{"paper_id":"14eeea81-5eec-4cdf-9bf6-14a564b7d035","body_text":"1 \n \nHigh-throughput robotic isolation of human iPS cell clones reveals frequent \nhomozygous induction of identical genetic manipulations by CRISPR-Cas9 \n \nAuthors \nGou Takahashi 1,†, Minato Maeda 1,2, Kayoko Shinozaki 1,2, Gakuro Harada 3, Saburo Ito 3, \nYuichiro Miyaoka1,2,4,* \n \n1Regenerative Medicine Project, Tokyo Metropolitan Institute of Medical Science, Setagaya, \nTokyo 156-8506, Japan \n2Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, \nBunkyo, T okyo 113-8510, Japan \n3Yamaha Motor Co., Ltd., Iwata, Shizuoka 438-8501, Japan  \n4Graduate School of Humanities and Sciences, Ochanomizu University, Bunkyo, T okyo \n112-8610, Japan \n \n†Current address: Department of Animal Science, Tokyo University of Agriculture, Atsugi, \nKanagawa 243-0034, Japan \n*Correspondence: miyaoka-yi@igakuken.or.jp\n (Y.M.), X account: @YuichiroMiyaoka \n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n2 \n \nSUMMARY \nGenome editing in human iPS cells is a powerful approach in regenerative medicine. \nCRISPR-Cas9 is the most common genome editing tool, but it often induces byproduct \ninsertions and deletions in addition to the desired edits. Therefore, genome editing of iPS \ncells produces diverse genotypes. Existing assays mostly analyze genome editing results in \ncell populations, but not in single cells. However, systematic profiling of genome editing \noutcomes in single iPS cells was lacking. In this study, we developed a method for \nhigh-throughput iPS cell clone isolation based on the precise robotic picking of cell clumps \nderived from single cells grown in extracellular matrices. We analyzed over 1,000 \ngenome-edited iPS cell clones and found that homozygous editing was much more frequent \nthan heterozygous editing. We also observed frequent homozygous induction of identical \ngenetic manipulations, including insertions and deletions. Our new cloning method and \nfindings will facilitate the application of genome editing to human iPS cells.  \n \nINTRODUCTION \nGenome editing has revolutionized our ability to study and cure genetic disorders (Gaj et al., \n2013). In particular, genome editing in induced pluripotent stem (iPS) cells allows the \ndevelopment of human cell-based isogenic disease models and potential cell therapies \n(Hockemeyer and Jaenisch, 2016). Clustered regularly interspaced short palindromic \nrepeats (CRISPR)-associated protein 9 (Cas9), which relies on its ability to cleave genomic \nDNA with target sequences, is the most widely used genome editing tool (Jinek et al., 2012). \nThe double-strand breaks at target sites induced by CRISPR-Cas9 mainly evoke two DNA \nrepair pathways. One is non-homologous end-joining (NHEJ), which joins two broken ends \nof DNA with diverse insertions and deletions (indels) at the joined sites. The other is \nhomology-directed repair (HDR), which repairs broken DNA by recombination between \ngenomic DNA and template DNA with sequence homology (Gaj et al., 2016). Therefore, we \ncan achieve precise genetic manipulation via HDR by providing cells with donor DNA with \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n3 \n \nthe intended sequences. In general, NHEJ is much more frequent than HDR, and both HDR \nand NHEJ can be concurrently induced in the same cell. As a result, even when we attempt \nto induce specific genetic manipulations via HDR in iPS cells, the resulting clones have \ndiverse genotypes. Therefore, it is important to understand the diverse allelic combinations \nand frequencies of genome editing in iPSCs. \nDespite the importance of monitoring genome editing outcomes in individual cells, most \nassays analyze genome editing results in cell populations, but not in single cells, for \nexample, the T7E1 assay and pooled amplicon sequencing (Germini et al., 2018). As a new \napproach, we previously isolated more than 2,600 clones of genome-edited human cultured \ncells (HEK293T, HeLa, and PC9 cells) using an automated single-cell dispensing system \n(Takahashi and Miyaoka, 2023). By analyzing the genotypes of these isolated clones, we \nfound a strong binary tendency of genome editing induced by CRISPR-Cas9; that is, \nindividual cells are often either not edited at all or all target alleles are fully edited (Takahashi \net al., 2022). \nHowever, owing to the high mortality, we could not apply the same single-cell dispensing \nsystem to human iPS cells as human cultured cell lines. Efficient methods to isolate human \niPS cell clones are in high demand, not only for the analysis of genome editing results, but \nalso for the isolation of iPS cell lines with desired genetic manipulations. Recent studies \nhave reported promising additives to avoid cell death in iPS cells, but they are not yet \nconclusive (Chen et al., 2021). Therefore, in this study, we developed a method that utilizes \na cell-handling robot to efficiently isolate a large number of clones from genome-edited iPS \ncell pools. Using this method, we obtained more than 1,000 genome-edited iPS cell clones \nand analyzed their genotypes. We found that the same genetic manipulations (HDR and \nvarious indels) were homozygously induced in human iPS cells. Our new approach to \nefficiently isolate human iPS cell clones and profiles of genome editing outcomes in