited for the analysis of the specificity of individualized target-specific programmable nucleases, especially for cells with limited sample volume and availability.

[0016] The in vitro methods Digenome-Seq (Kim et al., 2015, Nat Methods 2(3):237-243), CIRCLE-Seq (Tsai et al., 2017, Nat Methods 14(6):607-614), and SITE-Seq (Cameron et al., 2017, Nat Methods 14(6):600-606) identify nonspecific sites by labeling sites cleaved by CRISPR-Cas RGNs in isolated genomic DNA in vitro.

[0017] In Digenome-Seq, genomic DNA is cut in a test tube using target-specific programmable nucleases. DNA adapters are ligated to the breakage sites where primers can bind, which are then used for high-throughput sequencing. The main drawback of this approach is that the reference genome must also be sequenced, which must be done on expensive high-performance sequencers (e.g., HiSeq) due to the lack of an enrichment step. This makes the approach very expensive. Also, a relatively large amount of genomic input DNA must be used. For the method, see published European patent application EP 3 219 810 A1.

[0018] In CIRCLE-Seq, genomic DNA is first fragmented by sonication, and the individual fragments are ligated into ring-shaped molecules. In the next step, the circular DNA is cut by CRISPR-Cas RGN to obtain linearized molecules. Adapters are ligated to the ends of these linearized molecules, which can then be used for sequencing and identification of non-specific sites. The main drawback of this method is that it requires the use of a large amount of genomic input DNA. For this method, refer to U.S. Patent No. 9,850,484.

[0019] In SITE-Seq, genomic DNA is cleaved in vitro using target-specific programmable nucleases. Biotinylated adapters are ligated to the cleaved sites, allowing for enrichment of molecules after enzymatic fragmentation. Adapters are then ligated at their ends, which can be used for off-target sequencing and identification. The method also requires relatively large amounts of genomic input DNA.

[0020] The main drawback of these in vitro methods is that they require the use of relatively large amounts of genomic input DNA (approximately 8–25 μg depending on the method). This makes off-target analysis difficult or even impossible in precious cell samples with limited sample volume. Furthermore, in vitro methods are generally characterized by a high false positive rate (low specificity). That is, many of the non-specific sites cleaved in vitro are not cleaved in cells. This is because, for example, the concentration of target-specific programmable nucleases in cells is low or these sites are inaccessible due to chromatin organization (Kim et al., 2019, Annu Rev Biochem 88:191-220).

[0021] In summary, the use of these methods in an individualized manner on cells with limited sample amounts is limited mainly because current in vitro assays require large amounts of input DNA or cell-based methods cannot be easily transferred to certain cells, e.g., primary cells.

[0022] What is needed is a method for detecting non-specific sites of one or more target-specific programmable nucleases in a genome that allows for personalized detection of non-specific sites in genomic DNA of precious cell samples of limited sample volume and availability. Summary of the Invention

[0023] The present inventors have found a method for in vitro detection of non-specific sites of one or more target-specific programmable nucleases in a genome (Abnoba-Seq), which allows detection of non-specific sites in the field of genome editing in precious cell samples with limited sample volume. In contrast to established in vitro methods such as Digenome-Seq, CIRCLE-Seq, and SITE-Seq, the method of the present invention requires minimal amounts of genomic input DNA (<1 μg, equivalent to approximately 150,000 cells or less). This allows testing of one or more target-specific programmable nucleases on genomic DNA derived from limited supply of very small cell samples, such as primary cells, biopsy material, or gene-edited grafts. Also, the method of the present invention can be performed using basic (low-cost) NGS sequencers and does not require expensive high-performance sequencers such as HiSeq sequencing systems. Finally, the method is able to identify nonspecific sites with a sensitivity comparable to that of established in vitro methods and with a lower false positive rate, implying significantly greater specificity than comparable methods.

[0024] In a second aspect, the present invention also provides a method for detecting non-specific sites in vivo using the method for detecting non-specific sites of one or more target-specific programmable nucleases in genomes in vitro (Double-Abnoba-Seq). The method thus allows the measurement of the in vivo specificity of one or more target-specific programmable nucleases in valuable cell samples of limited sample volume and availability, in other words the actual off-target activity of one or more target-specific programmable nucleases in cells of a specific low sample volume. This means that the method according to the second aspect of the invention can be used to test target-specific programmable nuclease activity in vivo and demonstrate the actual in vivo activity of a specific target-specific programmable nuclease in a specific tissue or of limited sample volume and availability. In a third aspect, the present invention provides a kit for carrying out the method according to the first aspect of the invention or the method according to the second aspect of the invention.

[0025] Accordingly, the present invention relates to the subject matter defined in the following items [1] to

[84] .

[0026] [1] A method for detecting off-target sites of one or more target-specific programmable nucleases in a genome in vitro, comprising the steps of: (a) providing an input sample comprising genomic DNA, randomly fragmenting the genomic DNA into dsDNA fragments, ligating blocking adaptors to the ends of the dsDNA fragments, and removing dsDNA fragments containing unblocked ends to prepare purified end-blocked dsDNA fragments; (b) preparing a DNA library of cleaved and amplified dsDNA fragments by: b1) cleaving the purified end-blocked dsDNA fragments with one or more target-specific programmable nucleases to obtain cleaved dsDNA fragments; b2) ligating an affinity adaptor to the cleavage site of the cleaved dsDNA fragment to obtain an affinity adaptor-modified dsDNA fragment; b3) enriching the affinity adaptor-modified dsDNA fragments using the affinity adaptors to obtain enriched affinity adaptor-modified dsDNA fragments; and b4) The enriched affinity adaptor-modified dsDNA fragments are amplified by PCR amplification to obtain a DNA library of cleaved and amplified dsDNA fragments. (c) sequencing the DNA library of the cleaved and amplified dsDNA fragments to obtain sequence reads; (d) aligning the sequence reads to a reference genome to obtain aligned sequence reads; and (e) determining putative cleavage sites and identifying cleavage sites by eliminating false positive cleavage sites, wherein the putative cleavage sites are determined by searching for sequence regions in the reference genome where the sequence read start and sequence read end portions of the aligned sequence reads match, and identifying putative cleavage sites in these sequence regions as sites where the sequence read start and sequence read end portions match.

[0027] [2] The method according to claim 1, wherein the genomic DNA in the input sample prepared in step (a) is human genomic DNA.

[0028] [3] The method according to any one of the preceding claims, wherein the genomic DNA is isolated genomic DNA.

[0029] [4] The method of any one of the preceding claims, wherein the input sample comprises less than 1 μg, preferably less than 900 ng, more preferably less than 800 ng, more preferably about 700 ng or less, even more preferably about 650 ng, even more preferably about 600 ng, or preferably about 500 ng, or preferably about 400 ng, or preferably about 300 ng of genomic DNA.

[0030] [5] The method according to any one of the preceding claims, wherein the genomic DNA is fragmented by sonication, acoustic shearing, centrifugal shearing, point-sink shearing, needle shearing, or enzyme-based fragmentation, preferably by sonication.

[0031] [6] The method according to any one of the preceding claims, wherein in step a1), the genomic DNA is randomly fragmented into dsDNA fragments of an average or median fragment size, preferably the median, of 100 to 2000 bp, preferably 200 to 800 bp, more preferably about 300 to 600 bp, even more preferably about 400 to 500 bp, even more preferably about 400 bp, even more preferably 400 bp.

[0032] [7] Step a) a1) randomly fragmenting the genomic DNA to defined lengths to provide dsDNA fragments; a2) protecting the ends of the dsDNA fragments by ligating blocking adapters to obtain end-blocked dsDNA fragments; and a3) contacting the sample containing end-blocked dsDNA fragments with one or more exonucleases to remove non-end-blocked dsDNA fragments to obtain purified end-blocked DNA fragments. 2. The method according to any one of the preceding claims, comprising:

[0033] [8] The method according to any one of the preceding claims, wherein in step a), the blocking adaptor is ligated by enzymatic or chemical ligation, preferably by enzymatic ligation.

[0034] [9] In step a), ligating blocking adapters to the ends of the dsDNA fragments, e1) preparing the dsDNA fragment for blocking adaptor ligation; and e2) Ligating blocking adaptors to the ends of the dsDNA fragments to obtain end-blocked dsDNA fragments. 2. The method according to any one of the preceding claims, comprising:

[0035]

[10] The method of claim 9, wherein preparing the dsDNA fragments for blocking adaptor ligation comprises one or both of end repairing and dA tailing the fragmented dsDNA.

[0036]

[11] The method according to any one of the preceding claims, wherein in step a), the blocking adapter is a DNA-based adapter, preferably a dsDNA-based adapter.

[0037]

[12] The method of any one of the preceding claims, wherein the blocking adaptor comprises one or more modifications that prevent ligation of the end of the blocking adaptor joined to the dsDNA fragment after the blocking adaptor has been ligated to the dsDNA fragment, and one or more modifications that prevent exonuclease cleavage of the dsDNA fragment from the end of the blocking adaptor joined to the dsDNA fragment after the blocking adaptor has been ligated to the dsDNA fragment.

[0038]

[13] The method of claim 11 or 12, wherein the dsDNA based blocking adapter comprises one or more phosphorothioate bonds, one or more internal deoxyuridines, one or more C3-spacers, one or more dideoxy-G / dideoxy-C, or other spacers, and combinations thereof, preferably a combination of one or more C3-spacers and preferably two phosphorothioate bonds, preferably on each strand of the dsDNA based blocking adapter, preferably at the end of the dsDNA based blocking adapter that is not used to ligate the dsDNA based blocking adapter to the dsDNA fragment.

[0039]

[14] The method of any one of claims 11 to 13, wherein the dsDNA based blocking adapter comprises one or more C3-spacers, one or more dideoxy-G / dideoxy-C, or other spacers, preferably one or more C3-spacers, preferably on each strand of the dsDNA based blocking adapter, preferably at an end of the dsDNA based blocking adapter that is not used to ligate the dsDNA based blocking adapter to the dsDNA fragment.

[0040]

[15] The method of any one of claims 11 to 14, wherein the dsDNA based blocking adapter comprises two or more phosphorothioate bonds, preferably at least one on each strand of the dsDNA based blocking adapter, more preferably at least two on each strand of the dsDNA based blocking adapter, preferably at an end of the dsDNA based blocking adapter that is not used to ligate the dsDNA based blocking adapter to the dsDNA fragment.

[0041]

[16] The method of any one of the preceding claims, wherein in step a), the blocking adapter comprises a primer site suitable for use in PCR priming.

[0042]

[17] The method of any one of claims 12 and 16, wherein in step a2), the blocking adapter is a dsDNA-based adapter comprising or consisting of a forward sequence as defined in SEQ ID NO: 01 and a reverse sequence as defined in SEQ ID NO: 02.

[0043]

[18] The method of any one of the preceding claims, wherein in step a), removing dsDNA fragments comprising unblocked ends comprises contacting the sample comprising the end-blocked dsDNA fragments with one or more exonucleases to remove the dsDNA fragments and obtain purified end-blocked DNA fragments.

[0044]

[19] The method of any one of the preceding claims, wherein in step a), the one or more exonucleases comprise one or more of exonuclease III, exonuclease V, exonuclease VIII, lambda exonuclease, T5 exonuclease, T7 exonuclease, nuclease BAL-31, and combinations thereof, preferably exonuclease III, lambda exonuclease, or combinations thereof, optionally further comprising one or more selected from exonuclease I, thermostable exonuclease I, exonuclease T, exonuclease VII, RecJf, mungbean nuclease, nuclease P1, and combinations thereof.

[0045]

[20] Any one of the preceding methods, wherein in step a), the one or more exonucleases consist of a mixture of exonuclease I, exonuclease III, and lambda exonuclease, for example, in an enzyme unit ratio of 2:20:5.

[0046]

[21] The method of any one of the preceding claims, wherein in step a), preferably after contacting the sample containing the end-blocked dsDNA fragments with one or more exonucleases, the free ends of the remaining unblocked DNA fragments are dephosphorylated with a dephosphorylation reagent.

[0047]

[22] The method of claim 21, wherein in step a) the dephosphorylation reagent comprises one or more phosphatases, preferably comprising or consisting of one or more of Antarctic phosphatase, calf intestinal alkaline phosphatase, Quick-CIP, shrimp alkaline phosphatase, and combinations thereof.

[0048]

[23] The method of any one of the preceding claims, wherein in step b1) less than 400 ng, preferably less than 300 ng, more preferably less than 200 ng, even more preferably less than 100 ng, such as about 90 ng, about 80 ng, about 70 ng, about 60 ng, or about 50 ng, or less than about 50 ng, such as about 40 ng, of purified end-blocked dsDNA fragments are cleaved with one or more target-specific programmable nucleases to obtain cleaved dsDNA fragments.

[0049]

[24] The method according to any one of the preceding claims, wherein in step b1) purified end-blocked dsDNA fragments are used which correspond to less than 1,500,000 monoploid genomes, preferably less than 1,500,000 monoploid genomes, more preferably less than 80,000 monoploid genomes, more preferably less than 60,000 monoploid genomes, even more preferably less than 50,000 monoploid genomes, such as about 45,000, about 30,000, about 20,000 or about 15,000 monoploid genomes.

[0050]

[25] The method of any one of the preceding claims, wherein the purified end-blocked dsDNA fragments are cleaved by contacting a sample containing the purified end-blocked dsDNA fragments with one or more target-specific programmable nucleases.

[0051]

[26] The method of any one of the preceding claims, wherein in step b1), the one or more target-specific programmable nucleases are selected from TALENs, ZFNs, meganucleases, megaTALs, Fokl-dCaas9, CRISPR / Cas RNA-guided nucleases (CRISPR / Cas RGNs), Argonaute-family protein-based DNA-guided DNA nucleases, or mutants thereof, or combinations thereof.

[0052]

[27] The method of claim 26, wherein in step b1), the one or more target-specific programmable nucleases comprise, preferably consist of, one or more CRISPR / Cas RGNs, each preferably used as a preassembled ribonucleotide protease (RNP), in which the Cas endonuclease is bound to a guide RNA (gRNA), for example in a molar ratio of 1:3.

[0053]

[28] The method of claim 27, wherein the gRNA is a single gRNA (sgRNA) comprising one RNA or a double gRNA (dgRNA) comprising two RNAs as components, preferably a crRNA (CRISPR RNA) and a tracrRNA (trans-activating crRNA), preferably a sgRNA.

[0054]

[29] The method according to any one of claims 26 to 28, wherein the Cas endonuclease of the CRISPR / Cas RGN is Cas9 endonuclease or a mutant of Cas9 endonuclease, or Cpf1 endonuclease or a mutant of Cpf1 endonuclease.

[0055]

[30] The one or more target-specific programmable nucleases are selected from the group consisting of Cas9 endonuclease or a mutant of Cas9 endonuclease, Cpf1 endonuclease or a mutant of Cpf1 endonuclease, and combinations thereof, wherein preferably, the Cas9 endonuclease or a mutant of Cas9 endonuclease is selected from the group consisting of Streptococcus, Neisseria, Pasteurella, Francisella, and Campylobacter, such as S. pyogenes, Staphylococcus aureus, Neisseria meningitidis, Streptococcus thermophilus, or Treponema denticola, and preferably, the Cpf1 endonuclease or a mutant of Cpf1 endonuclease is selected from the group consisting of Lachnospira, Butyrivibrio, Peregrinibacteria, 30. The method according to any one of claims 26 to 29, comprising, and preferably consisting of, a bacterium selected from the group consisting of the genera Acidominococcus, Porphyromonas, Prevotella, Francisella, Candidatus methanoplasma, and Eubacterium, e.g., Candidatus Paceibacter, Acidaminococcus sp. BV3L6, or Lachnospiraceae bacterium ND2006.

[0056]

[31] The method of claim 26, wherein the meganuclease is selected from the group consisting of l-Scel, l-Ceul, Pl-Pspl, Pl-Scel, l-SceIV, l-Csml, l-Panl, l-Scell, I-Ppol, l-Scelll, l-Crel, l-Tevl, l-Tevll, and l-Tevlll.

[0057]

[32] Step b2) of ligating affinity adapters to obtain affinity adapter-modified dsDNA fragments, e1) preparing cleavage sites for the cleaved dsDNA fragments for ligation; and e2) Ligating affinity adaptors to the cleavage sites of the cleaved dsDNA fragments 2. The method according to any one of the preceding claims, comprising:

[0058]

[33] The method of claim 32, wherein preparing cleavage sites of the cleaved dsDNA fragments for ligation comprises one or both of end repairing and dA tailing the fragmented dsDNA.

[0059]

[34] The method of any one of the preceding claims, wherein in step b2), the affinity adapter comprises a primer site suitable for use in PCR priming.

[0060]

[35] The method according to any one of the preceding claims, wherein in step b2), the affinity adapter is a dsDNA-based adapter, preferably a dsDNA-based adapter comprising one or more modifications for affinity purification.

