CoChat
Sign Up
Cited by 471

The DOI matches its Crossref record; 11 of the 11 references it cites were found in a registry or an index.

Precise Correction of the Dystrophin Gene in Duchenne Muscular Dystrophy Patient Induced Pluripotent Stem Cells by TALEN and CRISPR-Cas9

Li, Hongmei Lisa, Fujimoto, Naoko, Sasakawa, Noriko, Shirai, Saya, Ohkame, Tokiko, Sakuma, Tetsushi, Tanaka, Michihiro, Amano, Naoki, Watanabe, Akira, Sakurai, Hidetoshi, Yamamoto, Takashi, Yamanaka, Shinya, Hotta, Akitsu·2015·Stem Cell Reports
DOI 10.1016/j.stemcr.2014.10.013PMC PMC4297888
Open this paper in CoChat — free account

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1016/j.stemcr.2014.10.013

Cites

11 of 11 references verified0 citation errors · 0 not checkable

Statistics

Figures

Graphical abstract
Graphical abstract CC BY · PubMed Central
Figure 1 Visualization of the Unique Sequences in the Human Genome (A) Example of the mapped unique k-mer sequences at the exon 45 region of the dystrophin gene. The mapped k-mers are indicated in the bottom panel and the coverage of the k-mers in each base position is indicated by the gray histogram. Within this region, unique sequences of 14- to 16-mer were identified. (B) Pseudocolor dot plot for the depth of unique k-mers and the number of potential off-target sites. CRISPR-sgRNA targeting sequences (23 bp with “NGG” PAM, n = 84,947) were randomly selected from the human genome and the depth of the unique k-mers and number of potential off-target sites (with up to 3 bp mismatches allowance) for each sgRNA sequence were calculated. Note that higher depth correlated with fewer potential off-target sites. See also Table S1.
Figure 1 Visualization of the Unique Sequences in the Human Genome (A) Example of the mapped unique k-mer sequences at the exon 45 region of the dystrophin gene. The mapped k-mers are indicated in the bottom panel and the coverage of the k-mers in each base position is indicated by the gray histogram. Within this region, unique sequences of 14- to 16-mer were identified. (B) Pseudocolor dot plot for the depth of unique k-mers and the number of potential off-target sites. CRISPR-sgRNA targeting sequences (23 bp with “NGG” PAM, n = 84,947) were randomly selected from the human genome and the depth of the unique k-mers and number of potential off-target sites (with up to 3 bp mismatches allowance) for each sgRNA sequence were calculated. Note that higher depth correlated with fewer potential off-target sites. See also Table S1.CC BY · PubMed Central
Figure 2 TALEN and CRISPR-sgRNA Are Equally Active for Targeting the Dystrophin Gene in iPSCs (A) The three strategies used to restore the reading frame of the dystrophin protein. (B) We designed the Platinum-TALEN pair and CRISPR-sgRNA1 within the peak of the unique k-mer sequences, as indicated by the gray histogram, and in front of the de novo premature stop codon, as indicated by the red hexagon. (C) The activities of Platinum-TALEN and CRISPR-sgRNA1 were analyzed by a restriction enzyme (XcmI) sensitivity assay. The XcmI digestion site (CCANNNNNNNNNTGG) was located in the spacer of the TALEN and next to the PAM sequence of the CRISPR-sgRNA1. The intensity percentage of the undigested band (703 bp) was used to calculate the mutation efficiency. (D) The frequency of deletions (black bar) and insertions (gray bar) was analyzed by deep sequencing for the region flanking the target site. The percentages of sequence reads corresponding to deletions, insertions, and mutated XcmI sites (white bar) are plotted. No TF, nontransfected control. See also Figure S2.