human \niPS cells will greatly contribute to regenerative medicine. \n \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n4 \n \nResults \nDevelopment of an efficient robotic isolation method for genome-edited iPS cell \nclones grown in Matrigel domes \nFirst, we developed a method to isolate clones from genome-edited iPS cells using a \ncell-handling robot, CELL HANDLER (Yamaha Motor) (Figure 1A). In this method, iPS cells \nwere transfected with a plasmid (px459-HypaCas9) to express HypaCas9 and \nsingle-stranded donor DNA to induce a pathogenic point mutation (ATP7B R778L, GRN \nR493X, or RBM20 R636S) (Kato-Inui et al., 2018). The puromycin-resistant gene was \nco-expressed with HypaCas9 via the T2A peptide so that the transfected cells were selected \nusing puromycin and then dispersed into single cells. We formed domes composed of \nMatrigel and allowed these single iPS cells to grow to maximize the number of single cells \ngrowing in a well of a 6-well plate (Figure 1B and 1C). Placing cells in a 3-dimensional \nstructure also made the robotic cell-picking process less damaging and more efficient than \nwith 2-dimensional cultures. Culturing these single cells for approximately 1 week resulted in \nthe formation of cell clumps with a diameter of 100\n– 150 μ m, which were suitable for being \npicked by CELL HANDLER (Figure 1D). CELL HANDLER captured multiple images with an \nautomatic focus to scan for cell clumps throughout the Matrigel domes. Specialized software \nprocessed these images to recognize and distinguish cell clumps by major diameter, \ncircularity, and neighbor distance (Table S1). We measured these criteria for typical cell \nclumps derived from single cells (Figure 1D). Cell clumps that met these criteria were \nselected and transferred into a new 96-well plate for expansion in canonical \ntwo-dimensional culture. After 2-4 weeks, we isolated genome-edited iPS cell clones that \nmaintained the expression of pluripotency marker genes (Figure S1A). The overall cloning \nefficiency was approximately 60.1% (1372 clones out of 2255 wells in 36 96-well plates) \nusing CELL HANDLER (Table 1 and Figure S1B).  \n \nAssessment of clonality of iPS cells isolated by robotic picking \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n5 \n \nTo assess the clonality of the isolated iPS cells by robotic picking, we established iPS cell \nlines expressing EGFP, mCherry, or EBFP. We then mixed these three cell lines and \nperformed robotic picking (Figure S1C). We repeated this cloning process twice and isolated \n28 and 43 clones. We confirmed that none of the clones contained 2 or more colors (Table \nS2). We then analyzed the number of alleles in isolated iPS cells after editing RBM20, GRN, \nand ATP7B. We investigated the genome editing outcomes by amplicon sequencing as \npreviously reported (Takahashi et al., 2022), and classified the resulting alleles into 4 types: \nwild-type (WT), NHEJ, HDR, and NHEJ+HDR, using CRISPResso2 (Clement et al., 2019). \nThis allelic classification revealed that 96.3% of the isolated clones had 1 or 2 allelic types \non average at the 3 target sites (Figure 1E). These results suggest that the majority of the \nisolated clones were derived from single cells. In contrast, 3.7% of clones harbored three or \nmore allelic types, and these clones were excluded from subsequent analyses. \n \nHomozygous NHEJ is the most common genotype in genome-edited iPS cells \nBecause iPS cells are diploid, there are 10 different genotypes: WT/WT, WT/NHEJ, \nWT/HDR, WT/NHEJ+HDR, NHEJ/NHEJ, NHEJ/HDR, NHEJ/NHEJ+HDR, HDR/HDR, \nHDR/NHEJ+HDR, and NHEJ+HDR/NHEJ+HDR. We classified the genotypes of isolated \niPS clones and found that NHEJ/NHEJ, WT/WT, and WT/NHEJ were the most common, \nsecond, and third most common genotypes, respectively, in all 3 target genes (Figure 2A).  \nWe previously reported that genome editing in cultured cells, such as HEK293T, \noccurred in a binary manner; that is, all targeted sequences were either edited or not edited \n(Takahashi et al., 2022). Therefore, we examined whether the genome editing of iPS cells \nalso occurred in a binary manner. For this purpose, we investigated the proportion of WT/WT, \nWT/NHEJ, and NHEJ/NHEJ clones. We created diagrams to visualize the genotypes of the \nisolated clones (Figure 2B, S2A, and S2B). We also generated model diagrams to represent \ngenotypes of the same number of clones if all allele types were randomly distributed, as \ndescribed previously (Takahashi et al., 2022). We obtained the overall frequencies of the WT, \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n6 \n \nNHEJ, HDR, and NHEJ+HDR alleles from the genotypes of the isolated clones (Figure 2C, \nS2C, and S2D). These overall allelic frequencies were nearly identical to those of the pooled \npopulations of genome-edited iPS cells before clone isolation, suggesting that our robotic \nclone-picking method does not cause any bias in the genotypes of the isolated cells (Figure \n2A and S3). We uniformly and randomly redistributed these alleles in the same number of \nclones as the isolated ones, which served as models for comparison (Figure 2D, S4A, S4B, \nand S4C). Relative to these models, we generally observed more isolated WT/WT and \nNHEJ/NHEJ clones, although the differences were not statistically significant for RBM20 \nand ATP7B (Figure 2E). The proportion of WT/NHEJ clones was significantly lower in the \nisolated clones than in the model for all three edited genes, indicating that genome editing in \nhuman iPS cells was also binary (Figure 2F).  \n \nFrequent homozygous induction of identical indels by NHEJ in iPS cells  \nSo far in our study, we have classified all different indels into a single category: “NHEJ”. \nAlthough NHEJ induced diverse indels, there were several specific indels with notable \nfrequencies. Therefore, we re-genotyped the isolated genome-edited iPS clones by \ndistinguishing the frequent indels in each gene.  \n We re-analyzed the amplicon sequencing data, identified the top eight most \nfrequent indel alleles, and classified 12 alleles in the isolated iPS clones genome-edited in \nRBM20, GRN, and ATP7B, which represented approximately 70% of the overall allelic \nfrequency (Figure 3A, 3B, S5A-B, S6B, and S6D). We found that many iPS clones were \nhomozygous for identical indels. In particular, several clones were homozygous for an indel \nat a low frequency (Figure S6A, S6C, and S6E). For example, we isolated one clone with a \nhomozygous 1-bp deletion in RBM20 No.3 (overall frequency was only 1.3%) and two \nclones with a homozygous 11-bp deletion from RBM20 No.1 and 3 (1.4%) (Figures 3A and \n3B). Therefore, we compared the proportion of clones with homozygous alleles for the 3 \ngenes. We examined the frequencies of iPS cell clones homozygous for the top three \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n7 \n \nfrequent indel alleles in all three target genes and found that they were all higher than those \nof the models, although some of them were not statistically significant (Figures 3C, 3D, S6A, \nS6C, and S6E). Overall, genome editing in iPS cells tends to result in identical sequence \nmanipulations in both copies of the target sequences in single cells. \n \nRobotic isolation of genome-edited iPS cell lines with rare genotypes  \nBecause our new method based on robotic handling of cells allowed high-throughput \nisolation of genome-edited iPS cell clones, we tested whether this method allowed us to \nestablish iPS clones with rare genotypes. As shown in Figure 3B, genome editing in iPS \ncells produced mainly non-edited cells or cells homozygous for NHEJ, and cells with \nheterozygous genotypes were relatively rare. Therefore, we characterized 2 clones with \nheterozygous NHEJ in GRN, whose mutations are associated with frontotemporal lobar \ndegeneration (Chen-Plotkin et al., 2011). Based on our amplicon sequencing analysis, 1 \nclone (GRN-Het 1-bp ins) had a heterozygous 1-bp insertion, and the other clone (GRN-Het \nΔ19-bp) had a heterozygous 19-bp deletion (Figure 4A). Both mutations were expected to \ncause nonsense-mediated decay owing to frameshift mutations (Figure 4A). Because we \nextracted genomic DNA from isolated clones while freezing a portion of these clones, we \nthawed and recovered these two clones. Sanger sequencing and digital PCR confirmed that \nthe recovered clones maintained the identified heterozygous mutations (Figure 4B, 4C, and \nS7). In addition, these clones maintained the expression of the pluripotency markers SOX2 \nand OCT4 (Figure 4D). These results indicate that our iPS cell cloning method allows the \nestablishment of clones with rare genotypes, which is highly useful for genome editing in iPS \ncells to study and cure diseases. \n \nDiscussion \nSince complete control of genome editing has not yet been achieved, we must rely on \nsingle-cell cloning to establish genome-edited iPS cell lines with desired genotypes. In this \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n8 \n \nstudy, we achieved high cloning efficiency using iPS cell culture in a 3-dimensional Matrigel \ndome and robotic isolation of cell clumps using CELL HANDLER. Small molecule additives, \nsuch as the CEPT cocktail, have been developed to avoid cell death induced by the \nsingularization of human iPS cells to enhance single-cell cloning (Tristan et al., 2023). \nTherefore, a combination of our robotic handling method and these small-molecule additives \nmay further enhance the single-cell cloning of iPS cells.  \nWe previously reported that genome editing in cultured human cells is induced in a binary \nmanner, where all target alleles are either completely edited or not edited at all. We found \nthis dichotomous effect in human iPS cells as well, as we observed fewer WT/NHEJ \nheterozygously edited clones relative to the mathematical expectation (Figure 2F). Patients \nwith some genetic disorders are heterozygous for mutations. Our findings indicate that \nestablishing human iPS cells with these heterozygous mutations using CRISPR-Cas9 is \nchallenging. Recently, Kawamata et. al. reported the same binary trend in genome editing \nand demonstrated that decreasing cleavage activity allows for heterozygous editing \n(Kawamata et al., 2023). Therefore, to achieve heterozygous editing of human iPS cells, it \nmay be necessary to lower CRISPR-Cas9 activity. \nIn this study, we investigated and distinguished indels with different sequences. Deletions \nwere the predominant genome editing outcome for all target genes (Figures 3A, S5A, and \nS5B). We also observed 1-bp insertions in the cleavage regions of the three