[0061]

[36] The method of claim 35, wherein the one or more modifications for affinity purification are selected from biotin and biotin-based modifications, preferably biotin and biotin-based modifications, and combinations thereof, including biotin, biotin-dT (biotinylated deoxythymidine), biotin-TEG (15 atom triethylene glycol spacer) desthiobiotin, double biotin, photocleavable (PC) biotin, biotin azide, nitrilotriacetic acid (NTA), DNP TEG (2,4-dinitrophenyl), digoxigenin, protein tags, and combinations thereof.

[0062]

[37] The method of any one of the preceding claims, wherein the affinity adapters comprise one or more modifications that allow for specific enrichment of affinity adapter-modified dsDNA fragments after ligation of the affinity adapters to the cleavage sites of the cleaved dsDNA fragments.

[0063]

[38] The method according to any one of the preceding claims, wherein in step b3), the affinity adapter-modified dsDNA fragments are enriched using affinity adapter binding molecules that can be immobilized on a solid support.

[0064]

[39] The method according to claim 38, wherein in step b3) beads, preferably magnetic beads, are used as the solid support.

[0065]

[40] The method according to claim 39, wherein the magnetic beads are polymer-coated magnetic beads, preferably selected from the group consisting of Dynabeads Myone Streptavidin C1, Dynabeads M-270, Dynabeads M-280, or Dynabeads MyOne Streptavidin T1.

[0066]

[41] The method according to any one of claims 38 to 40, wherein in step b3), the affinity adaptor binding molecule is a protein, preferably selected from streptavidin, avidin, an antibody specific for a part of the affinity adaptor or a functional fragment thereof (e.g. an anti-digoxigenin antibody), a digoxigenin binding protein, or a mutant thereof; more preferably streptavidin, avidin, or a mutant thereof.

[0067]

[42] The method according to any one of claims 38 to 41, wherein in step b3) the affinity adapter binding molecule is streptavidin, preferably immobilized on magnetic beads, preferably polymer-coated magnetic beads, more preferably Dynabeads.

[0068]

[43] The method of any preceding claim, wherein the affinity adapter is a dsDNA-based adapter comprising one or more biotin or biotin-based modifications, and the affinity adapter binding molecule is streptavidin or a variant thereof, preferably immobilized on a magnetic bead, preferably a polymer-coated magnetic bead.

[0069]

[44] In step b4), obtaining a DNA library of cleaved and amplified dsDNA fragments comprises: e1) a first PCR amplification step using enriched affinity adaptor-modified dsDNA fragments bound to a solid support via an adaptor-binding molecule as a template and a primer specific to the affinity adaptor for linear PCR amplification; and e2) a second PCR amplification step in which, after said first step, the enriched affinity adaptor-modified dsDNA fragments bound to the solid support via the adaptor-binding molecules are removed from the reaction mixture and a primer specific for the blocking adaptor is added as a second primer. 2. The method according to any one of the preceding claims, comprising a two-step PCR amplification of:

[0070]

[45] In step c), for sequencing the DNA library of cleaved and amplified dsDNA fragments, c1) preparing a DNA library of the cleaved and amplified dsDNA fragments for sequencing, if necessary, by ligating sequencing adapters to both ends of the dsDNA fragments; c2) The method according to any one of the preceding claims, wherein the DNA library of cleaved and amplified dsDNA fragments produced is sequenced, preferably by next generation sequencing (NGS), to obtain said sequence reads.

[0071]

[46] The method of claim 45, wherein in step c1), producing a DNA library of cleaved and amplified dsDNA fragments for sequencing comprises preparing a DNA library of cleaved and amplified dsDNA fragments for ligating sequencing adapters to both ends, preferably by one or both of end repairing and dA tailing the DNA library of cleaved and amplified dsDNA fragments.

[0072]

[47] The method of claim 45 or 46, wherein in step c1), sequencing adaptors are ligated to both ends of the dsDNA fragments, the sequencing adaptors being, for example, Illumina sequencing adaptors.

[0073]

[48] ​​The method according to any one of claims 45 to 47, wherein in step c2) droplet digital PCR (ddPCR), preferably with a ddPCR Library Quantification Kit for Illumina TruSeq, is used.

[0074]

[49] The method according to any one of claims 45 to 48, wherein in step c2) Illumina sequencing is carried out, for example, on a MiSeq instrument (Illumina), using, for example, the MiSeq Reagent Kit v2.

[0075]

[50] The method according to any one of claims 1 to 44, wherein in step b4), each primer used for PCR amplification contains a primer site compatible for sequencing.

[0076]

[51] The method of claim 50, wherein the primer site in each primer used in the PCR amplification comprises a sequencing primer binding sequence, preferably an NGS primer binding sequence suitable for sequencing.

[0077]

[52] In step d), first, d1) removing blocking adaptors, affinity adaptors, and any sequencing adaptor sequences from the sequence reads to obtain trimmed sequence reads; and d2) mapping the trimmed sequence reads to a reference genome to obtain aligned sequence reads; The method of any preceding claim, further comprising aligning the sequence reads to a reference genome to obtain aligned sequence reads.

[0078]

[53] The method according to any one of the preceding claims, wherein the genomic DNA is human genomic DNA and the reference genome is a human reference genome selected from, for example, hg16, hg17, hg1, hg19, and hg38, preferably hg19 or hg38.

[0079]

[54] The method of any one of the preceding claims, wherein determining the cleavage site is not based on sequence homology to an on-target cleavage site of the one or more target-specific programmable nucleases.

[0080]

[55] The method of any one of the preceding claims, wherein the putative cleavage site consists of a short DNA sequence defined at a specific site in the reference genome.

[0081]

[56] The method of any one of the preceding claims, wherein the putative cleavage sites have sequence lengths that are set so that there is no difference between different putative cleavage sites.

[0082]

[57] The method of any one of the preceding claims, wherein the cleavage site consists of a short DNA sequence defined at a specific site in the reference genome.

[0083]

[58] The method of any one of the preceding claims, wherein the cleavage sites have sequence lengths that are set so that there is no difference between different cleavage sites.

[0084]

[59] The method of claim 56 or 58, wherein the putative cleavage site and / or cleavage site has a set sequence length, preferably a sequence length of 4 nt.

[0085]

[60] The method according to any one of the preceding claims, wherein in step e), each sequence region is a group of sequences having a defined sequence length, preferably about 20 to 1000 nt, more preferably about 50 to 500 nt, even more preferably at least about 50 nt, even more preferably at least about 80 nt, even more preferably at least about 100 nt.

[0086]

[61] The method according to claim 60, wherein in step e), each sequence region is a group of sequences of a defined sequence length, preferably of about 100 nt or a multiple thereof, more preferably of a sequence length of 100 nt or a multiple thereof.

[0087]

[62] The method of any one of the preceding claims, wherein in step e) the putative cleavage sites are determined as specific sites in sequence regions where the aligned sequence read starts and ends coincide, and also as sites within these sequence regions where a local maximum of the sum of the aligned sequence read starts and ends occurs.

[0088]

[63] The method of any one of the preceding claims, wherein in step e), in each sequence region with non-zero read start and read end coverage, one or more putative cleavage sites are determined as sites where aligned sequence read start and sequence read end portions coincide and the sum of the number of sequence read start and sequence read end portions has a local maximum.

[0089]

[64] In step e), the putative cleavage site is - dividing the reference genome (the DNA sequence of the reference genome) containing the aligned sequence reads into sequence groups of defined length, preferably at least about 100 nt; and - for each group of sequences comprising at least one sequence read start and at least one sequence read end, determining one or more putative cleavage sites as those with a local maximum of the sum of the sequence read start and sequence read end, and optionally minimizing the distance to the center of the group of sequences. (Here, the sequence read start portion consists of the first 1 to 5 nt, preferably the first 2 to 4 nt, more preferably the first 3 nt of the aligned sequence read, and the sequence read end portion consists of the last 1 to 5 nt, preferably the last 2 to 4 nt, more preferably the last 3 nt.) The method according to any one of the preceding claims, wherein the

[0090]

[65] In step e), the cleavage site is e1) dividing the (DNA sequence of) the reference genome containing the aligned sequence reads into sequence groups of set sequence lengths, preferably about 100 nt in length; e2) preferably measuring read coverage for each set of sequences, discarding sequences with zero read coverage, and preferably combining sequences with consecutive read coverage >0 into a consecutive set of sequences based on sequence order in the reference genome; e3) dividing each set of sequences into windows of a set width, preferably 4 nt, with a set step size less than or equal to the window width, preferably 2 nt; e4) determining the beginning, end, and background coverage of each window by grouping as follows: - grouping sequence read start portions falling within said window region into start coverage groups; - grouping sequence read ends falling within said window regions into end-covering groups; - grouping sequence reads that fall within the window region but do not belong to the start and end coverage groups into a background coverage group; For each window, add up the number of reads in the start coverage group, the end coverage group, and the background coverage group; e5) for each group of sequences that includes at least one sequence read start and at least one sequence read end, determining one or more putative cleavage sites by maximizing the sum of start and end coverage and minimizing the distance to the center of the group of sequences; and e6) Eliminating false positive cleavage sites (Here, the sequence read start portion preferably consists of the first 1 to 5 nt, preferably the first 2 to 4 nt, more preferably the first 3 nt of the aligned sequence read, and the sequence read end portion preferably consists of the last 1 to 5 nt, preferably the last 2 to 4 nt, more preferably the last 3 nt.) 2. The method of any one of the preceding claims, wherein the antibody is identified by

[0091]

[66] The method according to item 64 or 65, wherein the width of the group of sequences is about 20 to 1000 nt, preferably about 50 to 500 nt, more preferably at least about 50 nt, even more preferably at least about 80 nt, and most preferably at least about 100 nt.

[0092]

[67] The method according to any one of items 64 to 66, wherein in step e), each sequence region is a group of sequences of a defined sequence length, preferably of about 100 nt or a multiple thereof, more preferably of a sequence length of 100 nt or a multiple thereof.

[0093]

[68] Step e1) comprises dividing the reference genome (the DNA sequence of the reference genome) containing the aligned sequence reads into a set of sequences, each having a length of about 100 nt; and Step e2) comprises, for each set of sequences of step e1), measuring the coverage of reads, discarding sequences with zero coverage of reads, and preferably combining sequences with consecutive coverage of reads >0 into a consecutive set of sequences based on the sequence order of the reference genome. 68. The method according to any one of items 65 to 67, wherein in the contiguous set of sequences, preferably each sequence group consists of one or more sequences of step e1).

[0094]

[69] The method according to any one of claims 65 to 68, wherein in the consecutive set of sequence groups of step e2), the width of each sequence group is defined by the sum of the consecutive sequence groups of step e1) between the two discarded sequence groups of step e2) that are closest to each other in the sequence of the reference genome, such that each sequence group of e2) starts with the sequence of the reference genome having the sequence group of step e1), immediately follows the first sequence group discarded in step e2), and ends with the sequence group of step e1) that is immediately before the next sequence group discarded in step e2) in the sequence of the reference genome.

[0095]

[70] Step e4) also determines the start, end and background coverage of each sequence group by grouping as follows: - Grouping the start of sequence reads that fall into sequence cluster regions into start coverage clusters; - Grouping the ends of sequence reads that fall into sequence cluster regions into end-covering clusters; - grouping sequence reads that fall into a sequence group region but do not belong to the start and end coverage groups into a background coverage group; 67. A method according to any one of claims 65 to 66, comprising adding up, for each sequence group, the number of read data in the start covering group, the end covering group, and the background covering group, respectively.

[0096]

[71] The sequence read start portion is composed of nt no. 1 to nt no. 5 or nt no. 2 to nt no. 6, preferably nt no. 1 to nt no. 4 or nt no. 2 to nt no. 5, more preferably nt no. 1 to nt no. 3 or nt no. 2 to nt no. 4 of the aligned sequence read end portion; The method according to any one of items 63 to 70, wherein the sequence read end portion preferably consists of nt no.1 to nt no.5 or nt no.2 to nt no.6, preferably nt no.1 to nt no.4 or nt no.2 to nt no.5, more preferably nt no.1 to nt no.3 or nt no.2 to nt no.4 of the aligned sequence read end portion, such that any sequence read end portion of the aligned sequence reads aligned to the same contiguous sequence of the reference genome points in the opposite direction to any sequence read start portion of the aligned sequence reads.

[0097]

[72] The method according to any one of the preceding claims, wherein the putative cleavage sites determined in step e) consist of a specific window of set sequence width, preferably a window of (approximately) 4 nt.

[0098]

[73] The method according to any one of claims 64 to 72, wherein the putative cleavage site measured in step e) corresponds to a specific window of nt of set width in step e3), preferably a window of nt of 4 nt.

[0099]

[74] The method of any preceding claim, wherein false positive cleavage sites are eliminated based on the ratio of background coverage to signal coverage at the putative cleavage site, and / or the ratio of starting coverage to end coverage at the putative cleavage site, and / or site coverage, and / or sequence group coverage.

[0100]

[75] 1. The ratio of background coverage to signal coverage does not exceed an acceptable limit, preferably about 0.5; and / or 2. the ratio of initiating coating to end coating at the putative cleavage site falls within a certain numerical range, preferably within about 1 / 5 to 5 / 1; and / or 3. A minimum site coverage, preferably 6 or greater; and / or 4. The sequence coverage is minimal, preferably 10 or more; 10. The method of any preceding claim, wherein false positive cleavage sites are excluded by selecting only putative cleavage sites.

[0101]

[76] False positive cleavage sites 1. The ratio of signal coverage to background coverage at the putative cleavage site is less than or equal to 0.5; and 2. The ratio of initiating coating to end coating at the putative cleavage site is between 1 / 5 and 5; and 3. Site coverage is 6 or more; and 4. The sequence coverage is 10 or more. 2. The method of any preceding claim, wherein only putative cleavage sites are excluded by selection.

[0102]

[77] The method of any one of the preceding claims, wherein false positive cleavage sites are preferably excluded or further rejected by using a machine learning model trained on false positive sites excluded by the method of any one of claims 61 to 63, preferably using as features the ratio of background coverage to signal coverage, the ratio of starting coverage to end coverage, the width of the sequence population, the relative distance between false positive sites, and / or the ratio of false positive sites to sequence population.

[0103]

[78] The method of claim 77, wherein a random forest model is constructed, for example, by obtaining the area under curve 1 using 3-fold 10 replicate cross-validation.

[0104]

[79] The method of any preceding claim, wherein a non-specific cleavage site is identified as a cleavage site in step e) that is not a specific cleavage site of the one or more target-specific programmable nucleases.

[0105]

[80] i) providing a first sample comprising genomic DNA of cells that have been treated with one or more target-specific programmable nucleases, and a second sample comprising genomic DNA of cells of a related genetic background, preferably cells having the same genetic background and that have not been treated with said one or more target-specific programmable nucleases; ii) detecting in vitro the non-specific sites in each of the two samples by using one or more target-specific programmable nucleases of step i) in the method of any one of claims 1 to 79; and iii) comparing the non-specific sites of the two samples to identify non-specific sites identified only in the sample comprising genomic DNA of cells not treated with the one or more target-specific programmable nucleases, but not in the sample comprising genomic DNA of cells treated with the one or more target-specific programmable nucleases, as in vivo non-specific sites of the one or more target-specific programmable nucleases. A method for detecting a non-specific site in a genome in vivo, comprising:

[0106]

[81] The method of claim 80, wherein the cells of related genetic background are cells of the same genetic background, preferably cells having the same genetic code, preferably cells derived from the same tissue.

[0107]

[82] The method of claim 80 or 81, wherein the cells of related genetic background are cells of the same cell lineage.

[0108]

[83] The method according to any one of claims 80 to 82, wherein the cells used in both samples are eukaryotic cells, e.g. human cells.

[0109]

[84] A kit for carrying out the method of any one of claims 1 to 79 or any one of claims 80 to 83. [Brief description of the drawings]

[0110] [Figure 1] Overview of the Abnoba-Seq workflow: Genomic DNA was randomly sheared to a median fragment size of approximately 400 bp, end-protected by ligation of blocking adapters (grey), and unprotected fragments were removed by exonuclease treatment. End-blocked DNA fragments were then cleaved with CRISPR-Cas9, and the genome was successfully cleaved by PCR amplification of specific sites as in-step controls (IPCs). The cleaved fragments were tagged by ligation of biotin-adapters (light grey) enriched on streptavidin (StA) beads, and then amplified by PCR as follows: first, linear PCR on StA beads with a single primer that binds to the biotin-adapter, then exponential PCR after adding a second primer that binds to the blocking adapter, and finally, libraries were produced and subjected to next-generation sequencing (NGS).