Figure 2 TALEN and CRISPR-sgRNA Are Equally Active for Targeting the Dystrophin Gene in iPSCs (A) The three strategies used to restore the reading frame of the dystrophin protein. (B) We designed the Platinum-TALEN pair and CRISPR-sgRNA1 within the peak of the unique k-mer sequences, as indicated by the gray histogram, and in front of the de novo premature stop codon, as indicated by the red hexagon. (C) The activities of Platinum-TALEN and CRISPR-sgRNA1 were analyzed by a restriction enzyme (XcmI) sensitivity assay. The XcmI digestion site (CCANNNNNNNNNTGG) was located in the spacer of the TALEN and next to the PAM sequence of the CRISPR-sgRNA1. The intensity percentage of the undigested band (703 bp) was used to calculate the mutation efficiency. (D) The frequency of deletions (black bar) and insertions (gray bar) was analyzed by deep sequencing for the region flanking the target site. The percentages of sequence reads corresponding to deletions, insertions, and mutated XcmI sites (white bar) are plotted. No TF, nontransfected control. See also Figure S2.CC BY · PubMed Central
Figure 3 Correction of the Dystrophin Gene by Exon Skipping and Frameshifting Solely by Nuclease Treatment (A) Summary of the exon-skipping (ES) and in-frame (IF) correction approaches using the indicated nucleases without the donor template. The DMD-c1 iPSC clone was derived by a retroviral method and the DMD-iPSC #1 (CiRA00111) clone was derived by an integration-free episomal vector method. Sanger sequence analyses were performed for all picked clones to identify subclones that had indels. The number of ES and IF clones is indicated. (B) The deletion patterns induced by TALENs and CRISPR-sgRNAs were analyzed by deep sequencing of the target site. The numbers around the pie charts indicate the deletion sizes found in more than 5% of sequence reads, and the numbers with an asterisk indicate deletions that could restore the reading frame of the dystrophin gene. The boxed numbers indicate the deletion patterns flanked with local microhomologies. (C) For each of the sequence reads obtained by deep sequencing, microhomology motifs on both sides of the deletion were retrieved and the distribution of the microhomology sizes was plotted. See also Figure S3 and Table S2.
Figure 3 Correction of the Dystrophin Gene by Exon Skipping and Frameshifting Solely by Nuclease Treatment (A) Summary of the exon-skipping (ES) and in-frame (IF) correction approaches using the indicated nucleases without the donor template. The DMD-c1 iPSC clone was derived by a retroviral method and the DMD-iPSC #1 (CiRA00111) clone was derived by an integration-free episomal vector method. Sanger sequence analyses were performed for all picked clones to identify subclones that had indels. The number of ES and IF clones is indicated. (B) The deletion patterns induced by TALENs and CRISPR-sgRNAs were analyzed by deep sequencing of the target site. The numbers around the pie charts indicate the deletion sizes found in more than 5% of sequence reads, and the numbers with an asterisk indicate deletions that could restore the reading frame of the dystrophin gene. The boxed numbers indicate the deletion patterns flanked with local microhomologies. (C) For each of the sequence reads obtained by deep sequencing, microhomology motifs on both sides of the deletion were retrieved and the distribution of the microhomology sizes was plotted. See also Figure S3 and Table S2.CC BY · PubMed Central
Figure 4 Correction of the Full-Length Dystrophin Protein by Exon 44 Knockin (A) Schematic overview depicting the exon 44 knockin strategy. Top line: structure of the dystrophin gene with the exon 44 deletion. The vertical blue arrow indicates the TALEN/CRISPR cut site and the red bar indicates the intron 45 probe used for Southern blot analysis. Note that we introduced silent mutations at the nuclease-targeting site within exon 45 of the donor template. (B) Summary of the exon 44 knockin experiments. The picked clones were first screened by genomic PCR for targeted knockin of the donor template and then by Southern blotting with EcoRI digestion and the intron 45 probe to confirm no additional integration. Successfully targeted clones were further treated with Cre to remove the drug selection cassette flanked by the loxP elements. (C) Southern blot of the knockin clones (clone names: TKII15, TKII17, and TKIE27 for TALEN-mediated knockin clones; CKIC2, CKIC4, and CKIC6 for CRISPR-mediated knockin clones) showing successful targeting at the designated site. Subsequent Cre treatment excised the loxP-flanked drug selection cassette. The probe in intron 45 was used to detect EcoRI-digested genomic DNA fragments. See also Figures S3B and S3C.