genes. As \nreported in previous studies, these 1-bp insertions presumably occur because Cas9 \noccasionally cleaves DNA sequences with single-base protrusions (Chakrabarti et al., 2019; \nLemos et al., 2018; Shen et al., 2018). Moreover, many of the observed deletions were \nderived from microhomology-mediated end-joining (MMEJ), as previously reported by other \ngroups (Chen et al., 2019; Kim et al., 2018; van Overbeek et al., 2016; Shen et al., 2018). \nThe most frequent 8-bp deletion in RBM20 was a typical example of MMEJ, in which a 5-bp \nmicrohomology (CCGGT) mediated the deletion (Figure 3A). \nInterestingly, these NHEJ (including MMEJ) alleles were found to be homozygously induced \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n9 \n \nin many clones (Figure 3D). The DNA repair mechanism is cell cycle-dependent and NHEJ \nis typically active throughout the cell cycle, whereas HDR is only active in the S/G2 phase \n(Danner et al., 2017). One possibility is that 1 of the 2 alleles was first edited by NHEJ and \nthe other allele was edited by HDR using the NHEJ allele as a homologous template, \nresulting in 2 identical NHEJ alleles. Further studies are needed to confirm this hypothesis.  \nIn summary, our study provides new insights into genome editing in human iPS using \nCRISPR-Cas9. Single-cell cloning remains the most straightforward strategy for isolating \ngenome-edited iPS cell clones. However, the manual selection of cell colonies requires \nconsiderable work and time. Moreover, numerous culture plates are required to grow iPS \ncell colonies for picking, while keeping these colonies separated from each other in \n2-dimensional culture. In our method, CELL HANDLER takes the burden of colony picking, \nand the 3-dimensional culture allows a large number of clones to grow as cell clumps. These \nadvantages enable the large-scale establishment of genome-edited iPS clones, which will \nbe highly valuable for studying genetic disorders and developing gene therapies using iPS \ncells. \n \nExperimental Procedures \n \nResource availability \nCorresponding author  \nFurther information and requests for resources and reagents should be directed to and will \nbe fulfilled by the corresponding author (miyaoka-yi@igakuken.or.jp). \n \nMaterials availability \nAll reagents and materials used in this manuscript are available upon request or prepared \nfor availability from commercial sources. GRN  heterozygous knockout iPS cell lines \ngenerated in this study will be made available on request, but we may require a payment \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n10 \n \nand/or a completed material transfer agreement if there is potential for commercial \napplication. \n \nData and code availability \nThe amplicon sequencing data analyzed in this study will be shared upon request. Raw \namplicon sequencing data are available at the DDBJ Sequence Read Archive. \n \nPlasmids and single-stranded DNA (ssDNA) donor \nThe px459-HypaCas9 plasmid used in this study to express HypaCas9 and the \npuromycin-resistant gene has been described previously (Kato-Inui et al., 2018) (Addgene \nPlasmid #108294). The single-stranded DNA donors and gRNAs used in this study have \nalso been reported previously (Takahashi et al., 2022). The sequences are listed in Tables \nS3.  \n \nMaintenance of iPS cells \nWTC11 iPS cells (Coriell Institute for Medical Research, GM25256) were used for all \nexperiments. iPS cells were maintained on thin-coated GFR Matrigel Matrix (Corning, \n356231) in mTeSR Plus (STEMCELL Technologies, ST-100-0276) medium supplemented \nwith 1% penicillin-streptomycin (P/S) (Nacalai tesque, 26253-84). Cells were dissociated \nusing the Accutase (Nacalai tesque, 12679-54), to passaged wells, we added Y-27632 (final \nconcentration 10 uM), a Rho-associated kinase inhibitor (Ri) (FUJIFILM Wako Pure \nChemical, 034-24024), to promote cell survival.  \n \nTransfection \nWTC11 iPS cells were seeded at 4x10^4 cells/well in a Matrigel-coated 24-well plate. The \nnext day, at least one hour before transfection, the medium was replaced with 500 \nμ L of \nfresh mT eSR Plus with P/S and Ri. Lipofectamine Stem (Thermo Fisher, STEM00003) was \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n11 \n \nused to transfect 400 ng/well of px459-HypaCas9 and 100 ng/well of ssDNA according to \nthe manufacturer’s instructions. After 1 and 2 days, the medium was replaced with 500 μ L of \nfresh mTeSR Plus with P/S. At 4-6 days after transfection, when the cells reached \nconfluence, we dissociated the iPS cells into single cells and embedded them into Matrigel \ndomes. \n \nFormation of Matrigel domes containing single iPS cells \nGFR Matrigel (Corning, 356231) was kept on ice until use. Confluent iPS cells on a 24-well \nplate were detached from the plate using 100 \nμ L/well of Accutase. Then, 400 μ L of PBS was \nadded to suspend the cells, which were centrifuged at 300 × g for 3 min. The supernatant \nwas removed and 1 mL of mTeSR Plus with P/S and Ri was added to gently suspend the \ncells. After counting the number of cells using Countess II (Thermo Fisher Scientific), the cell \nsuspension was diluted to 10-30 cells/\nμ L. Matrigel (400 μ L) was dispensed into a new 1.5 \nmL tube on ice, to which 100 μ L of diluted cell suspension was added and mixed gently. \nThen, 50 μ L of the cell and Matrigel mixture was aliquoted into each well of a 6-well plate \n(5-6 domes/well) (Figure 1B). Immediately after placing aliquots, the bottom of the plate was \nwarmed by hand for 2-3 min to allow the domes to start to gel. The plate was then inverted \nand kept at 37°C for 2 h to solidify the Matrigel domes. After solidification, 3 mL of mTeSR \nPlus with P/S and Ri were added and kept in a CO\n2 incubator.  \n \nRecognition, picking, and seeding of iPS cell clumps by CELL HANDLER \nRecognition, picking, and seeding of iPS cell clumps by CELL HANDLER (Yamaha Motor) \nwhen the cell clumps reached a size of 100-200 \nμ m diameters. CELL HANDLER acquired a \ntotal of 28 focus images in the Z-axis for a Matrigel dome to select cell clumps to pick. If cell \nclumps matched the parameters for picking, CELL HANDLER carried out the picking and \nseeding (Figure 1D and Table S1). The picked cell clumps were automatically transferred to \nthe wells of a Matrigel-coated 96-well plate containing 100 \nμ L of preheated mT eSR Plus with \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n12 \n \nP/S and Ri medium. CELL HANDLER captured images of the cell clumps before and after \npicking to verify the successful acquisition of cell clumps from the Matrigel domes. \nSupplemental movies (Video S1) show picking processes using CELL HANDLER. \n \nExpansion and freezing of isolated iPS cell clones \niPS clones were cultured for 2-3 weeks after robotic picking to ensure sufficient clone \nexpansion. The medium was removed, and 30 \nμ L/well of Accutase was added to detach and \nresuspend the cells. Half of the cell suspension was transferred to a new 96-well plate and \nused for genome extraction as described previously (Miyaoka et al., 2014). The other half of \nthe cell suspension in the original 96-well plate was mixed with 75 μ L of 10% dimethyl \nsulfoxide (DMSO) and 90% fetal bovine serum. The mixtures were layered with 75 μ L of \nmineral oil, and the plate was sealed with Parafilm for cryopreservation at -80°C. The two \nclone lines in Fig. 4A were selected by amplicon sequencing analysis as described below, in \nwhich clones with heterozygous knockout of the GRN gene were selected. To confirm that \nthese clones were pure, genotypes were identified by quantifying allele frequencies using \ndigital PCR, as described previously (Miyaoka et al., 2014). \n \nPreparation of multiplexed amplicon sequencing libraries \nMultiplexed amplicon sequencing libraries were prepared using 2 rounds of PCR, as \ndescribed previously (Takahashi et al., 2022). After the second PCR, the amplified DNA \nfragments from the pooled 48 samples were purified by gel extraction using the NucleoSpin \nGel and PCR Clean-up Midi kit (TaKaRa, 740986.20). We repeated this DNA purification \nprocess every 48 samples. The purified DNAs was mixed in equal molar ratios, and the \nlibrary was prepared according to Illumina's instructions. DNA concentrations of the mixed \nlibraries were quantified using the GenNext NGS Library Quantification Kit (Toyobo, \nNLQ-101). After quantification, PhiX Control v3 (Illumina, FC-110-3001) was added at a final \nconcentration of 20% for amplicon sequencing. Sequencing was performed with MiSeq \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n13 \n \n(Illumina) using the MiSeq v2 reagent kit (Illumina, MS-102-2003) or MiSeq Reagent kit v2 \nNano (Illumina, MS-103-1001) according to the manufacturer’s instructions. \n \nAmplicon sequencing data analysis and allele classification \nFastq files generated by MiSeq were imported into CLC Genomics Workbench (QIAGEN). \nAdapter sequences were removed and demultiplexed using the DNA Index. The data were \nthen analyzed using CRI SPResso2 (https://git hub.com/pinellolab/CRISPResso2) in \nCRISPResso Batch mode (Clement et al., 2019). CRISPResso2 was installed as \nrecommended using a Docker containerization system. In this study, all reads identified as \nambiguous by a CRISPResso2 analysis were classified as NHEJ. \nMathematical models for the distribution of clone genotypes, assuming that the WT, NHEJ, \nHDR, and HDR+NHEJ alleles were randomly induced in all target alleles, were built by \ndistributing these edited alleles to the isolated iPS cell clones at their observed overall \nfrequencies, as described previously (Takahashi et al., 2022). We also distinguished the \nmajor NHEJ alleles based on amplicon sequencing data. These 8 NHEJ alleles, together \nwith WT, HDR, and HDR+NHEJ alleles, accounted for approximately 60-70% of the total \nalleles. Other NHEJ alleles were grouped and labeled \"NHEJ.other\". \n \nImmunocytochemistry \niPS cells were gently washed in PBS and fixed in 4% PFA (EMS; 50-980-487) for 15 min. \nSamples were washed 3 times in PBS and incubated with blocking buffer (3% goat normal \nserum and 0.2% Triton-X in PBS) for 1 h at room temperature. Samples were then \nincubated overnight with the primary antibody (Table S4) at the appropriate concentration in \nprimary blocking buffer at 4°C overnight, washed 3 times in DPBS, and incubated with \nsecondary antibodies (Table S4) diluted in blocking buffer at room temperature for 1 h. The \nsamples were incubated with Hoechst 33342 (Nacalai tesque, 04929-82) for 10 min at room \ntemperature and washed twice with PBS. Fluorescent images were captured using a \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n14 \n \nKeyence BZ-X800 all-in-one microscope (Keyence) and analyzed using BZ-X Analyzer \nsoftware (Keyence).  \n \nStatistics \nTransfection was performed in triplicate (three biological replicates). Values are displayed as \nthe mean ± standard error (S.E.). Statistical significance between the 2 groups was \nassessed by a non-paired 2-tailed Student’s t-test and is displayed in the figures with \nasterisks as follows: *p < 0.05; **p < 0.01. NS: not significantly different (p > 0.2). \n \nSupplemental Information \nSupplemental information can be found online. \nFigures S1-S7 \nTables S1-S4 \nVideos S1 \n \nAuthor Contributions \nG.T. and Y .M. conceived of the study and designed the experiments. G.T. and M.M. \ntransfected iPS cells and encapsulated a single iPS cell in the Matrigel dome. G.H. \nperformed pilot experiments to set up the parameters for CELL HANDLER. S.I. performed \ncell clamp picking using CELL HANDLER presented in this study. G.T., M.M., and K.S. \nclassified the genotypes obtained using amplicon sequencing. G.T. and K.S. created the \nNHEJ model. Y .M. supervised the project. G.T. and Y .M. wrote the manuscript with the help \nof all the authors. \n \nAcknowledgments \nThis work was supported by the Japan Society for the Promotion of Science (JSPS) \nGrant-in-Aid for Challenging Research (Pioneering) 20K21409, Grant-in-Aid for Scientific \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n15 \n \nResearch (B) 20H03442 and 24K02028, Interstellar Initiative Beyond from Japan Agency for \nMedical Research and Development (AMED) (23jm0610092h0001), Takeda Science \nFoundation Medical Research Grant, Sumitomo Foundation Grant for Basic Science \nResearch Projects, Ichiro Kanehara Foundation Grant to Y .M., and JSPS Grant-in-Aid for \nEarly-Career Scientists (18K15054 and 22K15386). We thank Y . Utsumi and R. Abe \n(Yamaha Motor Co., Ltd.) and all lab members for their helpful discussions. \n \nDeclaration of interests \nG.H. and S.I. are employees of Yamaha Motor Co. Ltd. All other authors declare no conflicts \nof interest in association with the present study. \n \n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n16 \n \nReferences \nChakrabarti, A.M., Henser-Brownhill, T., Monserrat, J., Poetsch, A.R., Luscombe, N.M., and \nScaffidi, P . (2019). Target-Specific Precision of CRISPR-Mediated Genome Editing. Mol. \nCell 73, 699–713.e6. \n \nChen, W., McKenna, A., Schreiber, J., Haeussler, M., Yin, Y ., Agarwal, V., Noble, W.S., and \nShendure, J. (2019). Massively parallel prof iling and predictive modeling of the outcomes of \nCRISPR/Cas9-mediated double-strand break repair. Nucleic Acids Res. 47, 7989–8003. \n \nChen, Y ., Tristan, C.A., Chen, L., Jovanovic, V.M., Malley, C., Chu, P.-H., Ryu, S., Deng, T., \nOrmanoglu, P ., Tao, D., et al. (2021). A versatile polypharmacology platform promotes \ncytoprotection and viability of human pluripotent and differentiated cells. Nat. Methods 18, \n528–541. \n \nClement, K., Rees, H., Canver, M.C., Gehrke, J.M., Farouni, R., Hsu, J.Y ., Cole, M.A., Liu, \nD.R., Joung, J.K., Bauer, D.E., et al. (2019). CRISPResso2 provides accurate and rapid \ngenome editing sequence analysis. Nat. Biotechnol. 37, 224–226. \n \nDanner, E., Bashir, S., Yumlu, S., Wurst, W., Wefers, B., and Kühn, R. (2017). Control of \ngene editing by manipulation of DNA repair mechanisms. Mamm. Genome 28, 262–274. \n \nGaj, T., Gersbach, C.A., and Barbas, C.F. (2013). ZFN, TALEN, and CRISPR/Cas-based \nmethods for genome engineering. Trends Biotechnol. 31, 397–405. \n \nGaj, T., Sirk, S.J., Shui, S.-L., and Liu, J. (2016). Genome-Editing Technologies: Principles \nand Applications. Cold Spring Harb. Perspect. Biol. 8. \n \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n17 \n \nGermini, D., Tsfasman, T., Zakharova, V.V., Sjakste, N., Lipinski, M., and Vassetzky, Y . \n(2018). A comparison of techniques to evaluate the effectiveness of genome editing. Trends \nBiotechnol. 36, 147–159. \n \nHockemeyer, D., and Jaenisch, R. (2016). Induced pluripotent stem cells meet g enome \nediting. Cell Stem Cell 18, 573–586. \n \nJinek, M., Chylinski, K., Fonfara, I., H auer, M., Doudna, J.A., and Charpentier, E. (2012). A \nprogrammable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. \nScience 337, 816–821. \n \nKato-Inui, T., Takahashi, G., Hsu, S., and Miyaoka, Y . (2018). Clustered regularly \ninterspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 with \nimproved proof-reading enhances homology-directed repair. Nucleic Acids Res. 46, \n4677–4688. \n \nKawamata, M., Suzuki, H.I., Kimura, R., and Suzuki, A. (2023). Optimization of Cas9 activity \nthrough the addition of cytosine extensions to single-guide RNAs. Nat. Biomed. Eng. 7, \n672–691. \n \nKim, S.-I., Matsumoto, T., K agawa, H., Nakamura, M., Hirohata, R., Ueno, A., Ohishi, M., \nSakuma, T., Soga, T., Yamamoto, T ., et al. (2018). Microhomology-assisted scarless \ngenome editing in human iPSCs. Nat. Commun. 9, 939. \n \nLemos, B.R., Kaplan, A.C., Bae, J.E., Ferrazzoli, A.E., Kuo, J., Anand, R.P., Waterman, D.P., \nand Haber, J.E. (2018). CRISPR/Cas9 cleavages in budding yeast reveal templated \ninsertions and strand-specific insertion/deletion profiles. Proc. Natl. Acad. Sci. USA 115, \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n18 \n \nE2040–E2047. \n \nMiyaoka, Y ., Chan, A.H., Judge, L.M., Yoo, J., Huang, M., Nguyen, T.D., Lizarraga, P.P., So, \nP .