[0111] [Diagram 2]End-blocking test of DNA fragments Ligation of blocking adapters to DNA fragments (PCR amplicons), followed by exonuclease treatment and subsequent capillary electrophoresis are shown above, and the percentage of end-blocked DNA fragments (PCR amplicons) is shown in the table below these graphs. A: Exonuclease treatment: Ligation of blocking adapters (75 pmol) to 189 bp PCR amplicons (400 ng), followed by digestion for 30 min with a combination of exonucleases in a two-step reaction with Exonuclease III (200 U), Lambda Exonuclease (10 U), and Exonuclease I (20 U). B: Titration of blocking adapter concentration: Ligation of different concentrations of blocking adapters to 189 bp PCR amplicons (400 ng). PCR amplicons alone, those ligated to a blocking adaptor at one end, and those ligated to blocking adaptors at both ends are represented by a 189 bp peak, a 215 bp peak, or a 241 bp peak, respectively, with a 225 bp peak in between, and the 15 bp and 5,000 bp peaks (first and last peaks), fragments of the alignment markers; percentages were calculated as the mean AUC.

[0112] [Diagram 3] Intra-step control A: U20S cells. PCR to amplify genome-specific sites after in vitro cleavage of end-blocked genomic DNA with CRISPR-Cas nucleases targeting RNF2, FANCF, VEGFA, and HEKs4. B: HSCs. PCR to amplify genome-specific sites after in vitro cleavage of end-blocked genomic DNA with CRISPR-Cas nuclease targeting VEGFA. +RNP, cleaved sample; -RNP, untreated sample; t, position of expected PCR amplification product.

[0113] [Figure 4]Bioinformatic Pipeline To identify target sites, the bioinformatic pipeline was performed in the following steps: (1) calculate sequence population coverage, (2) discard empty sequence populations, (3) divide sequence populations into a sliding window of 4 nt with 2 nt steps, (4) identify putative cleavage sites that have read start or read end at the same site, and (5) filter out false positive sites by applying the following filter ratios: background coverage to signal coverage ≤ 0.5, start coverage to end coverage ratio ≥ 1 / 5 to ≤ 5, site coverage (sum of start and end reads) ≥ 6, and sequence population coverage ≥ 10.

[0114] [Diagram 5] Alignment patterns of NGS reads at specific sites. NGS reads were aligned to the human reference genome (hg19) and visualized with the Integrative Genomics Genome Viewer (IGV). The chromosomal location of each target site is indicated by a bar. Single reads are represented as arrows with their orientation indicated. Annotated genes (exons as boxes, introns as lines) are shown at the bottom. A: Genomic DNA of U20S cells. B: Genomic DNA of HSCs.

[0115] [Figure 6] (Figure 6-1, Figure 6-2) Top 20 Target Sites of Tested Targets Alignment showing the sequences of the top 20 identified target sites, including the specific sites (*) identified for CRISPR-Cas nucleases targeting FANCF, RNF2, VEGFA, and HEKs4 (Figure 6-1, Figure 6-2). The coverage (number of reads) is shown next to the sequence. The top row shows the sequence of the specific site including the PAM (NGG). In the bottom row, the difference from each specific site is highlighted in gray. Positive and negative numbers indicate bulges in the DNA or gRNA, respectively. N is any nucleotide.

[0116] [Figure 7]Comparison of Abnoba-Seq with GUIDE-Seq and CIRCLE-Seq. Venn diagram showing the overlap of identified target sites between Abnoba-Seq, GUIDE-Seq, and CIRCLE-Seq.

[0117] [Figure 8] (Figure 8-1, Figure 8-2) Abnoba-Seq applied to human hematopoietic stem cells. A: Top 20 (non)specific sites. Alignment showing the sequences of the top 20 identified target sites, including the specific sites (*) identified for CRISPR-Cas nucleases targeting VEGFA. Coverage (number of reads) is shown next to the sequences. The top row shows the sequences of the specific sites including PAM. In the bottom row, the differences from the specific sites are highlighted in gray. Positive or negative numbers indicate bulges in the DNA or gRNA, respectively. N: any nucleotide. B: Shared and distinct non-specific sites. Venn diagram showing the overlap of identified non-specific sites between U20S cells and HSC (Figure 8-1). C, D: Qualitative comparison. Venn diagram showing the overlap of the top 20 or top 50 identified target sites between U20S and HSC (Figure 8-2).

[0118] [Figure 9] Targeted amplicon sequencing of the top 20 non-specific sites. U20S cells and HSCs were transduced with RNPs targeting VEGFA or HEKs4. Editing was assessed by targeted amplicon sequencing of the specific and top 20 non-specific sites. A: Editing efficiency with RNPs targeting VEGFA. For each (non-)specific site (y-axis), the bars indicate the percentage of edited alleles on a logarithmic scale (x-axis) in either U20S cells (black) or HSCs (grey). The chromosomal location of each site is indicated. ON, specific site; OT, non-specific site. B: Editing efficiency with RNPs targeting HEKs4. For each (non-)specific site (y-axis), the bars indicate the percentage of edited alleles on a logarithmic scale (x-axis) in HSCs (grey). The chromosomal location of each site is indicated on the right. ON, specific site; OT, non-specific site; nd, not determined due to sequencing failure. Detailed Description of the Invention

[0119] Abnoba-Seq In a first aspect, the present invention relates to a method for detecting non-specific sites of one or more target-specific programmable nucleases in a genome in vitro, which allows to detect all cleavage sites of one or more target-specific programmable nucleases in a genome in vitro.Unless otherwise limited, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art.

[0120] The method of the first aspect of the present invention has several advantages compared to known methods for determining the non-specific site of target-specific programmable nuclease.The method does not use homology with specific site / sequence to identify non-specific site, as used by known in vitro methods, for example, and in particular, similarity with actual target sequence does not play a role in detecting non-specific site, and the method in this respect is completely unbiased.This results in a method with comparable sensitivity but higher selectivity compared to other known in vitro methods.

[0121] Advantageously, the method of the present invention can be carried out with a small amount of input genomic DNA.This allows the testing of one or more target-specific programmable nucleases on precious cell samples with limited sample volume and availability, for example, by measuring the specificity of target-specific programmable nucleases in primary cells, biopsy material, or gene-edited grafts.In addition, it does not require expensive high-performance sequencers such as HiSeq, but can be carried out using basic (low-cost) NGS sequencers.

[0122] The term "target-specific programmable nuclease" is known to those skilled in the art. Preferably, the term "target-specific programmable nuclease" used in the context of the present invention refers to all forms of nucleases that can recognize and cleave a specific site on a desired genome. In particular, this can include, but is not limited to, ZFN, TALEN, meganuclease, Argonaute-family protein-based DNA-guided DNA nuclease, and CRISPR / Cas RGN. The terms "CRISPR / Cas RGN", "ZFN", and "TALEN" can refer to one nuclease or more than one nuclease of that type, for example, CRISPR / Cas RGN can refer to one CRISPR / CasRNA-guided nuclease or more than one (two, three, etc.) CRISPR / CasRNA-guided nuclease, and similarly, ZFN can refer to one Zn finger nuclease or more than one Zn finger nuclease, etc. In a preferred embodiment, the target-specific programmable nuclease is a nuclease that specifically cleaves a target in double-stranded DNA (dsDNA), i.e., it targets a specific target site or sequence (at a specific site) for cleavage, and preferably introduces a dsDNA break in the DNA.

[0123] The term "target site" as used in the context of the present invention is known to those skilled in the art and can preferably refer to specific and non-specific sites. The term "specific site" as used in the context of the present invention is known to those skilled in the art. It can preferably refer to a specific site or sequence that is targeted by a target-specific programmable nuclease for nucleic acid cleavage, preferably DNA cleavage, and can be freely selected within a nucleotide sequence, preferably a DNA sequence such as a genomic DNA sequence, depending on the purpose.

[0124] Of course, as will be understood by those skilled in the art, a target-specific programmable nuclease may target a nucleic acid, preferably DNA such as genomic DNA, for cleavage at one or more sites other than the specific site in addition to the specific site. These other sites are non-specific sites. The term "non-specific site" used in the context of the present invention is known to those skilled in the art. Preferably, the term "non-specific site" refers to a site that is not identical in the nucleic acid sequence to the target site or sequence (specific site) of the target-specific programmable nuclease. That is, it can refer to a site other than the specific site that is targeted for cleavage by the target-specific programmable nuclease.

[0125] Preferably, the non-specific site is not identical in nucleic acid sequence to the specific site, but is a site targeted for cleavage by the target-specific programmable nuclease in vitro. In certain embodiments, the non-specific site is not identical in nucleic acid sequence to the specific site, but is a site targeted for cleavage by the target-specific programmable nuclease in vitro and in vivo (i.e., in a cell). By targeted cleavage, it is preferably meant that the target-specific programmable nuclease directly or indirectly binds to a nucleic acid, e.g., DNA, using a specific site or sequence, and then cleaves the nucleic acid at or near the specific site or sequence.

[0126] Step a) Preparation of purified end-blocked dsDNA fragments The method of the first aspect of the invention comprises the steps of (a) providing an input sample comprising genomic DNA, randomly fragmenting the genomic DNA into dsDNA fragments, ligating blocking adaptors to the ends of the dsDNA fragments, removing dsDNA fragments containing unblocked ends, and preparing purified end-blocked dsDNA fragments.

[0127] The genomic DNA in the input sample is not particularly limited, as long as it is suitable for use in the method of the present invention. Preferably, the genomic DNA is isolated genomic DNA, which means that the genomic DNA is extracted from cells and preferably purified. That is, it is separated to some extent, preferably as far as possible, from proteins and other cellular components and / or reagents used during extraction. The source of the genomic DNA is not particularly limited. The genomic DNA may be derived from cells of a particular organism, for example, from a single or multicellular organism, from a multicellular organism from cells of a particular tissue, or from a particular cell type. In one embodiment, the genomic DNA is at least one derived from a eukaryotic cell (e.g., an animal, a mammal, an insect, a plant, a fungus, an insect, a bird, a fish, an amphibious organism, a worm, or a cnidarian).

[0128] In a preferred embodiment, the genomic DNA contained in the input sample provided in step (a) is human genomic DNA, for example from primary cells or biopsy material. Such cells may or may not be cells of a particular tissue or cells of a particular cell type of a particular tissue.

[0129] The amount of genomic DNA in the input sample is not particularly limited. In a preferred embodiment of the present invention, the input sample of genomic DNA contains less than 1 μg, preferably less than 900 ng, more preferably less than 800 ng, even more preferably less than about 700 ng, even more preferably about 650 ng or less, even more preferably about 600 ng or less, or preferably about 500 ng, or preferably about 400 ng of genomic DNA.

[0130] The term "about" as used herein in conjunction with a value or range indicates that a given quantity value or range can include an amount within 10% of the stated value or range, or optionally within 5% of the value or range, or in some embodiments, within 1% of the value or range. Of course, the term "about" also includes the exact value or exact range of the value or range with which it is combined.

[0131] According to the present invention, the method of randomly fragmenting genomic DNA into dsDNA fragments is not particularly limited. Methods of randomly fragmenting genomic DNA are known in the art. Genomic DNA can be randomly fragmented, for example, by shearing, i.e., mechanically breaking DNA, or by enzymatic fragmentation. According to one embodiment of the present invention, genomic DNA is fragmented by ultrasonication (sonication), acoustic shearing, centrifugal shearing, point sink shearing, needle shearing, or enzyme-based fragmentation, preferably by ultrasonication. Preferably, genomic DNA is fragmented to a defined length (median or average). The specific length of genomic DNA to be fragmented into dsDNA fragments is not limited. In a particular embodiment of the first aspect of the invention, the genomic DNA is randomly fragmented into dsDNA fragments having a median fragment size of about 300-600 bp, preferably 200-800 bp, more preferably about 300-600 bp, more preferably about 400-500 bp, even more preferably about 400 bp. In a preferred embodiment of the invention, the genomic DNA is randomly fragmented into dsDNA fragments having a median fragment size of about 300-600 bp, more preferably 300-600 bp, more preferably about 400 bp, even more preferably about 400 bp. The term "bp" refers to one or more base pairs, i.e., one or more pairs of complementary nucleotides in a double-stranded (ds) nucleic acid such as dsDNA or dsRNA. Genomic DNA in the input sample is randomly fragmented to obtain dsDNA fragments, and then blocking adapters are ligated to the ends of the dsDNA fragments and dsDNA fragments containing unblocked ends are removed to prepare purified end-blocked dsDNA fragments.

[0132] The term "blocking adapter" as used herein refers to a molecule, preferably a molecule comprising dsDNA, that blocks further ligation, as well as exonuclease cleavage (digestion) from an end of a dsDNA fragment after the blocking adapter has been ligated to that end. The term "end-blocked" in reference to a dsDNA fragment refers to a dsDNA fragment having a blocking adapter ligated to both ends.

[0133] The blocking adaptor is not particularly limited as long as it fulfills the above-mentioned role. Methods for preparing blocking adaptors are known in the art. That is, after ligation, it blocks further ligation and exonuclease cleavage from the end of the DNA fragment to which it is ligated. Preferably, the blocking adaptor is a DNA-based adaptor, more preferably a dsDNA-based adaptor (i.e., a dsDNA-based blocking adaptor), which means that it contains dsDNA. Such dsDNA-based blocking adaptors can contain other modifications, including modified nucleotides and other molecules that are not nucleotides. Such dsDNA-based blocking adaptors preferably have a sequence length that allows a primer site to be incorporated into the blocking adaptor for PCR amplification.

[0134] According to a preferred embodiment of the invention, the blocking adaptor, preferably a dsDNA-based blocking adaptor, comprises one or more modifications that prevent ligation and one or more modifications of the end where the blocking adaptor is ligated to the dsDNA fragment after ligation of the blocking adaptor to the dsDNA fragment, and that prevent exonuclease cleavage (digestion) of the dsDNA fragment from the end where the blocking adaptor is ligated to the dsDNA fragment after ligation of the blocking adaptor to the dsDNA fragment.

[0135] As is known in the art, modification of the blocking adapter after ligation of the blocking adapter to a dsDNA fragment prevents ligation of the end at which the blocking adapter is ligated to the dsDNA fragment, and dsDNA-based blocking adapters preferably include one or more C3-spacers, one or more dideoxy-G / dideoxy-C, or other spacers, preferably on each strand of the dsDNA-based blocking adapter, and preferably at the end of the dsDNA-based blocking adapter that is not used to ligate the dsDNA-based blocking adapter to the dsDNA fragment (i.e., the end of the blocking adapter that is not directly adjacent to the end of the dsDNA fragment to which the blocking adapter is ligated after ligation).

[0136] Modifications of the blocking adapter after ligation to the dsDNA fragment are also known in the art to prevent exonuclease cleavage of the dsDNA fragment from the end where the blocking adapter is ligated to the dsDNA fragment. dsDNA-based blocking adapters include one or more phosphorothioate linkages, one or more C3-spacers, and combinations thereof. In a preferred embodiment, the blocking adaptor is a dsDNA based blocking adaptor that includes one or more, preferably two or more, preferably at least two (e.g., two, three, four, etc.) phosphorothioate linkages on each strand of the DSDNA based blocking adaptor, preferably near the ends, e.g., among the last 3-5 nucleotides of the DSDNA based blocking adaptor that are not used to ligate the DSDNA based blocking adaptor to a DSDNA fragment (i.e., the end of the blocking adaptor that is not directly adjacent to the end of the dsDNA fragment to which the blocking adaptor is ligated after ligation), and one or more C3 spacers, preferably on each strand of the DSDNA based blocking adaptor, preferably at least the end of the DSDNA based blocking adaptor that is not used to ligate the DSDNA based blocking adaptor to a DSDNA fragment.

[0137] The dsDNA-based form of the blocking adapter may further comprise modifications, including, for example, one or more internal deoxyuridines, preferably at least two on each strand, preferably adjacent to the end of the adapter that is not used to link the dsDNA-based blocking adapter to a dsDNA fragment. A preferred blocking adapter is a dsDNA-based blocking adapter that comprises one or more C3-spacers in combination with one or more phosphorothioate linkages, and optionally one or more internal deoxyuridines, at least one on each strand of the dsDNA-based blocking adapter, preferably at the end of the dsDNA-based blocking adapter that is not used to link the dsDNA-based blocking adapter to a dsDNA fragment.

[0138] Furthermore, the blocking adapter preferably comprises a primer site suitable for use in PCR priming, meaning a nucleotide sequence at which a PCR primer can be annealed to amplify the dsDNA fragment to which the blocking adapter is ligated. Preferably, the sequence of the dsDNA-based blocking adapter and / or the primer site in the blocking adapter is not present in the target genome (i.e. the sequence of the genomic DNA in the input sample and / or the reference genome). In a particular embodiment of the invention, the blocking adapter is a dsDNA-based blocking adapter comprising the forward sequence defined in SEQ ID NO:01 and the reverse sequence of SEQ ID NO:02. In a further particular embodiment of the invention, the blocking adapter is a dsDNA-based blocking adapter consisting of the sequences of SEQ ID NO:01 and SEQ ID NO:02.