Figure 4 Correction of the Full-Length Dystrophin Protein by Exon 44 Knockin (A) Schematic overview depicting the exon 44 knockin strategy. Top line: structure of the dystrophin gene with the exon 44 deletion. The vertical blue arrow indicates the TALEN/CRISPR cut site and the red bar indicates the intron 45 probe used for Southern blot analysis. Note that we introduced silent mutations at the nuclease-targeting site within exon 45 of the donor template. (B) Summary of the exon 44 knockin experiments. The picked clones were first screened by genomic PCR for targeted knockin of the donor template and then by Southern blotting with EcoRI digestion and the intron 45 probe to confirm no additional integration. Successfully targeted clones were further treated with Cre to remove the drug selection cassette flanked by the loxP elements. (C) Southern blot of the knockin clones (clone names: TKII15, TKII17, and TKIE27 for TALEN-mediated knockin clones; CKIC2, CKIC4, and CKIC6 for CRISPR-mediated knockin clones) showing successful targeting at the designated site. Subsequent Cre treatment excised the loxP-flanked drug selection cassette. The probe in intron 45 was used to detect EcoRI-digested genomic DNA fragments. See also Figures S3B and S3C.CC BY · PubMed Central
Figure 5 Multiple Whole-Genome Analyses Revealed No Severe Off-Target Mutagenesis (A) Summary of the karyotyping analyses of the corrected clones. Conventional Giemsa staining was used for chromosome counting (50 metaphases were counted), and high-resolution G-banding was applied to detect chromosomal rearrangements (at least 20 metaphases were analyzed). (B) SNP array analysis identified de novo CNVs in the corrected clones. To determine associations between detected CNVs and potential off-target sites, the distances between the edges of the detected CNVs and potential nuclease target sites are plotted for each nuclease (red cross for sgRNA1, green diamond for Golden-TALEN, and blue × mark for Platinum-TALEN). Since the likelihood of the distance distribution depends on how many genomic sites are selected, we also selected random genomic positions and calculated the distance for each CNV (gray circles). A zero value in the y axis indicates the edge of the CNVs and a negative value indicates the inside of the CNVs. If any given nuclease off-target site is associated with the detected CNVs, the target site should approximate zero distance. (C) The number of single nucleotide variants (SNVs) and small indels detected by exome sequencing. Only de novo SNVs and indels that did not exist in the original DMD fibroblasts or DMD iPSCs are plotted. Unmodified control clones underwent the same electroporation and subcloning process as the other corrected clones and were sequenced at similar passage numbers. See also Figures S4–S6 and Tables S3–S5.
Figure 5 Multiple Whole-Genome Analyses Revealed No Severe Off-Target Mutagenesis (A) Summary of the karyotyping analyses of the corrected clones. Conventional Giemsa staining was used for chromosome counting (50 metaphases were counted), and high-resolution G-banding was applied to detect chromosomal rearrangements (at least 20 metaphases were analyzed). (B) SNP array analysis identified de novo CNVs in the corrected clones. To determine associations between detected CNVs and potential off-target sites, the distances between the edges of the detected CNVs and potential nuclease target sites are plotted for each nuclease (red cross for sgRNA1, green diamond for Golden-TALEN, and blue × mark for Platinum-TALEN). Since the likelihood of the distance distribution depends on how many genomic sites are selected, we also selected random genomic positions and calculated the distance for each CNV (gray circles). A zero value in the y axis indicates the edge of the CNVs and a negative value indicates the inside of the CNVs. If any given nuclease off-target site is associated with the detected CNVs, the target site should approximate zero distance. (C) The number of single nucleotide variants (SNVs) and small indels detected by exome sequencing. Only de novo SNVs and indels that did not exist in the original DMD fibroblasts or DMD iPSCs are plotted. Unmodified control clones underwent the same electroporation and subcloning process as the other corrected clones and were sequenced at similar passage numbers. See also Figures S4–S6 and Tables S3–S5.CC BY · PubMed Central