-L., and Conklin, B.R. (2014). Isolation of single-base genome-edited human iPS cells \nwithout antibiotic selection. Nat. Methods 11, 291–293. \n \nvan Overbeek, M., Capurso, D., Carter, M.M., Thompson, M.S., Frias, E., Russ, C., \nReece-Hoyes, J.S., Nye, C., Gradia, S., Vidal, B., et al. (2016). DNA Repair Profiling \nReveals Nonrandom Outcomes at Cas9-Mediated Breaks. Mol. Cell 63, 633–646. \n \nShen, M.W., Arbab, M., Hsu, J.Y ., Worstell, D., Culbertson, S.J., Krabbe, O., Cassa, C.A., \nLiu, D.R., Gifford, D.K., and Sherwood, R.I. ( 2018). Predictable and precise template-free \nCRISPR editing of pathogenic variants. Nature 563, 646–651. \n \nTakahashi, G., and Miyaoka, Y . (2023). Large-scale single-cell cloning of genome-edited \ncultured human cells by On-chip SPiS. STAR Protocols 4, 102364. \n \nTakahashi, G., Kondo, D., Maeda, M., Morishita, Y ., and Miyaoka, Y . (2022). Genome editing \nis induced in a binary manner in single human cells. IScience 25, 105619. \n \nTristan, C.A., Hong, H., Jethmalani, Y., Chen, Y ., Weber, C., Chu, P.-H., Ryu, S., Jovanovic, \nV.M., Hur, I., Voss, T.C., et al. (2023). Efficient and safe single-cell cloning of human \npluripotent stem cells using the CEPT cocktail. Nat. Protoc. 18, 58–80. \n \n \nFigure Legends \nFigure 1. Robotic isolation of genome-edited human iPS cell clones. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n19 \n \n(A) Schematic representation of the robotic isolation of genome-edited human iPS cell \nclones. After transfection of HypaCas9 and donor DNA, and selection with puromycin, \ngenome-edited single iPS cells were cultured in Matrigel domes to form cell clumps (top left \nand top center). Cell clumps were isolated by robotic picking by CELL HANDLER (top right). \nClones were genotyped by amplicon sequencing (bottom left, center, and right). (B) \nBright-field and fluorescence images of Matrigel domes in wells and EGFP-positive iPS cells \ngrowing in Matrigel domes. Five to six Matrigel domes were created in each well of a 6-well \nplate. Scale bar: left and right panels = 2 mm, center panel = 1 cm). (C) Tracking images of \na single iPS cell forming a cell clump in a Matrigel dome from day 1 to 5. Scale bar: 10 \nμ m. \n(D) Picking and seeding of cell clumps by CELL HANDLER. CELL HANDLER captured \nimages before and after picking. In the “Before picking” image, the black arrows indicate \ntarget clumps recognized by CELL HANDLER (outlined with green lines). In the “After \npicking” image, the magenta square indicates the position of a cell clump picked by CELL \nHANDLER. The “Dispensed” image shows the cell clump immediately after seeding into the \n96-well plate (red arrow). Scale bar: 300 \nμ m. (E) Proportion of the isolated iPS cell clones \nwith different allelic numbers. We analyzed the number of allelic types in isolated clones. \nClones with \n≥  3 allelic types were removed from further analysis, as those clones may not \nbe derived from single genome-edited iPS cells. \n \nFigure 2. Genotypes of genome-edited iPS cell clones. \n(A) Proportion of genotypes of genome-edited human iPS cell clones. Almost no WT/HDR \nand WT/NHEJ+HDR clones were isolated. (B) Genome editing outcomes in isolated clones \nderived from single human iPS cells edited by Cas9. RBM20 editing outcomes are shown \n(No.1 to No.3). Each bar represents 1 clone, and the genotypes of WT (green), NHEJ (blue), \nHDR (red), and HDR + NHEJ (purple) in 1 clone are also shown in each bar. (C) Total allelic \nfrequencies of WT (green), NHEJ (blue), HDR (red), and HDR + NHEJ (purple) in \ngenome-edited iPS cells in the 3 experiments shown in (B).  (D) Model of the distributions of \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n20 \n \nclones with different genotypes assuming different alleles are randomly distributed at the \nobserved frequencies. The model for RBM20 No.3 is shown. (E) Comparison of the \nproportions of WT/WT and NHEJ/NHEJ clones between the mathematical models and the \nobserved cells in RBM20,  GRN, and ATP7B. Student’s t-test was used to evaluate \ndifferences. **p <0.01, *p <0.05 and NS: not significantly different (p >0.2).  (F) Comparison \nof the proportions of WT/NHEJ clones between the mathematical models and the observed \ncells in RBM20, GRN, and ATP7B. Student’s t-test was used to evaluate differences. **p \n<0.01 and *p <0.05 \n \nFigure 3. Profiles of various indels induced by NHEJ. \n(A) Sequences and frequencies of different alleles generated by RBM20 R636S editing. The \n10 most frequently observed alleles after RBM20 editing are shown. Black and red underline \nindicates guide RNA and PAM sequences, respectively. Red triangles and red dotted lines \nindicate the cleavage site by Cas9. Blue and red characters indicate unedited and \nsubstituted nucleotides, respectively. Blue boxes highlight the microhomology sequence \n(CCGGT). (B) Genome editing outcomes with distinguished NHEJ sequences in isolated \nclones derived from single human iPS cells edited by Cas9. RBM20 editing outcomes are \nshown (No.1 to No.3) (C) Model of the distributions of clones with different genotypes \nassuming different alleles are randomly distributed at the observed frequencies. The model \nfor RBM20 No. 3 is shown.