[0139] The blocking adaptor is ligated to the end of the dsDNA fragment by methods known in the art. For example, the blocking adaptor can be ligated to the end of the dsDNA fragment by enzymatic or chemical ligation using ligation protocols known in the art. Enzymatic ligation can be performed using known DNA ligases, which facilitate the joining of DNA strands by catalyzing the formation of phosphodiester bonds, including blunt-end and sticky-end ligation. In a preferred embodiment, the blocking adaptor is ligated to the end of the dsDNA fragment by enzymatic ligation.

[0140] Preferably, ligating a blocking adaptor to the end of a dsDNA fragment comprises: 1. preparing a dsDNA fragment for blocking adaptor ligation; and 2. ligating a blocking adaptor to the end of the dsDNA fragment to obtain an end-blocked dsDNA fragment. Methods for preparing a dsDNA fragment for blocking adaptor ligation are known to those skilled in the art. Preparing a dsDNA fragment for blocking adaptor ligation can include one or both of end repair and dA tailing (incorporation of non-templated deoxyadenosine 5'-monophosphate (dAMP) into the 3' end of a blunt-ended DNA fragment).

[0141] In a specific embodiment of the present invention, a dsDNA fragment is prepared for blocking adapter ligation by end repair, and a dsDNA fragment is prepared that does not have an overhang and contains a 5' phosphate group and a 3' hydroxyl group, and a dA tailing of the blunt-ended DNA fragment thus prepared is prepared. A dsDNA-based blocking adapter having a dT overhang at the 3' end of one end of the adapter can be ligated to the end of the dsDNA fragment thus prepared.

[0142] After the blocking adaptors are ligated to the ends of the dsDNA fragments, the dsDNA fragments containing unblocked ends are removed to prepare purified end-blocked dsDNA fragments. In the dsDNA fragments with unblocked ends, one or both ends are not ligated to a blocking adaptor. Methods for removing such dsDNA fragments with unblocked ends are known to those skilled in the art.

[0143] In a preferred embodiment, removing dsDNA fragments comprising unblocked ends comprises contacting the sample with one or more exonucleases (after ligation of blocking adaptors to the dsDNA fragments) to remove the dsDNA fragments and obtain purified, end-blocked DNA fragments. Suitable exonucleases are known in the art and include, for example, one or more of exonuclease III, exonuclease V, exonuclease VIII, lambda exonuclease, T5 exonuclease, T7 exonuclease, nuclease BAL-31, and combinations thereof, preferably exonuclease III, lambda exonuclease, or combinations thereof. In some cases, one or more single-stranded DNA (ssDNA) exonucleases are used, for example, selected from exonuclease I, thermostable exonuclease I, exonuclease T, exonuclease VII, RecJf, mung bean nuclease, nuclease P1, and combinations thereof.

[0144] In certain embodiments, the one or more exonucleases consist of a mixture of Exonuclease I, Exonuclease III, and Lambda Exonuclease, for example, in an enzyme unit ratio of 2:20:5.

[0145] In a preferred embodiment of the method of the first aspect of the invention, step a) comprises the steps of: a1) randomly fragmenting genomic DNA to defined lengths to provide dsDNA fragments; a2) protecting the ends of the dsDNA fragments by ligating blocking adaptors to obtain end-blocked dsDNA fragments; and a3) contacting the sample containing the end-blocked dsDNA fragments with one or more exonucleases to remove unblocked dsDNA fragments and obtain purified end-blocked DNA fragments.

[0146] In another particular embodiment of the method of the first aspect of the invention, in step a), preferably after contacting the sample containing end-blocked dsDNA fragments with one or more exonucleases to remove the dsDNA fragments, the free ends of the remaining unblocked DNA fragments are dephosphorylated with a dephosphorylation reagent, in which embodiment, preferably in step a), said dephosphorylation reagent consists of one or more phosphatases, preferably comprising or consisting of one or more phosphatases selected from the group consisting of antacid phosphatase, calf intestinal alkaline phosphatase, Quick-CIP, shrimp alkaline phosphatase, and combinations thereof.

[0147] Step b) Preparation of a DNA library of cleaved and amplified dsDNA fragments The method of the first aspect of the present invention comprises a second step, namely (b) of preparing a DNA library of cleaved and amplified dsDNA fragments by: b1) cleaving the purified end-blocked dsDNA fragments with one or more target-specific programmable nucleases to obtain cleaved dsDNA fragments, b2) ligating affinity adapters to the cleavage sites of the cleaved dsDNA fragments to obtain affinity adapter-modified dsDNA fragments, b3) enriching the affinity adapter-modified dsDNA fragments using the affinity adapters to obtain enriched affinity adapter-modified dsDNA fragments, and b4) amplifying the enriched affinity adapter-modified dsDNA fragments by PCR amplification to obtain a DNA library of cleaved and amplified dsDNA fragments.

[0148] According to the present invention, the amount of purified end-blocked dsDNA fragments in step b1) is not particularly limited. According to a particular embodiment, in step b1), less than 400 ng, preferably less than 300 ng, more preferably less than 200 ng, even more preferably less than 100 ng, for example about 90 ng, about 80 ng, about 70 ng, about 60 ng, less than 60 ng, about 50 ng, or 50 ng of purified end-blocked dsDNA fragments, alternatively less than 50 ng, for example 40 ng of purified end-blocked dsDNA fragments, are cleaved with one or more target-specific programmable nucleases to obtain cleaved dsDNA fragments. In a preferred embodiment, less than 100 ng, preferably about 50 ng, of purified end-blocked dsDNA fragments are used in step b1).

[0149] According to another particular embodiment of the invention, in step b1), purified endblocked dsDNA fragments are used which correspond to less than 1,500,000 monoploid genomes, preferably less than 1,500,000 monoploid genomes, more preferably less than 80,000 monoploid genomes, even more preferably less than 60,000 monoploid genomes, even more preferably less than 50,000 monoploid genomes, such as about 45,000, about 30,000, about 20,000 or about 15,000 monoploid genomes.

[0150] The purified end-blocked dsDNA fragments can be cleaved with one or more target-specific programmable nucleases to obtain cleaved dsDNA fragments by contacting a sample containing the purified end-blocked dsDNA fragments with one or more target-specific programmable nucleases.

[0151] The one or more target-specific programmable nucleases used in step b1) are not particularly limited, as long as they can be used in the method of the present invention, i.e., are suitable for use as target-specific programmable nucleases for target-specific cleavage of genomic DNA, e.g., for genome editing, and preferably for detection of non-specific sites in vitro and in vivo using the method of the present invention. Suitable target-specific programmable nucleases include TALEN, ZFN, Meganuclease, MegaTAL, Fokl-dCas9, Argonaute-family protein-based DNA nuclease, CRISPR / Cas RGN, or mutants thereof, or combinations thereof.

[0152] The TALEN that can be used in the method of the present invention is not particularly limited as long as it can be used in the method of the present invention. TALEN refers to one or more fusion proteins containing a DNA binding domain and a nucleotide cleavage domain, preferably a Fokl cleavage domain, derived from a transcription activator-like effector (TALE) protein. TALE proteins contain multiple 33-35 amino acid repeat domains (TALE repeat arrays), each of which recognizes a single base pair (see, for example, Gaj et al., 2013, Trends Biotechnol. 31(7):397-405). Methods for producing engineered TALE repeat arrays and fusing them to nucleases are known in the art (e.g., Reyon et al., 2012, Nature Biotechnology 30, 460-465; Bogdanove and Voytas, 2011, Science 333, 1843-1846; Bogdanove et al., 2010, Curr Opin Plant Biol 13, 394-401; Scholze and Boch, 2011, J. Curr Opin Microbiol 2011; Boch et al., 2009, Science 326, 1509-1512; Cermak et al., 2011, Nucleic Acids Res 39,e82; Tesson et al., 2011, Nat Biotechnol 29,695-696; Sander et al., 2011, Nat Biotechnol 29, 697-698; and Zhang et al., 2011, Nat Biotechnol 29, 149-153).

[0153] The meganuclease that can be used in the method of the present invention is not particularly limited, as long as it can be used in the method of the present invention. Meganucleases are sequence-specific endonucleases derived from various organisms such as bacteria, yeast, algae, and plant organelles. Endogenous meganucleases have been described that have customized DNA binding sites with recognition sites of about 12-30 bp; meganuclease recognition sites of 18 bp and 24 bp in length, and can be used in the method (see, for example, Silva et al., 2011, Current Gene Therapy, 11:11-27; Arnuld et al., 2011, Protein Engineering Design & Selection, 24:27-31). Meganucleases can be selected, for example, from meganucleases of the LAGLIDADG family, the GIY-YIG family, the His-Cyst box family, and the HNH family. Representative meganucleases include l-Scel, l-Ceul, Pl-Pspl, Pl-Scel, l-Scelv, l-Csml, l-Panl, l-Scell, l-Ppol, l-Scelll, l-Crel, l-Tevl, l-Tevll, and l-Tevll.

[0154] The ZFN that can be used in the method of the present invention is not particularly limited, so long as it is suitable for use in the method of the present invention. ZFN is an artificially engineered nuclease that includes a Zn-finger DNA binding domain and a DNA cleavage domain. The DNA binding properties of individual Zn-finger DNA binding domains can be artificially engineered by randomizing amino acids at α-helical positions involved in DNA binding and by using selection methodologies such as phage display to identify desired mutants that can bind to the desired DNA target site (Rebar et al., 1994, Science, 263:671; Choo et al., 1994 Proc Natl Acad Sci. USA, 91:1163; Jamieson et al., 1994, Biochemistry 33:5689; Wu et al., 1995 Proc Natl. Acad. Sci. USA, 92:344). Such recombinant zinc-finger DNA-binding domains can be fused to functional domains, such as nucleases, to introduce targeted modifications into the genomes of model organisms, plants and human cells (Carroll, 2008, Gene Ther., 15:1463-68, Cathomen, 2008, Mol. Ther. 16:1200-07; Wu et al., 2007, Cell. Mol. Life Sci., 64:2933-44).

[0155] The selection of a particular target site / sequence and the design and construction of ZFNs are well known to those of skill in the art and are described in detail, for example, in U.S. Patent Nos. 7,888,121, 8,409,861, 6,479,626, 6,903,185, and 7,153,949. ZFNs can include any combination of suitable linkers between each Zn-fingers of the protein. Zinc-finger DNA binding domains can be engineered to bind to a sequence of choice (Beerli et al., 2002, Nat. Biotechnol., 20:135-141; Segal et al., 2003, Biochemistry, 42:2137-48; Mandell et al., 2006, Nucleic Acids Res., 34:W516-523; Carroll et al., 2006, Nat. Protoc. 1:1329-41; Bae et al., 2003, Nat. Biotechnol., 21:275-280; Wright et al., 2006, Nat.Protoc. 1:1637-52; Maeder et al., 2008, Mol. Cell, 31:294-301; Joung et al., 2010, Nat. Methods, 7:91-92). In a preferred embodiment, ZFNs are prepared as described in WO 2011 / 017293 and WO 2004 / 099366. Further suitable zinc finger DNA binding domains are described in U.S. Patent Nos. 6,511,808, 6,013,453, 6,007,988, and 6,503,717.

[0156] The MegaTAL and Fokl-dCas9 that can be used in the method of the present invention are not particularly limited, as long as they are suitable for use in the method of the present invention. MegaTAL is an artificial fusion protein of meganuclease with TAL effector (see, for example, Boissel et al., 2014, Nucl. Acids Res. 42(4):2591-2601; Boissel and Scharenberg, 2015, Methods Mol Biol. 2015; 1239:171-96). Fokl-dCas9 is a fusion protein that includes inactivated Cas9 and Fokl nuclease for sequence targeting (Guilinger et al., 2014, Nat Biotechnol. Jun;32(6):577-582; WO 2014 / 144288; and WO 2014 / 204578).

[0157] The Argonaute-family protein-based DNA-guided DNA nuclease that can be used in the method of the present invention is not particularly limited, as long as it can be used in the method of the present invention.The method of using the Argonaute-family protein-based DNA-guided DNA nuclease for genome editing is known in the art (for example, see WO 2017139264A1; Wei et al., 2016, Genes & Diseases, Vol.3, Iss. 3, p. 169-170).

[0158] The CRISPR / Cas RNA-guided nuclease (CRISPR / Cas RGN) that can be used in the method of the present invention is not particularly limited, as long as it can be used in the method of the present invention. That is, it means that they are at least suitable for use as a target-specific programmable nuclease for target-specific cleavage of genomic DNA, and are suitable for use in detecting non-specific sites, preferably in vitro and in vivo, using the method of the present invention for genome editing. CRISPR / Cas RGN refers to the combination of Cas9 endonuclease and gRNA. Cas9 endonuclease can bind to gRNA to form Cas9-gRNA ribonucleotide protein (RNP). Methods for preparing CRISPR / Cas RGN specific to a target sequence are well known in the art.

[0159] The Cas endonuclease of the CRISPR / Cas RGN used in the method of the present invention is not particularly limited, as long as it is suitable for use in the CRISPR / Cas RGN for genome editing and suitable for use in the detection of non-specific sites, preferably in vitro and in vivo, using the method of the present invention. In a specific embodiment of the present invention, the Cas endonuclease of the CRISPR / Cas RGN is Cas9 endonuclease or a mutant thereof, or Cpf1 endonuclease or a mutant thereof. The Cas endonuclease can be induced through a simple nucleotide (nt) complementarity of about 20 nucleotides between the engineered gRNA and the complementary strand of the target genomic DNA sequence adjacent to the PAM. As used herein, the term "nt" refers to one or more (e.g., one, two, three, four, etc.) nucleotides of a DNA or RNA molecule.

[0160] Cas9 endonuclease, such as S. pyogenes (hereinafter SpyCas9), can be used (Shen et al., 2013, Cell Res; Jinek et al., 2013, Elife 2, e00471; Hwang et al., 2013, Nat Biotechnol 31, 227-2229(2013); Cong et al., 2013, Science 339, 819-823; Mali et al., 2013, Science 339, 823-826; Cho et al., 2013, Nat Biotechnol 31, 230-223). Engineered CRISPR endonucleases from Prevotella and Francisella 1 Cpf1 (CRISPR-associated endonucleases in Prevotella and Francisella 1) can also be used (Zetsche et al., 2015, Cell 163, 759-771; and Makarova et al., 2015, Nat Rev Microbiol 13, 722-736).

[0161] In some embodiments, the methods of the invention involve the use of wild-type or mutant Cas9 endonuclease from the genera Streptococcus, Neisseria, Pasteurella, Francisella, Campylobacter (e.g., from S. pyogenes, Staphylococcus aureus, Neisseria meningitidis, Streptococcus thermophilus, or Treponema denticola); or from the genera Lachnospira, Butyrivibrio, Peregrinibacteria, Acidominococcus, Porphyromonas, Prevotella, Francisella, Candidatus methanoplasma, and Eubacterium (e.g., Candidatus Paceibacter, Acidaminococcus sp. BV3L6, or Lachnospiraceae bacterium). Wild-type or mutant Cpf1 endonuclease from (from ND2006) is used, either bacterially encoded or codon-optimized for expression in, for example, mammalian cells and / or modified in its PAM recognition specificity and / or its genomic specificity. A number of variants have been described (e.g., WO 2016 / 141224, PCT / US2016 / 049147, Kleinstiver et al., 2016, Nat Biotechnol., 34(8):869-74; Tsai and Jung, 2016, Nat Rev Genet., 17(5):300-12; Kleinstiver et al., 2016, Nature., 529(7587):490-5; Shmakov et al., 2015, Mol Cell., 5;60(3):385-97; and Tsai et al., 2014, Nat Biotechnol., 32(6):569-76).

[0162] The term "gRNA" or "guide RNA" as used herein refers to a target DNA-specific RNA that can bind to Cas endonuclease to form CRISPR / Cas RGN ribonucleotide protein (RNP) and guide Cas protein to a target site / sequence. Any type of gRNA can be used in the methods of the present invention as long as it is suitable for use with CRISPR / Cas RGN for genome editing and is suitable for use in the detection of non-specific sites, preferably in vitro and in vivo, using the methods of the present invention. The gRNA can be, for example, a single guide RNA (sgRNA) or a double gRNA (dgRNA) comprising a crRNA (CRISPR RNA) / tracrRNA (transactivating crRNA) pair of RNA molecules. According to certain embodiments of the present invention, the gRNA is a sgRNA comprising one RNA or a dgRNA comprising two RNAs, preferably one crRNA and one tracrRNA, as components.

[0163] One or more target-specific programmable nucleases target a specific site or nucleotide sequence (specific site). The specific site is not particularly limited and may include one or more specific sites or sequences.

[0164] According to one embodiment of the present invention, in step b1), the one or more target-specific programmable nucleases are selected from Transcription Activator-Like Effector Nucleases (TALENs), Zn-finger Nucleases (ZFNs), Meganucleases, MegaTALs, Fokl-dCas9, Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) / Cas RNA-guided nucleases (CRISPR / Cas RGNs), or mutants thereof, or combinations thereof.

[0165] In a particular embodiment of the invention, in step b1), the one or more target-specific programmable nucleases preferably comprise, preferably consist of, one or more CRISPR / Cas RGNs, each used as preassembled ribonucleotide protease (RNP) with the Cas endonuclease bound to the gRNA, preferably in a molar ratio of 1:3. The Cas endonuclease may, for example, be a recombinant protein. According to an exemplary embodiment, about 1.6 pmol of Cas endonuclease and about 4.8 pmol of sgRNA, preferably pre-integrated in the Cas endonuclease / gRNA RNP, are used to cleave about 50 ng of purified end-blocked dsDNA fragments. For larger or smaller amounts of purified end-blocked dsDNA fragments, the Cas endonuclease and sgRNA used can, for example, be proportionally adapted.

[0166] Intra-step controls may be used to monitor the on-target activity of one or more target-specific programmable nucleases. According to one embodiment of the present invention, PCR amplification with site-specific primers for the specific sites of one or more target-specific programmable nucleases is used as an in-step control.

[0167] The affinity adapters used in step b2) of the method of the first aspect of the invention are not particularly limited, as long as they allow the ligation of the affinity adapters to the cleavage sites of purified end-blocked dsDNA fragments cleaved by one or more target-specific programmable nucleases (i.e., the cleaved dsDNA fragments) using the affinity adapters for specific enrichment of the affinity adapter-modified dsDNA fragments thus prepared. Such affinity adapters are known in the art. Preferably, the affinity adapters contain one or more modifications that allow specific enrichment of the affinity adapter-modified dsDNA fragments after ligation of the affinity adapters to the cleavage sites of the cleaved dsDNA fragments.

[0168] The affinity adapters used in the methods of the present invention are preferably DNA-based adapters, preferably dsDNA-based adapters (i.e., dsDNA-based affinity adapters), that contain one or more modifications suitable for affinity purification. Such dsDNA-based affinity adapters preferably have a sequence length that allows the incorporation of primer sites for PCR amplification into the affinity adapter.

[0169] Modifications suitable for affinity purification are known in the art and include, for example, biotin and biotin-based modifications, where biotin or biotin-based modifications are attached to the 5' or 3' end of the DNA-based affinity adaptor or to a specific nucleotide (or nucleotide derivative), such as deoxythymidine, of the DNA-based affinity adaptor. Suitable biotin and biotin-based modifications include, for example, biotin, biotin dT (biotinylated deoxythymidine base), biotin TEG (15-atom triethylene glycol spacer), desthiobiotin TEG, double biotin, photocleavable (PC) biotin, biotin azide, nitrilotriacetic acid (NTA), DNP, TEG (2,4-dinitrophenyl), and combinations thereof. Other suitable modifications include digoxigenin, protein tags, and combinations thereof. Preferably, the one or more modifications are selected from biotin and biotin-based modifications, and combinations thereof, preferably at least two modifications, In certain embodiments, the biotin modification is added to only one strand of the dsDNA-based affinity adapter.

[0170] Furthermore, the affinity adapter preferably comprises a primer site adapted for use in PCR priming, which means a nucleotide sequence to which a PCR primer can be annealed to amplify the cleaved dsDNA fragment to which the affinity adapter is ligated. Preferably, the sequence of the affinity adapter, preferably a dsDNA-based affinity adapter, and / or the primer site in the affinity adapter does not exist in the target genome (i.e., the sequence of genomic DNA in the input sample and / or reference genome).

[0171] In a preferred embodiment, the affinity adapter is a dsDNA-based affinity adapter comprising one or more biotin or biotin-based modifications. In a particular preferred embodiment, the affinity adapter is a DNA-based adapter comprising one or more, preferably at least two, internal biotin-dTs. In a further particular embodiment of the invention, the affinity adapter is a dsDNA-based affinity adapter comprising the sequence defined in SEQ ID NO:03 and the sequence of SEQ ID NO:04. In a further particular embodiment of the invention, the affinity adapter consists of the sequences of SEQ ID NO:03 and SEQ ID NO:04.

[0172] The affinity adaptor is ligated to the cleavage site of the cleaved dsDNA by a method known in the art. The cleavage site of the cleaved dsDNA fragment refers to the cleavage site in the dsDNA fragment cleaved by one or more target-specific programmable nucleases. For example, the affinity adaptor can be ligated to the cleavage site of the cleaved dsDNA fragment by enzymatic or chemical ligation using ligation protocols known in the art. Enzymatic ligation can be performed using known DNA ligases. In a preferred embodiment, the affinity adaptor is ligated to the cleavage site of the cleaved dsDNA by enzymatic ligation.

[0173] Preferably, ligating affinity adaptors to the cleavage sites of the cleaved dsDNA fragments includes: 1. preparing the cleaved dsDNA fragments for affinity adaptor ligation; and 2. ligating affinity adaptors to the cleavage sites of the cleaved dsDNA fragments to obtain affinity adaptor-modified dsDNA fragments. Methods for preparing cleaved dsDNA fragments for affinity adaptor ligation are known in the art. Preparing the cleaved dsDNA fragments for affinity adaptor ligation may include one or both of end repairing and dA tailing the cleaved dsDNA.

[0174] In a specific embodiment of the present invention, the cleavage site of the cleaved dsDNA fragment is prepared by end repair to prepare a cleaved dsDNA fragment without an overhang and containing a 5' phosphate group and a 3' hydroxyl group and a dA tailing for affinity adaptor ligation. A dsDNA-based affinity adaptor having a dT overhang at the 3' end of one end of the adaptor can be ligated to the end of the cleaved dsDNA fragment thus prepared.

[0175] After ligation, the affinity adapter-modified dsDNA fragments are enriched using affinity adapters to obtain enriched affinity adapter-modified dsDNA fragments. In this context, "enriching affinity adapter-modified dsDNA fragments" refers to specific, transient or permanent binding to affinity adapter-modified dsDNA fragments via affinity adapters, and separation from dsDNA fragments that are not affinity adapter-modified, preferably and optionally other components. Thus, how affinity adapter-modified dsDNA fragments are enriched depends on the affinity adapters used (i.e., the type of affinity tag used in the affinity adapters). Protocols for enriching affinity adapter-modified dsDNA fragments using specific affinity adapters are known in the art.

[0176] According to a preferred embodiment, the affinity adaptor binding molecule, preferably immobilizable on a solid support, is used to enrich the affinity adaptor-modified dsDNA fragments. The affinity adaptor binding molecule refers to a molecule, such as a protein, an antibody, etc., that specifically binds (reversibly or irreversibly) to an affinity adaptor. The affinity adaptor binding molecule can be immobilized on a solid support before or after binding to the affinity adaptor. The nature of the affinity adaptor binding molecule depends on the affinity adaptor used. The affinity adaptor binding molecule can be, for example, a protein. The solid support is not particularly limited and includes, for example, beads such as magnetic beads. In a preferred embodiment, magnetic beads, preferably polymer-coated magnetic beads such as Dynabeads, are used as the solid support.

[0177] According to one embodiment, the affinity adaptor binding molecule is preferably a protein selected from streptavidin, avidin; an antibody or functional fragment thereof specific for a portion of the affinity adaptor, such as an anti-digoxigenin antibody; a digoxigenin binding protein; or a variant thereof; more preferably streptavidin, avidin, or a variant thereof. The antibody or functional fragment thereof specific for a portion of the affinity adaptor preferably specifically binds to one or more modifications suitable for affinity purification on the affinity adaptor, such as digoxigenin or a protein tag.

[0178] According to a preferred embodiment, the affinity adaptor comprises one or more biotin or biotin-based modifications, and the affinity adaptor binding molecule is streptavidin, avidin or a variant thereof, preferably streptavidin or a variant thereof. Protocols for enriching DNA modified with one or more biotin or biotin-based modifications are known in the art. Streptavidin or a variant thereof may be used immobilized on polymer-coated magnetic beads. For example, Dynabeads Myone Streptavidin C1, Dynabeads M-270, Dynabeads M-280, or Dynabeads Myone Streptavidin T1 may be used.

[0179] The enriched affinity adaptor-modified dsDNA fragments are amplified by PCR amplification to obtain a DNA library of cleaved and amplified dsDNA fragments. PCR amplification of dsDNA fragments, including affinity adaptor-modified dsDNA fragments, as described herein, is known in the art. In a preferred embodiment, PCR primers specific to portions of the blocking adaptor and / or affinity adaptor (i.e., complementary to a portion of the sequence of the blocking adaptor and / or affinity adaptor) are used for PCR amplification.

[0180] When the affinity adaptor-modified dsDNA fragments are enriched using affinity adaptor-binding molecules immobilized on a solid support, the affinity adaptor-modified dsDNA fragments bound to the solid support via the affinity adaptor-binding molecules can serve as templates for PCR amplification. Alternatively, the enriched affinity adaptor-modified dsDNA fragments are separated (e.g., released) from the affinity adaptor-binding molecules and / or the solid support prior to PCR amplification. In a preferred embodiment, PCR amplification is performed at least initially using the affinity adaptor-modified dsDNA fragments bound to the solid support via the affinity adaptor-binding molecules as templates. As used herein, the term "template" refers to the function of the enriched affinity adaptor-modified dsDNA fragments to function as DNA templates for direct synthesis of other DNA or RNA during PCR amplification.

[0181] According to a preferred embodiment of the present invention, obtaining a DNA library of cleaved and amplified dsDNA fragments in step b4) comprises two-step PCR amplification, preferably comprising: 1. a first PCR amplification step: performing linear PCR amplification using the enriched affinity adapter-modified dsDNA fragments bound to the solid support via the adapter binding molecule as a template and a primer specific to the affinity adapter; and 2. a second PCR amplification step following the first step: removing the enriched affinity adapter-modified dsDNA fragments bound to the solid support via the adapter binding molecule from the reaction mixture, and adding a primer specific to the blocking adapter as a second primer.

[0182] Step c) Sequencing The method of the first aspect of the invention further comprises the step (c) of sequencing the DNA library of cleaved and amplified dsDNA fragments to obtain sequence reads. How to sequence a DNA library of dsDNA fragments is known in the art.

[0183] As used herein, "sequencing" includes any method of determining the sequence of a nucleic acid. Sequencing methods for use in the present method are not particularly limited and include, for example, chain terminator (Sanger) sequencing and dye terminator sequencing. In a preferred embodiment, next generation sequencing (NGS) is used. Different NGS platforms use various assay chemistries, but they generally generate sequence data from multiple sequencing reactions run simultaneously on multiple templates. Typically, sequence data is collected using a scanner, then assembled and bioinformatically analyzed. In this manner, sequencing reactions are performed, read, assembled, and analyzed in parallel (see, for example, US 2014 / 0162897; and MacLean et al., 2009 Nature Rev. Microbiol. 7:287-296).

[0184] Some NGS methods require template amplification, while others do not. Methods that require amplification include pyrosequencing, the Solexa / Illumina platform, or supported oligonucleotide ligation and detection (SOLiD) (see, e.g., EP 0 946 752; EP 1634 963; EP 0 777 749; EP 1 117 827; and EP 0 777 749; and U.S. Patent Nos. 6,969,488 and 6,130,073). Methods that do not require amplification include, for example, single molecule sequencing such as nanopore sequencing, HeliScope single molecule sequencing, and real-time sequencing using zero-mode waveguide synthetic single molecule real-time (SMRT) DNA sequencing (see, for example, EP 2 007 908; EP 1 848 829; EP 2 100 971; and U.S. Patent Nos. 7,169,560; 7,501,245; 6,818,395; 7,501,245; and 7,170,050). In a preferred embodiment, Illumina sequencing is used.

[0185] In a preferred embodiment, in step c) for sequencing the DNA library of cleaved and amplified dsDNA fragments, in step c1), a DNA library of cleaved and amplified dsDNA fragments is prepared for sequencing, if necessary, by ligating sequencing adaptors to both ends of the dsDNA fragments, and in step c2), the prepared DNA library of cleaved and amplified dsDNA fragments is sequenced, preferably by NGS, to obtain sequence read data. In a particular embodiment, in step c1), preparing a DNA library of cleaved and amplified dsDNA fragments for sequencing comprises preparing a DNA library of cleaved and amplified dsDNA fragments for ligating sequencing adaptors to both ends, preferably by one or both of end repair and dA tailing of the DNA library of cleaved and amplified dsDNA fragments, followed by ligating sequencing adaptors, preferably to both ends of the dsDNA fragments, where said sequence adaptors are, for example, Illumina sequence adaptors. In step c2), preferably droplet digital PCR (ddPCR) may be used, for example using the ddPCR Library Quantification Kit for Illumina TruSeq.

[0186] Sequencing in step (c) does not require a high performance sequencer such as a HiSeq. Instead, a basic NGS sequencer can be used, such as a MiSeq instrument (Illumina) with the MiSeq Reagent Kit v2.

[0187] Step d) Aligning sequence reads to the reference genome The method of the first aspect of the present invention further comprises a step (d) of aligning the sequence to the reference genome to obtain aligned sequence read data. This alignment in context means to position the sequence read to the corresponding sequence in the reference genome, so that for a particular sequence, the corresponding sequence in the reference genome is most consistent with the sequence of the sequence read (i.e., ranging from incomplete match to sequence identity). The method of aligning the sequence read to the reference genome is known in the art. For example, any computer program may be used as long as the sequence read can be aligned as described above. They may be known programs already known in the relevant art, or programs tailored for the purpose. In a specific embodiment of the present disclosure, the sequence read is aligned to the reference genome using Bowtie2.

[0188] According to a preferred embodiment of the invention, the genomic DNA is human genomic DNA and the reference genome is a human reference genome selected from, for example, hg16 (NCBI Build 34), hg17 (NCBI Build 35), hg18 (NCBI Build 36.1), hg19 (GRCh37), and hg38 (GRCh38), preferably selected from hg19 and hg38.

[0189] In a particular preferred embodiment, in step d), the sequence reads are first aligned to the reference genome by trimming blocking adapter sequences, affinity adapter sequences, and any sequencing adapter sequences from the sequence reads in step d1) to obtain trimmed sequence reads; and then, in step d2), mapping the trimmed sequence reads to a reference genome to obtain aligned sequence reads.

[0190] Step e) Identification of cleavage sites The method of the first aspect of the invention further comprises a step (e) of identifying cleavage sites by determining putative cleavage sites and eliminating false positive cleavage sites, where the putative cleavage sites are determined by locating sequence regions in the reference genome that match both the beginning and end of the aligned sequence reads, and identifying putative cleavage sites as those of these sequence regions where the beginning and end of the sequence reads match.

[0191] Thus, the determination of the cleavage site is not based on sequence homology to the specific cleavage sites of one or more target-specific programmable nucleases.

[0192] The term "putative cleavage site" may include any cleavage site cleaved in vitro by one or more target-specific programmable nucleases, as well as false positive cleavage sites. False positive cleavage sites are sites that are measured as putative cleavage sites during cleavage site detection, rather than as sites cleaved in vitro by one or more target-specific programmable nucleases, and are therefore erroneously identified, i.e., false positives. The term "putative cleavage site" as used herein refers to a short nt sequence (i.e., a short DNA sequence in the reference genome), preferably of a set length (i.e., does not change from one putative cleavage site to another). In a preferred embodiment, the putative cleavage site measured in step e) consists of a nt sequence in a specific window of set sequence width (i.e., a window covering a defined sequence length of the reference genome that does not change from window to window), preferably of a window of about 6 nt or less, more preferably of a window of (about) 4 nt. In other words, in said preferred embodiment, the putative cleavage site corresponds to a DNA sequence in the reference genome covered by a specific window.

[0193] The sequence region in the reference genome where both the read start portion and the read end portion of the aligned sequence match is a region in the reference genome of variable or set sequence length that includes aligned sequence reads that include both the read start portion and the read end portion. That is, it is a sequence region where one or more aligned sequence reads start and one or more sequence reads end in the sequence region among all aligned sequence reads in the region aligned to the reference genome. In a preferred embodiment, each sequence region has a defined sequence length, preferably about 20 to 1000 nt, more preferably about 50 to 500 nt, even more preferably at least about 50 nt, even more preferably at least about 80 nt. Most preferably, it is a sequence group of at least (about) 100 nt. In another preferred embodiment, in step e), each sequence region is a sequence group of a defined sequence length, preferably about 100 nt or a multiple thereof, more preferably 100 nt or a multiple thereof (i.e., 100 nt, 200 nt, 300 nt, 400 nt, etc.).