\n (D) Comparison of the proportions of homozygous NHEJ clones \nwith the top 3 most frequent indels between the mathematical models and the actually \nobserved cells. Values ±S.E. are shown (n = 3). Student’s t-test was used to evaluate \ndifferences. **p <0.01, *p <0.05 and NS: not significantly different (p >0.2). \n \nFigure 4. Maintenance of pluripotency and genotypes after robotic isolation of iPS \nclones. \n(A) Two iPS cell lines with heterozygous NHEJ in GRN identified by amplicon sequencing. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n21 \n \nDNA sequences and predicted amino acid sequences are shown. These 2 indels were \nexpected to be nonsense mutations.  (B) Sanger sequencing of the 2 heterozygous NHEJ \niPS cell lines. (C)  Digital PCR analysis of the NHEJ frequency in the 2 heterozygous NHEJ \niPS cell lines. The 2 lines had approximately 50% wild-type and 50% NHEJ alleles. WT \ngenomic DNA is the control for digital PCR. (D) Immunocytochemistry of SOX2 and OCT4 in \nthe 2 GRN heterozygous KO iPS cell lines. Scale bar: 100 μ m. \n \n \nTable \n \nTable 1. iPS cell cloning efficiencies in 96-well plates by robotic picking \nTarget mutation Exp. Wells with clones/ \nPlated wells \nCloning efficiency \n(%) \n N1 204 / 288 70.8 \nRBM20 R636S N2 168 / 240 70.0 \n N3 165 / 288 57.3 \n Total 537 / 816 66.4 \n    \n N1 174 / 264 65.9 \nGRN R493X N2 159 / 240 66.3 \n N3 137 / 240 57.1 \n Total 470 / 744 62.3 \n    \n N1 181 / 288 62.8 \nATP7B R778L N2 103 / 233 44.2 \n N3 81 / 174 46.6 \n Total 365 / 695 53.0 \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\nFigure 1\nA\nDome 2\nDome 1\nBright filed\nEGFP+ iPS cell clumps in\nMatrigel domes in a 6-well plate\nDomeWell\n96%\nATP7BGRN\n99%\nClones with one or two alleles\nClones with three or more alleles\nRBM20\n94%\nDay1 Day3 Day5\nB\nD\nC\nE\n2. Single-cell culture in \nMatrigel domes\n3. Picking of cell clumps \nby CELL HANDLERTM\n1. Selection of genome-edited \ncells by puromycin\n5. Amplicon-seq 6. Classification of generated alleles\nATGTTAGAACGG…\nATGTTAGAACGG…\nATGTT-GAACGG…\nATGTTTGAACGG…\nWT\nNHEJ\nHDR\nNHEJ+HDR\nClone A\nAllele 1\nAllele 2\nPCR\nSelection\nssDNA Donor\nHypaCas9-T2A-PuroR\n4. Library preparation by PCR\nBefore picking After picking Dispensed\nin 2D culture platein Matrigel domes\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n13.4\n%\n8.3\n%72.7\n%\n2.0\n%\n2.2\n% 0.7\n%\n0.2\n%\n0.5\n%\n29.3\n%\n13.1\n%\n51.8\n%\n1.8\n%\n0.3\n%\n3.1\n%\n0.3\n%\n0.3\n%\n18.1\n%\n12.0\n%68.1\n%\n1.4\n%\n0.5\n%\nATP7BGRNRBM20\nProportion of genotypes of  genome-edited human iPS cell clones\nRBM20 No.1\nRBM20 No.2\nWT NHEJ NHEJ+HDR        HDR\n0%\n100%\n0%\n100%\n0%\n100%\nRBM20 No.3\n0%\n100%\nModel for RBM20 No.3 WT NHEJ NHEJ+HDR        HDR\nFigure 2\nA\nB\nD\nRBM20 GRN ATP7B\nNS\nObsvd Model(%)\nWT/WT Clone\nproportion\n0\n20\n40 NS*\n0\n50\n100\n(%)\nNS NS\n*\n0\n30\n60\nWT/NHEJ Clone\nproportion\n(%)\n*\n***\nNHEJ/NHEJ Clone \nproportion\nObsvd Model Obsvd Model\nE\nRBM20 GRN ATP7B RBM20 GRN ATP7B\nWT WT/NHEJ WT/HDR WT/NHEJ+HDR NHEJ\nNHEJ/HDR NHEJ/NHEJ+HDR HDR HDR/NHEJ+HDR NHEJ+HDR\nTotal Allelic Frequencies\nWT NHEJ NHEJ+HDR        HDR\nNo.1 No.2 No.3\n0\n50\n100\n(%)\nC\nF\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\n0%\n100%\n0%\n100%\n0%\n100%\n0%\n100%\nRBM20 No.1\nRBM20 No.2\nRBM20 No.3\nModel for RBM20 No.3\nFigure 3\nA\nB\nC\nWT Δ 8-bp_1 Δ 9-bp Δ 1-bp Δ 8-bp_2 Δ 8-bp_3\nΔ 11-bp 1-bp ins Δ 6-bp NHEJ.others NHEJ+HDR HDR\nD RBM20\n*\n0\n10\n20\n0\n4\n8\n0\n4\n8\nObsvd\nModel\nGRN ATP7B\n** NS\nP=0.0532\nNS ** *\nObsvd\nModel\nObsvd\nModel\n(%) (%) (%)\nΔ 8-bp_1\nWT\nΔ 9-bp\nΔ 1-bp\nΔ 8-bp_2\nHDR\nΔ 8-bp_3\nΔ 11-bp\n1-bp ins\nΔ 6-bp\nCGCGGTCTCGTAGTCCGGT--------CACTCTCCCCGAG\nCGCGGTCTCGTAGTCCGGTGAGCCGGTCACTCTCCCCGAG\nCGCGGTCTCGTAG---------CCGGTCACTCTCCCCGAG\nCGCGGTCTCGTAGTCCGGT-AGCCGGTCACTCTCCCCGAG\nCGCGGTCTCGTA--------AGCCGGTCACTCTCCCCGAG\nCGCGGTCTAGTAGTCCGGTGAGCCGGTCACTCTCCCCGAG\nCGCGGTCTCGTAGTCCGGTG--------ACTCTCCCCGAG\nCGCGGTCTC-----------AGCCGGTCACTCTCCCCGAG\nCGCGGTCTCGTAGTCCGGTGAAGCCGGTCACTCTCCCCGA\nCGCGGTCTCGTAGTCCGGTG------TCACTCTCCCCGAG\nCGCGGTCTCGTAGTCCGGTGAGCCGGTCACTCTCCCCGAG\n28.2%\n17.9%\n14.3%\n5.6%\n2.7%\n1.9%\n1.7%\n1.4%\n1.3%\n1.1%\nAllelic frequencies in RBM20 edited-iPS clones\nRef\nP=0.1413\nP=0.1594\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint \n\nGRN heterozygous knock-out iPS cell lines\nWT ARSCEKEVVSAQPATFLARSP…\nGRN-Het 1-bp ins ARSCEEGSGLCPACHLPGP*\nGRN-Het Δ 19-bp ARSLPSLPPSWPVALTWV*\nPredicted amino acid sequence\nWT GCTCGATCCTGCGAGAAGGAAGTGGTCTCT…\nGRN-Het 1-bp ins (WT/1-bp ins) GCTCGATCCTGCGAAGAAGGAAGTGGTCTCT…\nGRN-Het Δ 19-bp (WT/Δ 19-bp) GCTCGATC-------------------TCT…\n0\n50\n100 WT\nNHEJ\nGRN-Het\n1-bp ins\nGRN-Het\nΔ 19-bp\nWT\nAllelic frequencies(%)\nFigure 4\nA\nC\nB\nD\nCCTGCGAGAAGGAAG\nCCTGCGAAGAAGGAA\nWT\n1-bp ins\n1-bp ins\nGRN-Het 1-bp ins\n（ WT/1-bp ins）\nATCCTGCGAGAAGGAAGTGGTCTCT\nATCTCTGCCCAGCCTGCCACCTTCC\nΔ 19-bp\nWT\nΔ 19-bp\nGRN-Het Δ 19-bp\n（ WT/Δ 19-bp）\nGRN-Het Δ 19-bpGRN-Het 1-bp ins \nSOX2\n OCT4SOX2SOX2\n OCT4\nSOX2\nSequences from amplicon-sequencing\nNuclei Nuclei Nuclei Nuclei\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 19, 2024. ; https://doi.org/10.1101/2024.09.18.613641doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}