[0194] The sequence read start refers to one end of the aligned sequence read that is located opposite to the sequence read end. For sequence reads aligned to the same nucleotide strand of the reference genome, all sequence read start points in the same direction, and all sequence read end points in the opposite direction of the nucleotide strand of the reference genome sequence. That is, preferably, for a first aligned sequence read that is complementary to a first single-stranded sequence of the reference genome, the sequence read start is at the 3' end of the first aligned sequence read, and for a hypothetical second aligned sequence read that is exactly the same length and exactly complementary to the first aligned sequence read, the sequence read start is at the 5' end of the second aligned sequence read. This allows both aligned sequence reads to start at the same spatial position of the reference genome. Thus, according to certain embodiments of the present invention, any sequence read start of the sequence reads aligned to the same continuous sequence of the reference genome points in the opposite direction to any sequence read end of the sequence reads aligned to the same continuous sequence of the reference genome. That is, the sequence read start portions of all aligned sequence reads that are aligned to the same contiguous sequence of the reference genome are spatially located in the aligned sequence read array on one side of the sequence center of each of the individual aligned sequence reads, and the sequence read end portions of all aligned sequence reads that are aligned to the same contiguous sequence of the reference genome are spatially located in the aligned sequence read array on the opposite side (relative to the sequence read start portions) of the sequence center of each of the individual aligned sequence reads.

[0195] In the sequence regions thus identified, putative cleavage sites are identified as those in these sequence regions where the sequence read start and end coincide. In this context, the term "coincides" means that the sequence read start and end points pointing to each other are close together, i.e., within a few nt or in contact, i.e., aligned neatly so that they are not separated on the reference genome sequence from which they are read, or overlap by a few nt. Putative cleavage sites in these sequence regions can then be identified, for example, by finding the local maximum of the sum of the aligned sequence read start and end.

[0196] In a particular embodiment of the invention, in step e), for each sequence region with non-zero read start and read end coverage, one or more putative cleavage sites (e.g. exactly one putative cleavage site per sequence region, or two putative cleavage sites per sequence region, etc.) are determined as sites where the sequence read start and sequence read end of the aligned sequence reads coincide and where the sum of the number of sequence read start and the number of sequence read end is a local maximum. Non-zero read start and read end coverage means that such a sequence region is a region where one or more sequence read start and one or more sequence read end of the aligned sequence reads align.

[0197] According to a preferred embodiment of the invention, in step e), the putative cleavage sites are determined as follows: first, the reference genome (DNA sequence) containing the aligned sequence reads is divided into sequence groups of defined length, then, for each sequence group with non-zero coverage of the read start and read end, one or more putative cleavage sites are determined as sites with a local maximum of the sum of the sequence read start and sequence read end, and optionally, the distance to the center of the sequence group is minimized. Here, the sequence read start consists of the first 1-5 nt, preferably the first 2-4 nt, more preferably the first 3 nt of the aligned sequence read, and the sequence read end consists of the last 1-5 nt, preferably the last 2-4 nt, more preferably the last 3 nt of the aligned sequence read. In a specific embodiment of the invention, one putative cleavage site is determined for each sequence group.

[0198] In a preferred embodiment of the present invention, the putative cleavage sites measured in step e) consist of a specific window of set sequence width (i.e. a window covering a set sequence length of the reference genome that does not change for each putative cleavage site), preferably a window of 4 nt. In other words, in said preferred embodiment, the putative cleavage sites correspond to the DNA sequence in the reference genome covered by a specific window. Preferably, such a window is a defined width within the sequence group, for example a window of 4 nt, with a local maximum for the sum of the start and end parts of the sequence reads (aligned within said window). Optionally, the measurement of said window comprises as an additional step the minimization of the distance to the center of said sequence group.

[0199] A sequence group having non-zero read start and read end coverage refers to a sequence group in which one or more sequence read start portions and one or more sequence read end portions of the aligned sequences are aligned. In the above embodiment, the sequence group width (i.e., the nucleotide sequence length covered by the sequence group) is not particularly limited, and can be, for example, about 20 to 1000 nt, preferably about 50 to 500 nt, more preferably at least about 50 nt, and even more preferably at least about 80 nt. In a preferred embodiment, the sequence group width is at least about 100 nt, preferably at least 100 nt. In another embodiment, the sequence group width is about 100 nt or a multiple thereof, more preferably 100 nt or a multiple thereof (i.e., 100 nt, 200 nt, 300 nt, 400 nt, etc.).

[0200] According to a particular embodiment of the invention, in step e) the cleavage site is identified as follows: e1) dividing the reference genome containing the aligned sequence reads into sequence groups of a set sequence length, preferably about 100 nt in length; e2) preferably for each set of sequences, measuring the coverage of reads, discarding sequences with zero coverage of reads, and combining sequences with >0 coverage of reads into a set of contiguous sequences, preferably based on sequence order in the reference genome; e3) dividing each set of sequences into windows of a set (sequence) width, preferably 4 nt, with a step size equal to or smaller than the set width of the window, preferably smaller, preferably 2 nt; e4) Measure the beginning, end, and background coverage of each window by grouping as follows; - Grouping the start of sequence reads that fall within the window region into start coverage groups; - Grouping the ends of sequence reads that fall within the window region into end-covering groups; - Group sequence reads that fall within the window region but do not belong to the start and end coverage groups into a background coverage group. e5) for each group of sequences that includes at least one sequence read start and at least one sequence read end, determining one or more putative cleavage sites by maximizing the sum of start and end coverage and minimizing the distance to the center of the group of sequences; and e6) excluding false positive cleavage sites in which the sequence read start portion consists of the first 1-5 nt, preferably the first 2-4 nt, more preferably the first 3 nt of the aligned sequence read, and the sequence read end portion consists of the last 1-5 nt, preferably the last 2-4 nt, more preferably the last 3 nt of the aligned sequence read.

[0201] Preferably, each of the one or more putative cleavage sites measured in step e5) consists of nts of a specific window of set sequence width (i.e. a window covering a defined sequence length of the reference genome that does not change for each putative cleavage site), preferably of a window of 4 nt nts set in step e3). The step size used here preferably refers to the difference in nts between the first nt of the first window and the first nt of the second window. This is the window located immediately next to the first window along the sequence of the reference genome (i.e. the window located closest to the first window when moving in one direction from the first window along the sequence of the reference genome). The set step size refers to a step size that does not change between different windows located closest to each other, for example a step size of 2 nt.

[0202] In a particular preferred embodiment of the present invention, step e1) comprises dividing the reference genome consisting of sequenced sequences into sequence groups (preliminary sequence groups) of set sequence length, preferably about 100 nt in length, and step e2) comprises measuring the coverage of read data for each (preliminary) sequence group of e1), discarding sequence groups with a coverage of read data of 0, and combining sequence groups with a coverage of read data >0 into a continuous sequence group set (final sequence group), preferably based on the sequence order of the reference genome, wherein in the continuous sequence group set of step e2), each (final) sequence group preferably consists of one or more sequence groups of step e1) (i.e. 1, 2, 3, 4, 5, etc. preliminary sequence groups).

[0203] Preferably, in the particular preferred embodiment, in the set of consecutive sequence groups of step e2), each (final) sequence group has a sequence group width defined by the sum of the consecutive sequence groups of step e1) between the two discarded sequence groups of step e2) that are closest to each other in the sequence of the reference genome. Thereby, each (final) sequence group of step e2) starts with a sequence of the reference genome that includes the sequence group of step e1) (preliminary sequence group) and ends with a sequence group of step e1) that immediately follows the first sequence group discarded in step e2) and immediately precedes the next sequence group discarded in step e2) (i.e. the next sequence of the reference genome) (i.e. the end of the sequence group). That is, in the embodiment, the individual (final) sequence groups of step e2) may have different lengths (cover different lengths on the sequence of the reference genome). However, there is always more than one sequence group of step e1) (i.e. one, two, three, four, five, etc. sequence groups of step e1).

[0204] In another particular embodiment of the invention, step e4) comprises determining the start, end and background coverage of each group of sequences by grouping as follows: - Grouping the start of sequence reads that fall into sequence cluster regions into start coverage clusters; - Grouping the ends of sequence reads that fall into sequence cluster regions into end-covering clusters; - grouping sequence reads that fall into a sequence group region but do not belong to the start and end coverage groups into a background coverage group; The method further includes adding up, for each sequence group, the number of read data in the start covering group, the end covering group and the background covering group, respectively.

[0205] In step e3), the set window width (i.e., the set length and the constant length of the nucleotide sequence of the reference genome covered by each window) and the set step size are not particularly limited. In a preferred embodiment, the set window width, which is constant for all windows, is 8 nt or less, preferably 6 nt or less, most preferably 4 nt, and the set step size is less than or equal to the window width, preferably less than the window width, for example 6 nt or less, preferably 4 nt or less (e.g., 4, 3, 2, or 1 nt), most preferably 2 nt. In a particular embodiment, putative cleavage sites are measured in step e5) for each group of sequences. In a preferred embodiment of the present invention, the putative cleavage sites measured in step e) consist of a window of nt, preferably 4 nt, of a specific window (i.e., a sequence covered by a specific window) defined in step e3).

[0206] At a specific putative cleavage site, the aligned sequence read start and the sequence read end of the putative cleavage site may not be perfectly aligned with a specific position in the reference genome sequence. According to one embodiment, at a putative cleavage site, the aligned sequence read start starts + / -2 nt or less, preferably + / -1 nt or less, of a specific first position in the reference genome sequence, and the aligned sequence read end ends + / -2 nt or less, preferably + / -1 nt or less, of a specific second position in the reference genome sequence. Preferably, the two specific positions are separated from each other by a few nt (for example, only 1, 2 or 3 nt) in the sequence of the reference genome, or are directly adjacent to each other.

[0207] After determining the putative cleavage sites, false positive cleavage sites (also referred to herein as "false positive sites") are removed from the list of putative cleavage sites to identify the cleavage sites. As used herein, the term "cleavage site" refers to a short sequence of nt (i.e., a short sequence in the reference genome), preferably of a set length (i.e., that does not vary from putative cleavage site to putative cleavage site). In a preferred embodiment, the cleavage site is (approximately) 4 nt in sequence length. In a particular embodiment, the cleavage sites identified in step e) consist of a specific nt window of set sequence width (i.e., a window covering a set sequence length of the reference genome that does not vary from window to window), preferably a window of about 6 nt or less, more preferably a window of (approximately) 4 nt. In a preferred embodiment of the invention, false positive cleavage sites are removed based on the ratio of background coverage to signal coverage at the putative cleavage site, the ratio of start coverage to end coverage at the putative cleavage site, site coverage, and / or sequence group coverage. Preferably, the actual cleavage of the cleavage site of step e) by one or more target-specific programmable nucleases occurs within the sequence of that particular cleavage site, i.e. within the sequence of the cleavage site consisting of a particular window of nt of set sequence width, anywhere in the sequence covered by said particular window.

[0208] The start coverage at the putative cleavage site is the sum of the sequence read start parts aligned in the sequence region of the putative cleavage site, i.e., the sum of the aligned sequence reads having the sequence read start part aligned to the sequence region of the reference genome defined by the putative cleavage site. In a preferred embodiment, the start coverage at the putative cleavage site is the sum of the sequence read start parts aligned in the sequence region of the putative cleavage site, which is defined by a specific window with a sequence width set (i.e., the window does not change in different windows for different putative cleavage sites), preferably a window with a sequence width of 4 nt. In other words, in the preferred embodiment, the start coverage at the putative cleavage site is the sum of the aligned sequence reads having the sequence read start part aligned to the sequence region of the reference genome defined by a specific window (i.e., the window preferably set in step e3 above) representing the putative cleavage site), for example, a window of 4 nt.

[0209] Similarly, the end coverage at the putative cleavage site is the sum of the sequence read end parts aligned in the sequence region of the putative cleavage site, i.e., the sum of the aligned sequence reads having the sequence read end part aligned to the sequence region of the reference genome defined by the putative cleavage site. In a preferred embodiment, the end coverage at the putative cleavage site is the sum of the sequence read end parts aligned in the sequence region of the putative cleavage site, the putative cleavage site being defined by a specific window of sequence width set (i.e., the window does not change in different windows for different putative cleavage sites), preferably a window of sequence width of 4 nt. In other words, in the preferred embodiment, the end coverage at the putative cleavage site is the sum of the aligned sequence reads having the sequence read end part aligned to the sequence region of the reference genome defined by a specific window (i.e., the window preferably set in step e3 above) representing the putative cleavage site), e.g., a window of 4 nt.

[0210] Preferably, the start coverage of a particular window defining a particular putative cleavage site also includes all sequence read start portions that align only with a portion of the sequence of the sequence read start portion that is in the sequence region of the reference genome defined by the particular window. That is, it is sufficient that the sequence read start portion aligns with the sequence region of the particular window, for example, 1 nt or 2 nt (depending on the length of the sequence read start portion) to the sequence region of the particular window, in order to be included in the start coverage group of the particular window. Similarly, preferably, the end coverage of a particular window defining a particular putative cleavage site also includes all sequence read end portions that align only with a portion of the sequence of the sequence read end portion that is in the sequence region of the reference genome defined by the particular window. That is, it is sufficient that the sequence read end portion aligns with the sequence region of the particular window, for example, 1 nt or 2 nt (depending on the length of the sequence read end portion) to be included in the end coverage group of the particular window.

[0211] The background coverage at a putative cleavage site is the sum of sequence reads aligned at the sequence region of the putative cleavage site, but not belonging to the start coverage group and the end coverage group, i.e., not having a sequence read start or sequence read end aligned to the sequence region of the reference genome defined by the putative cleavage site. The putative cleavage site is optionally defined by a specific window of set sequence width (i.e., the width does not change for different windows corresponding to different putative cleavage sites), preferably a window of 4 nt sequence width (i.e., the window preferably set in step e3) above).

[0212] The term "site coverage" as used herein refers to the sum of start and end coverage at the putative cleavage site, i.e., the sum of aligned sequence read start and sequence read end that meet / match at the putative cleavage site. In a preferred embodiment, the term "site coverage" refers to the sum of start and end coverage at the putative cleavage site, i.e., the putative cleavage site is defined by a specific window of set sequence width (i.e., the width does not change in different windows corresponding to different putative cleavage sites), preferably a window of 4 nt sequence width. In other words, the term "site coverage" refers to the sum of sequence read start and sequence read end aligned to the reference genome within a sequence region of a specific window of set sequence width (i.e., the window preferably set in step e3) above) that represents a specific putative cleavage site, for example, of 4 nt. The term "sequence group coverage" as used herein refers to the sum of all aligned sequence reads in a sequence group (i.e., the sum of all aligned sequence reads that align to a sequence region of the reference genome of a specific sequence group).

[0213] In another preferred embodiment of the invention, false positive cleavage sites are eliminated by selecting only putative cleavage sites that: 1) have a ratio of background coverage to signal coverage not exceeding a certain threshold, preferably about 0.5; and / or 2) a ratio of starting coverage to end coverage at the putative cleavage site within a certain numerical range, preferably within about 1 / 5 to 5 / 1; and / or 3) a minimum site coverage, preferably about 6 or more; and / or 4) a minimum sequence group coverage, preferably 10 or more.

[0214] In certain embodiments of the invention, false positive cleavage sites are eliminated by selecting only putative cleavage sites where: 1) the ratio of background coverage to signal coverage at the putative cleavage site is less than or equal to 0.5, and 2) the ratio of start coverage to end coverage at the putative cleavage site is between 1 / 5 and 5 / 1, and 3) the site coverage is greater than or equal to 6, and 4) the sequence population coverage is greater than or equal to 10.

[0215] According to yet another specific alternative embodiment of the present invention, false positive cleavage sites are removed or further removed using a machine learning model. Here, the machine learning model is preferably trained on the false positive sites removed by the procedure described in the above embodiment, and is preferably trained using the ratio of background coverage to signal coverage, the ratio of starting coverage to end coverage, the sequence population width, the relative distance between false positive sites, and / or the ratio of false positive sites to sequence population as features. Methods for using machine learning models, including appropriate model selection, feature selection, and training, are known in the art. For example, a random forest model can be constructed using 3-fold 10-replicate cross-validation and obtaining the area under curve 1.

[0216] In one embodiment of the invention, each cleavage site of step e) constitutes one target site (i.e. one specific site or one non-specific site) of one or more target-specific programmable nucleases, whereby preferably each target site (i.e. specific site or non-specific site) consists of one specific cleavage site that is different for each target site (one of the individual cleavage sites identified in step e). In a particular embodiment, when the specific cleavage sites of step e) consist of a specific window of nts of a set sequence width, each target site (i.e. specific site or non-specific site) consists of a specific window of sequence of a set sequence width that defines one specific cleavage site, whereby each specific window that defines one specific cleavage site identified in step e) results in one target site.

[0217] According to another embodiment of the invention, the cleavage sites of step e) are comprised in the target sites (i.e. specific sites or non-specific sites) of one or more target-specific programmable nucleases, whereby each target site (i.e. specific site or non-specific site) comprises one cleavage site. As will be appreciated by the skilled artisan, the cleavage sites measured in step e), which in this embodiment may consist of a certain window of nts of the set sequence width of step e), do not necessarily have to comprise the entire target site sequence (i.e. the sequence of the specific site or non-specific site), since, for example, the specific site may further comprise, for example, a target sequence (i.e. a sequence that a particular target-specific programmable nuclease is designed to recognize) and / or a nuclease binding sequence (CRISPR / Cas RGN, e.g. a PAM sequence), and similarly, the non-specific site may in this embodiment further comprise, for example, a sequence that corresponds to a certain degree (even to a very small degree, e.g. 1 nt, or 2 nt, or 3 nt, etc.) with the target sequence and / or the nuclease binding sequence.

[0218] The CRISPR / Cas RGN target site may in this alternative embodiment comprise a target sequence and a PAM sequence in addition to the cleavage site. For example, when the one or more target-specific programmable nucleases used in step b) consist of one or more CRISPR / Cas RGNs, the target site may in this alternative embodiment comprise or consist of, for example, about 18-23 nt, preferably about 20 nt, of target sequence that is identical or at least homologous to a very low degree (i.e., for example, 1 nt, or 2 nt, or 3 nt, or 4 nt, or 5 nt, or 6 nt, or 7 nt, or 8 nt, etc. sequence homology) to the target sequence of the one or more target-specific programmable nucleases used in step b), and, immediately adjacent to the target sequence, a PAM sequence of, for example, 2-6 nt, preferably 3 nt. Thus, in this other embodiment, in certain preferred embodiments where the one or more target-specific programmable nucleases comprise one or more CRISPR / Cas RGNs, the target site, preferably the non-specific site, comprises or consists of, for example, a sequence of about 20-29 nt in length, preferably about 23 nt. Such, e.g., 23 nt sequences can be identified by identifying a cleavage site in step e) that is adjacent to or overlapping with: 1) a 2-6 nt, preferably 3 nt, PAM sequence, and 2) a target sequence that is at least somewhat, preferably very low, homologous to the target sequence of one of the one or more CRISPR / Cas RGNs used in step b) (e.g., at least 1 nt, or at least 2 nt, or at least 3 nt, or at least 4 nt, or at least 5 nt, or at least 6 nt, or at least 7 nt, or at least 8 nt, or at least 9 nt, or at least 10 nt, or at least 11 nt, or at least 12 nt, or at least 13 nt, etc., of sequence homology).The cleavage site is preferably located at a target site (i.e., a specific site or a non-specific site) that is close to the end of the target sequence immediately adjacent to the PAM sequence, e.g., a 23 nt long region, e.g., within a sequence range consisting of about the first 1-7 nt of the end of the target sequence immediately adjacent to the PAM sequence and about the first 1-2 nt of the end of the PAM sequence immediately adjacent to the target sequence.

[0219] According to a particular embodiment of the method of the first aspect of the invention, the non-specific cleavage sites of one or more target-specific programmable nucleases are identified as cleavage sites of step e) that are not non-specific cleavage sites of one or more target-specific programmable nucleases. According to another particular embodiment of the method of the first aspect of the invention, the non-specific cleavage sites of one or more target-specific programmable nucleases are identified as specific sites that comprise the cleavage sites of step e) that are not specific sites of one or more target-specific programmable nucleases. In practice, this can be done by identifying (e.g. by labeling) the target sites of one or more target-specific programmable nucleases in the list of cleavage sites identified in step e), which are known as a result of designing one or more target-specific programmable nucleases. This can be done by means known in the art.

[0220] Duplex-Abnoba-Seq In a second aspect, the present invention relates to a method for detecting non-specific sites in vivo using the method for in vitro detection of non-specific sites of one or more target-specific programmable nucleases in a genome of the first aspect (Dual-Abnoba-Seq). In particular, a second aspect of the invention relates to a method for detecting specific sites in a genome in vivo comprising the steps of: i) providing a first sample comprising genomic DNA of cells that have been treated with one or more target-specific programmable nucleases and a second sample comprising genomic DNA of cells of a related genetic background, preferably cells of the same genetic background, that have not been treated with one or more target-specific programmable nucleases; ii) detecting non-specific sites in vivo in each of the two samples using the method of the first aspect of the invention as defined in any one of the above embodiments and using one or more target-specific programmable nucleases; and 3) comparing the non-specific sites between the two samples and identifying non-specific sites identified only in the sample comprising genomic DNA of cells that have not been treated with one or more target-specific programmable nucleases and not identified in the sample comprising genomic DNA of cells that have been treated with one or more target-specific programmable nucleases as in vivo non-specific sites of the one or more target-specific programmable nucleases.

[0221] The method according to the second aspect of the invention allows for personalized measurement of the in vivo specificity of one or more target-specific programmable nucleases, in other words the ability to measure the actual off-target activity of one or more target-specific programmable nucleases in specific cells, i.e. the method according to the second aspect of the invention can be used to demonstrate target-specific programmable nuclease activity in vivo, for example in a specific cell type.

[0222] As the inventors have found, isolating genomic DNA from gene-edited and untreated cells and carrying out on these samples of genomic DNA a method for in vitro detection of one or more non-specific sites of target-specific programmable nucleases in the genome using the target-specific programmable nucleases used for gene editing in the gene-edited cells according to the first aspect of the invention allows for the measurement of actual off-target activity in the cells.

[0223] By comparing the non-specific sites identified in the gene-edited sample of genomic DNA with the non-specific sites in the sample of genomic DNA from untreated cells, the actual in vivo activity of one or more target-specific programmable nucleases in cells of a particular cell type or tissue can be measured using a minimal amount of genomic DNA, e.g., less than 1 μg per sample, as sites that are no longer identified in the sample of genomic DNA from gene-edited cells. This allows for the testing of off-target activity in vivo in precious and limited availability cell samples with very small sample amounts, e.g., less than 1 μg per sample.

[0224] Methods for treating cells with one or more target-specific programmable nucleases are known in the art (see, for example, Yin et al., 2017, Delivery techniques for genome editing. Nat Rev Drug Discov 16,387-399). One or more target-specific programmable nucleases are introduced into cells by means known in the art and at a concentration and form suitable for genome editing in vivo. For example, when the one or more target-specific programmable nucleases include one or more CRISPR / Cas RGNs, for example, gRNAs, Cas endonucleases, or both can be introduced into cells as purified proteins and / or nucleic acids or expressed transiently or stably in cells using methods known in the art. The one or more target-specific programmable nucleases may be one of those described above in the first embodiment of the invention.

[0225] In particular, a second aspect of the invention relates to a method for detecting non-specific sites in a genome in vivo, comprising the steps of: i) preparing a first sample comprising genomic DNA of cells treated with one or more target-specific programmable nucleases, and a second sample comprising genomic DNA of cells of a related genetic background, preferably cells of the same genetic background, not treated with one or more target-specific programmable nucleases. Preferably, related genetic background means that such cells have the same genetic background, meaning with as few differences as possible, ideally the genomic DNA of these cells is the same except for minor variations that may occur from one cell to another, e.g. due to random mutations. Preferably, the related genetic background of the cells are preferably cells originating from the same anatomical region in a multicellular organism (e.g. human body), e.g. the same tissue and / or the same cell type.

[0226] In a preferred embodiment of the method of the second aspect of the present invention, the cells of related genetic background are cells of the same genetic background, preferably cells of the same genetic code, preferably cells of the same anatomical origin, preferably cells originating from the same tissue and / or cell type. In another embodiment, the cells of related genetic background are cells of the same cell lineage. Cells of a cell lineage as used herein very generally refer to cells derived from a common ancestor cell that is the result of division of a dividing cell and, optionally, the result of cell differentiation. Preferably, cells of the same cell lineage are cells originating from the same tissue, preferably cells of the same cell type.

[0227] Preferably, the cells of related genetic background are cells derived from the same individual organism. The type of organism from which the cells of the first and second samples are derived is not particularly limited, so long as it is suitable for use in the method of the second aspect of the present invention. In one embodiment, the cells are eukaryotic cells (e.g., animals, mammals, insects, plants, fungi, insects, birds, fish, amphibians, worms, or cnidarians). In certain embodiments, the cells are human cells. For example, the cells may be primary cells.

[0228] For both samples, non-specific sites are detected in vitro independently using the method of the first aspect of the invention with one or more target-specific programmable nucleases in step i). Finally, the non-specific sites between the two samples are compared to identify non-specific sites that are identified only in the sample containing genomic DNA of cells not treated with one or more target-specific programmable nucleases and not in the sample containing genomic DNA of cells treated with one or more target-specific programmable nucleases as in vivo non-specific sites of said one or more target-specific programmable nucleases. Methods for comparing non-specific sites between two samples are known in the art. Preferably, this is done using a computer program that is existing or specially made for this purpose. For example, after normalization of sequence population coverage by sequence population width, cleavage sites can be selected according to their location in the genome and the results of edited cells written next to the results of untreated cells.

[0229] In a further aspect, the present invention relates to a kit for carrying out the method according to the first aspect of the invention as described in any of the above embodiments, and to a kit for carrying out the method according to the second aspect of the invention as described in any of the above embodiments. Methods for preparing such kits are known in the art.

[0230] The present invention is further illustrated by the following examples. EXAMPLES

[0231] Materials and methods applied in the examples Formation of ribonucleotide complexes For in vitro experiments, gRNA was transcribed in vitro from PCR-amplified products using the HiScribe T7 High Yield RNA Synthesis Kit (NEB) followed by cleaning with the MEGAclear kit (Thermo Fisher Scientific). For cell-based experiments, chemically modified gRNA (Synthego) was purchased. Cas9 protein was home-made according to a previously published protocol (Jinek et al., 2012, Science 337(6096):816-821). To form the RNP complex, Cas9 and gRNA were incubated at room temperature for 10 min in a 1:3 molar ratio.

[0232] Cell culture and gene transfer U20S cells were cultured in DMEM with low glucose and pyruvate (Gibco) supplemented with 10% FBS and 1% penicillin / streptomycin (Sigma-Aldrich). CD34+ cells were purified from human umbilical cord blood obtained with informed consent from obstetricians and gynecologists by Ficoll-based density gradient centrifugation followed by positive selection using the human CD34 MicroBead Kit UltraPure (Miltenyi Biotec). The CD34+ cells were cultured in CellGro SCGM (CellGenix®) supplemented with 20 ng / μl TPO, 60 ng / μl SCF, and 60 ng / μl Flt-3L (CellGenix®). Prior to electroporation, 12 ng / μl IL-3 was added to the CD34+ cell culture medium. Cells were transfected with RNPs using Nucleofector 4D (Lonza Group AG) using either the SE kit and program DN-100 for U2OS cells and the P3 kit and program CA-136 for CD34+ cells according to the manufacturer's instructions.

[0233] Genomic DNA isolation and fragmentation Cellular genomic DNA was isolated using a QIAamp DNA Blood Mini Kit (Qiagen). Genomic DNA was sheared to a median length of 400 bp using an M220 focused sonicator (Covaris) according to the manufacturer's instructions.

[0234] Abnoba-Seq workflow All primers used are listed in Table 1. NEBNext Ultra II Kit (NEB) and AMPure XP beads (Beckman Coulter) were used throughout the workflow. Production of end-blocked DNA fragment batches: Sheared genomic DNA (600 ng per reaction) was end-prepared and blocking adapters (112.5 pmol) were ligated using the NEBNext Ultra II Kit (NEB) for 30 min at 20°C. Unblocked fragments were removed by exonuclease treatment with Exonuclease III (200 U, NEB), Exonuclease I (20 U, NEB), and Lambda Exonuclease (50 U, NEB).

[0235] In vitro cleavage: 50 ng of end-blocked genomic DNA fragments were cleaved in vitro with 1.6 pmol of Cas9 protein (lab-made) and 4.8 pmol of gRNA in a volume of 25 μl using 1x NEB buffer 3 and 1x BSA for 1 hour at 37° C. The reaction was stopped by adding 4 μg of RNase and a stop solution containing 50% EDTA 0.5M pH 8.0, 30% glycerol, and 1% SDS, and incubated for 15 minutes at 37° C. A quality control PCR was performed using Q5 polymerase (NEB) to amplify the specific site.

[0236] Biotinylation of cleaved DNA fragments: The ends of the cleaved DNA fragments were prepared and ligated with biotin-adapters (75 pmol) using the NEBNext Ultra II kit. Biotinylated fragments were enriched via Dynabeads MyOne Streptavidin C1 (Thermo Fisher Scientific) according to the manufacturer's protocol. Then, the streptavidin-bound biotinylated DNA fragments were applied to linear PCR amplification using only biotin-adapter specific primers as templates. After 10 cycles, the PCR reaction was stopped at 4°C, the PCR reaction with the linear PCR product was transferred to a new tube, a primer specific to the blocking adapter was added, and PCR was continued for another 10 cycles. The amplified products were applied to DNA library preparations to prepare them for next-generation sequencing.

[0237] Next-generation sequencing: DNA libraries were quantified using ddPCR with a ddPCR Library Quantification Kit for Illumina TruSeq (Bio-Rad). Sequencing was performed on a MiSeq instrument (Illumina) with the MiSeq Reagent Kit v2 (300 cycles).

[0238] Off-target analysis: The first and later steps of sequencing data analysis were performed using Galaxy at usegalaxy.eu (Afgan et al., 2018, Nucleic Acids Res 46(W1):W537-W544). First, adapter sequences from Ilumia-adapters, blocking-adapters, and biotin-adapters were removed from the reads using Trim Galore with default settings. The removed reads were mapped to the human reference genome (hg19) with Bowtie2 with default settings. The aligned sequence reads were then further analyzed by first quantifying total read coverage within a 100 nt sliding window on the human genome using bedtools (Quinlan and Hall, 2010, Bioinformatics 26(6):841-842). Non-empty contiguous windows with coverage >0 were merged into sequence clusters, and empty windows were discarded. Sequence clusters were then split into 4 nt overlapping windows in steps of 2 nt. The first and last nucleotides (nt) of each read were used to calculate start and end coverage, respectively. Reads that did not belong to the start coverage group and did not belong to the end coverage group were used to calculate background coverage. For each sequence group with non-zero start and end coverage, putative cleavage sites were searched for by maximizing start and end coverage and minimizing the distance to the center of the sequence group. False positive sites were filtered out by applying the following criteria: background-to-signal ratio ≤ 0.5, start-to-end ratio ≥ 1 / 5 ~ ≤ 5, site coverage (start and end coverage) ≥ 6, and sequence group coverage ≥ 10. To further remove the remaining false positive sites, a machine learning model was trained on the RNF2 annotated sites based on the features of start-to-end ratio, background-to-signal ratio, sequence group width, relative distance, and site-to-sequence group ratio. Based on the literature and prior knowledge, the RNF2 sites identified by the used pipeline were classified as true or false positive. For two-thirds of these sites, a random forest model was constructed using 3-fold 10 replicate cross-validation to obtain area under the curve 1 for both the training and test sets.The trained model was then used to classify other target sites and filter out predicted false positive sites.

[0239] Visualization: To visualize the identified target sites, the Integrative Genomics Viewer (IGV) was used (Robinson et al., 2011, Nat Biotechnol 29(1):24-26; Thorvaldsdottir et al., 2013, Brief Bioinform 14(2):178-192).

[0240] Sequencing of target amplicons U20S and CD34+ cells isolated from umbilical cord blood were nucleated with VEGFA- and HEK4-targeting RNPs, and genomic DNA from treated and untreated cells was harvested on day 7 after gene transfer. The specific site, as well as the top 20 nonspecific sites of CRISPR-Cas9 nuclease, were amplified by PCR to an amplicon size of approximately 300 bp using Q5 polymerase (NEB). The PCR amplicons were purified using a QIAquick PCR Purification Kit (Qiagen), and the concentrations were measured using a Qubit Fluorometer (Thermo Fisher Scientific). 20 ng of each of the 21 amplicons were mixed per group (untreated / gene-edited, U20S / HSC) and applied to DNA library preparation using the NEBNext Ultra II kit (NEB). DNA libraries were quantified using ddPCR with ddPCR Library Quantification Kit for Illumina TruSeq (Bio Rad) and sequenced (500 cycles) on a MiSeq (Illumina) using MiSeq Reagent Kit v2. After NGS, reads were analyzed using CRISPResso (Pinello et al., 2016, Nat Biotechnol 34(7):695-97) with the following settings: set minimum average quality score to 30 (q30), filter reads based on phred33 quality score, trim Illumina adapter sequences from the reads, quantify insertions / deletions within a 40 nt window around the predicted cleavage site (w40), and ignore substitution events for quantification without considering SNPs. Editing efficiency was calculated using the following formula: Editing efficiency %=(% indel reads in gene edited cells)-(% indel reads in untreated cells).

[0241] Table 1: Oligonucleotides Sequences are shown in the 5' to 3' direction. Modifications: * = phosphorothioate bond, / 5'Phos / = 5' phosphorylation, / 5SpC3 / = 5' C3 spacer, / 3SpC3 / = 3' C3 spacer, / iBiodT / = internal biotin-deoxythymidine, / ideoxyU / = internal deoxyuridine. JPEG0007612225000001.jpg211151JPEG0007612225000002.jpg211150JPEG00076122250 00003.jpg203151JPEG0007612225000004.jpg212151JPEG0007612225000005.jpg138150

[0242] result Example 1: Abnoba-Seq workflow We hypothesized that cleaving fragmented, end-blocked genomic DNA using CRISPR-Cas in vitro would allow for subsequent tagging of the cleavage sites with biotinylated adapters, which would then facilitate enrichment of the cleaved DNA fragments and allow for final identification of the target sites after high-throughput sequencing (Figure 1). To generate a fragmented, end-blocked genomic DNA library, genomic DNA was randomly sheared to a size (median) of approximately 400 bp, and double-stranded (ds) DNA adapters with C3-spacers on both strands were ligated to the end-repaired, dA-tailed DNA fragments. DNA fragments with blocking adapters ligated to both ends, referred to as end-blocked fragments, accounted for more than 99% after treating the fragments with a combination of exonuclease III, exonuclease I, and λ exonuclease (Figure 2). Then, 50 ng of end-blocked DNA fragments corresponding to approximately 15,000 single genomes were cleaved with 1.6 pmol of RNP (SpyCas9 protein complexed with gRNA at a molar ratio of 1:3). The cleavage reaction produced free DNA ends at all target sites to which dsDNA-based affinity adapters with two internal biotinylated dT bases were ligated after end repair and dA tailing. As an in-step control, CRISPR-Cas9 on-target activity was monitored by PCR amplification with primers specific for specific sites. If the specific site was effectively cleaved, the PCR should return only a weak band (Figure 3). If the sample passed the in-step control, the CRISPR-Cas cleaved DNA fragments were enriched using streptavidin-loaded beads, which have a high affinity for biotin. These adapters also added target sequences for the PCR step used as follows: first, to release the biotinylated DNA fragments from streptavidin beads by linear amplification, then to produce dsDNA again, and finally, to fabricate sequencing libraries with Illumina adapters and subject them to paired-end next-generation sequencing (NGS). Importantly, the Abnoba-Seq pipeline is not based on a sequence homology-driven algorithm; as a result, it is entirely independent of the gRNA and PAM sequences.The bioinformatics pipeline relies on pattern recognition algorithms to search for reads that contain specific sequences in the aligned sequence reads, i.e., overlapping start (first 3 bases) and end (last 3 bases). Briefly, the whole genome was scanned to identify genomic regions containing putative cleavage sites in the end sequences. These sites are defined by the presence of multiple reads that either start or end at the same position (+ / - 3 nt). Finally, a machine learning algorithm based on random forests was deployed to remove false positive cleavage sites. The selected sequences were finally sorted according to their coverage (see Methods and Figure 4 for details).

[0243] Example 2: Validation of the Abnoba-Seq workflow To validate the method, we applied Abnoba-Seq to genomic DNA of the human cell line U20S, using four different CRISPR-Cas nucleases previously evaluated by GUIDE-Seq and CIRCLE-Seq (Tsai et al., 2015; Tsai et al., 2017). As expected, alignment of NGS reads to the human reference genome showed the location of the cleavage site with the most reads starting or ending three nucleotides upstream of the PAM in the human genome (Figure 5). We observed other reads that did not start or end at the cleavage site, but were part of a cluster of reads that included the target site. These DNA fragments could potentially start or end at the cleavage site, but were not sequenced in full length due to the maximum sequence read length of 150 nt.

[0244] Bioinformatic analysis revealed 42 target sites for CRISPR-Cas9 nuclease targeting FANCF, 51 sites for RNF2-targeting nuclease, 307 target sites for VEGFA-targeting nuclease, and 233 target sites for CRISPR-Cas9 nuclease targeting HEK site 4 (HEK4) (Figure 6). Note that for Abnoba-Seq, like GUIDE-Seq and CIRCLE-Seq, specific sites often returned the least number of reads. Most identified nonspecific sites showed a high degree of sequence homology to the specific sites. Interestingly, however, in addition to simple mispairing of nucleotides, Abnoba-Seq identified numerous nonspecific sites with bulges in either the gRNA or DNA target sites.

[0245] A side-by-side comparison of nonspecific sites identified by Abnoba-Seq with those previously found by GUIDE-Seq and CIRCLE-Seq showed a reasonable degree of overlap for RNF2-targeting nucleases but not FANCF, VEGFA, and HEK4, respectively (Figure 3). Between 58% and 100% of GUIDE-Seq and 11% and 21% of CIRCLE-Seq nonspecific sites were identified by Abnoba-Seq, respectively (Table 2). Specific sites were identified by all three methods. Meanwhile, a more qualitative comparison showed that the majority of the top 20 identified nonspecific sites were detected by all three methods. This suggests that these three assays can reliably detect frequently cleaved nonspecific sites.

[0246] In summary, Abnoba-Seq can identify commonly cleaved non-specific sites with substantially little genomic input DNA and is a suitable method to evaluate the specificity of target-specific programmable nucleases in precious cell samples with limited sample volume and availability.

[0247] Example 3: Aabnoba-Seq analysis on cell samples with limited sample volume and availability Table 2 JPEG0007612225000006.jpg30141

[0248] One major drawback of existing off-target analysis methods is that they cannot be easily applied to cell samples of limited sample volume and availability. As a proof of concept that Abnoba-Seq can be performed on genomic DNA of clinically relevant cell types with limited availability, we applied Abnoba-Seq to genomic DNA isolated from umbilical cord blood-derived CD34-positive hematopoietic stem and progenitor cells (HSCs). As visualized in the in-step control (Figure 3B), CRISPR-Cas9 nuclease targeting VEGFA showed high on-target cleavage activity, and Abnoba-Seq results for HSCs were comparable to previous ones performed on genomic DNA of U20S cells (Figure 6). A more detailed comparison revealed that the majority of identified sites were common between the two samples (Figure 8B). Considering the top 20 or top 50 target sites identified on the genomic DNA of HSCs, 95% (19 / 20) or 92% (46 / 50) were identified on the genomic DNA of U20S cells, respectively (Figure 8C, D), indicating that the reproducibility of this assay is high. In summary, Abnoba-Seq is a robust method that can be successfully applied to genomic DNA isolated from precious cell samples with limited sample volume and availability.

[0249] Example 4: Validation of CRISPR-Cas off-target activity in cells Analysis of in vitro activity of target-specific programmable nucleases reverts to the worst-case scenario, since (i) all potential genomic non-specific sites in vitro are easily accessible to CRISPR-Cas9, and (ii) nuclease concentrations can be much higher in vitro than in the nuclei of transduced cells. Therefore, the top 20 Abnoba-Seq predicted non-specific sites were analyzed for VEGFA and HEKs4 targeting CIRPSR-Cas nucleases by sequencing of targeted amplicons 7 days after transduction of U20S cells or primary HSCs. Overall editing efficiency, as indicated by indel frequencies, was 96% for VEGFA-specific sites in U20S cells, 61% for VEGFA-specific sites in HSCs, and 69% for HEKs4-specific sites in HSCs (Figure 9). Overall off-target activity in these cells was relatively low. The highest off-target activity was observed at VEGFA non-specific site (OT) 8 (20% edited allele in U20S cells and 10% edited allele in HSCs) and VEGFA OT-13 (5% edited allele in HSCs). Off-target activity at other predicted non-specific sites was low or absent. Some of the non-specific sites were only cleaved in U20S cells and not in primary HSCs (OT-12, OT-14, OT-16), suggesting cell type specificity. In summary, only about half of the non-specific sites identified in vitro were actually cleaved in cells. It should be noted that the majority of these examined sites were also predicted to be non-specific sites by CIRLCE-Seq and GUIDE-Seq. Also, importantly, the in vitro method for genome-wide and unbiased detection of non-specific sites is merely a predictive tool, although complex.

[0250] While many genome-wide methods for identifying CRISPR-Cas-related off-target activity are seemingly unbiased, bioinformatics pipelines for analyzing NGS data do introduce some bias, particularly with respect to the maximum number of mismatched base pairs allowed or the requirement for PAMs. Importantly, Abnoba-Seq, which identifies nonspecific sites, does not make any assumptions about the nature of the nonspecific sites, the degree of homology to specific sites, or the presence of specific PAMs. Thus, Abnoba-Seq identifies nonspecific sites that are excluded by other methods, including RNF2 OT7 with seven mismatched base pairs and two bulges, RNF2 OT37 with nine mismatched base pairs, VEGFA OT2 with three mismatched base pairs and one TGA PAM, VEGFA OT26 with three mismatched base pairs and one TGA PAM, or VEGFA OT75 with two mismatched base pairs and one TTG PAM. In summary, Abnoba-Seq is a truly unbiased in vitro assay to reveal CRISPR-Cas non-specific activity in vitro and in vivo. Because only minimal genomic DNA is required to perform the assay, Abnoba-Seq can be applied in a cell type-specific manner to precious cell samples with limited sample volume and availability.

Claims

1. 1. A method for detecting off-target sites of one or more target-specific programmable nucleases in a genome in vitro, the method comprising the steps of: (a) providing an input sample comprising genomic DNA, randomly fragmenting the genomic DNA into dsDNA fragments, ligating blocking adaptors to the ends of the dsDNA fragments, and removing dsDNA fragments containing unblocked ends to prepare purified end-blocked dsDNA fragments; (b) preparing a DNA library of cleaved and amplified dsDNA fragments by: b1) cleaving the purified end-blocked dsDNA fragments with one or more target-specific programmable nucleases to obtain cleaved dsDNA fragments; b2) ligating an affinity adaptor to the cleavage site of the cleaved dsDNA fragment to obtain an affinity adaptor-modified dsDNA fragment; b3) enriching the affinity adaptor-modified dsDNA fragments using the affinity adaptors to obtain enriched affinity adaptor-modified dsDNA fragments; and b4) The enriched affinity adaptor-modified dsDNA fragments are amplified by PCR amplification to obtain a DNA library of cleaved and amplified dsDNA fragments. (c) sequencing the DNA library of the cleaved and amplified dsDNA fragments to obtain sequence reads; (d) aligning the sequence reads to a reference genome to obtain aligned sequence reads; and (e) determining putative cleavage sites and identifying cleavage sites by eliminating false positive cleavage sites, wherein putative cleavage sites are determined by searching for sequence regions in the reference genome where both the sequence read start and sequence read end portions of aligned sequence reads match, and identifying putative cleavage sites in these sequence regions as sites where the sequence read start and sequence read end portions match.

2. 2. The method of claim 1, wherein the input sample contains less than 1 μg of genomic DNA.

3. Step (a) a1) randomly fragmenting the genomic DNA to defined lengths to provide dsDNA fragments; a2) protecting the ends of the dsDNA fragments by ligating blocking adaptors to obtain end-blocked dsDNA fragments, 1) preparing dsDNA fragments for blocking adaptor ligation; and 2) ligating blocking adaptors to the ends of the dsDNA fragments to obtain end-blocked dsDNA fragments. and a3) contacting the sample containing end-blocked dsDNA fragments with one or more exonucleases to remove non-end-blocked dsDNA fragments to obtain purified end-blocked DNA fragments. The method of claim 1 or 2, comprising:

4. 4. The method of any one of claims 1 to 3, wherein the blocking adaptor comprises one or more modifications that prevent ligation of the end of the blocking adaptor joined to the dsDNA fragment after the blocking adaptor has been ligated to the dsDNA fragment and one or more modifications that prevent exonuclease cleavage of the dsDNA fragment from the end of the blocking adaptor joined to the dsDNA fragment after the blocking adaptor has been ligated to the dsDNA fragment.

5. 5. The method of claim 1, wherein in step b1), the one or more target-specific programmable nucleases are selected from transcription activator-like effector nucleases (TALENs), Zn finger nucleases (ZFNs), meganucleases, megaTALs, Fokl-dCas9, clustered regularly interspaced short palindromic repeats (CRISPR) / Cas RNA-guided nucleases (CRISPR / Cas RGNs) or variants thereof, Argonaute-family protein-based DNA-guided DNA nucleases or variants thereof, or combinations thereof.

6. A method according to any one of claims 1 to 4, wherein in step b1), the one or more target-specific programmable nucleases are CRISPR / Cas RGN, wherein the Cas endonuclease of the CRISPR / Cas RGN is Cas9 endonuclease or a mutant of Cas9 endonuclease, or Cpf1 endonuclease or a mutant of Cpf1 endonuclease.

7. The method of any one of claims 1 to 6, wherein the blocking adapter and / or the affinity adapter are dsDNA-based adapters and contain a primer site suitable for use in PCR priming.

8. 8. The method according to claim 1, wherein the affinity adapters are dsDNA-based adapters comprising one or more biotin or biotin-based modifications, and in step b3), streptavidin or a variant thereof immobilized on magnetic beads is cleaved and used as an affinity adapter binding molecule for enriching the affinity adapter-modified dsDNA fragments.

9. Step b4) for obtaining a DNA library of the cleaved and amplified dsDNA fragments, 1. A first PCR amplification step using enriched affinity adaptor-modified dsDNA fragments bound to a solid support via an adaptor-binding molecule as a template and a primer specific to the affinity adaptor for linear PCR amplification; and 2. A second PCR amplification step in which, after the first step, the enriched affinity adaptor-modified dsDNA fragments bound to the solid support via the adaptor-binding molecule are removed from the reaction mixture, and a primer specific to the blocking adaptor is added as a second primer. The method according to any one of claims 1 to 8, comprising a two-step PCR amplification of:

10. In step (d), first, d1) removing blocking adaptors, affinity adaptors, and any sequencing adaptor sequences from the sequence reads to obtain trimmed sequence reads; and then d2) mapping the trimmed sequence reads to a reference genome to obtain aligned sequence reads; The method of any one of claims 1 to 9, wherein the sequence reads are aligned to a reference genome to obtain aligned sequence reads.

11. The method according to any one of claims 1 to 10, wherein the genomic DNA is human genomic DNA and the reference genome is human reference genome hg19 or hg38.

12. In step (e), the putative cleavage site is - dividing the reference genome containing the aligned sequence reads into groups of sequences of defined length; and - for each sequence group comprising at least one sequence read start and at least one sequence read end, one or more putative cleavage sites are determined as sites with a local maximum of the sum of sequence read start and sequence read end, optionally measured by minimizing the distance to the center of the sequence group, wherein a sequence read start consists of the first 1-5 nt of the aligned sequence reads and a sequence read end consists of the last 1-5 nt of the aligned sequence reads.

13. In step (e), the cleavage site is e1) dividing the reference genome containing the aligned sequence reads into sequence groups of a set sequence length of 100 nt; e2) measuring read coverage for each set of sequences, discarding sequences with zero read coverage, and combining sequences with consecutive read coverage >0 into a consecutive set of sequences based on their sequence order in the reference genome; e3) dividing each set of sequences into windows of a set width of 4 nt with a set step size equal to or smaller than the width of the windows of 2 nt; e4) Measure the beginning, end, and background coverage of each window by grouping as follows: - grouping sequence read start portions falling within said window regions into start coverage groups; - grouping sequence read ends falling within said window regions into end-covering groups; - grouping sequence reads that fall within the window region but do not belong to the start and end coverage groups into a background coverage group; For each window, add up the number of reads in the start coverage group, the end coverage group, and the background coverage group; e5) for each group of sequences that includes at least one sequence read start and at least one sequence read end, determining one or more putative cleavage sites by maximizing the sum of start and end coverage and minimizing the distance to the center of the group of sequences; and e6) The method according to any one of claims 1 to 12, wherein false positive cleavage sites are excluded and a sequence read start is identified by consisting of the first 1 to 5 nt of an aligned sequence read and a sequence read end is identified by consisting of the last 1 to 5 nt of an aligned sequence read.

14. False positive cleavage sites:

1. The ratio of background coverage to signal coverage does not exceed a threshold of 0.5; and / or 2. the ratio of starting coating to end coating at the putative cleavage site is within a numerical range between 1 / 5 and 5 / 1; and / or 3. A minimum site coverage of 6 or greater; and / or 4. The method of any one of claims 1 to 13, wherein putative cleavage sites are eliminated by selecting only those with a minimum sequence group coverage of 10 or more.

15. A method for detecting non-specific sites in a genome in vivo, comprising the steps of: i) providing a first sample comprising genomic DNA of cells that have been treated with one or more target-specific programmable nucleases, and a second sample comprising genomic DNA of cells of the same genetic background that have not been treated with said one or more target-specific programmable nucleases; ii) detecting in vitro non-specific sites in each of the two samples by using one or more target-specific programmable nucleases of step i) in the method of claims 1 to 14; and iii) comparing the non-specific sites of the two samples and identifying non-specific sites that are identified only in the sample containing genomic DNA of cells that have not been treated with the one or more target-specific programmable nucleases, but not in the sample containing genomic DNA of cells that have been treated with the one or more target-specific programmable nucleases, as in vivo non-specific sites of the one or more target-specific programmable nucleases.

Images