Figure 6 Restoration of the Dystrophin Protein in Differentiated Myogenic Cells (A) RT-PCR analysis for dystrophin cDNA from iPSCs and cells differentiated from the corrected clones toward skeletal muscle lineage. The original DMD patient and an IF clone (IFH30) with 1 bp insertion corresponded to the 452 and 453 bp PCR bands, respectively. The healthy control and knockin clones (TALEN-mediated TKII15 and CRISPR-mediated CKIC2) corresponded to the 600 bp bands; and the ES clone (ESH29) corresponded to the 276 bp band. (B) Sanger sequence analysis of dystrophin cDNAs from differentiated skeletal muscle cells. The IF clone (IFH30) exhibited a 1 bp insertion (A, black arrow), the ES clone (ESH29) exhibited the conjugation of exons 43 and 46 due to the skipping of exon 45, and knockin clones (TKII15 and CKIC2) exhibited the complete restoration of exon 44 in front of exon 45 as the healthy control. (C) Immunofluorescence staining of skeletal muscle cells differentiated from the corrected clones. A z axis section of the confocal microscopy image shows submembrane localization of the dystrophin protein in the healthy control and all corrected clones, but not in the uncorrected original DMD iPSCs. The cells were stained by DAPI, a marker of skeletal differentiation (myosin heavy chain [MHC]), in red and an antibody that detects the rod domain of dystrophin (DYS1) in green. Scale bar, 50 μm. (D) Western blot analysis to estimate the molecular weight of the dystrophin protein in the corrected clones. Expected molecular weight: 420 kDa for the reading-frame-corrected clone, 414 kDa for the exon-skipping clone, and 427 kDa for the exon 44 knockin clones and healthy control. An anti-dystrophin C terminus (amino acids 3661–3677) antibody was used to detect dystrophin protein, and an anti-α-SMA antibody was used as the sample loading control.
Figure 6 Restoration of the Dystrophin Protein in Differentiated Myogenic Cells (A) RT-PCR analysis for dystrophin cDNA from iPSCs and cells differentiated from the corrected clones toward skeletal muscle lineage. The original DMD patient and an IF clone (IFH30) with 1 bp insertion corresponded to the 452 and 453 bp PCR bands, respectively. The healthy control and knockin clones (TALEN-mediated TKII15 and CRISPR-mediated CKIC2) corresponded to the 600 bp bands; and the ES clone (ESH29) corresponded to the 276 bp band. (B) Sanger sequence analysis of dystrophin cDNAs from differentiated skeletal muscle cells. The IF clone (IFH30) exhibited a 1 bp insertion (A, black arrow), the ES clone (ESH29) exhibited the conjugation of exons 43 and 46 due to the skipping of exon 45, and knockin clones (TKII15 and CKIC2) exhibited the complete restoration of exon 44 in front of exon 45 as the healthy control. (C) Immunofluorescence staining of skeletal muscle cells differentiated from the corrected clones. A z axis section of the confocal microscopy image shows submembrane localization of the dystrophin protein in the healthy control and all corrected clones, but not in the uncorrected original DMD iPSCs. The cells were stained by DAPI, a marker of skeletal differentiation (myosin heavy chain [MHC]), in red and an antibody that detects the rod domain of dystrophin (DYS1) in green. Scale bar, 50 μm. (D) Western blot analysis to estimate the molecular weight of the dystrophin protein in the corrected clones. Expected molecular weight: 420 kDa for the reading-frame-corrected clone, 414 kDa for the exon-skipping clone, and 427 kDa for the exon 44 knockin clones and healthy control. An anti-dystrophin C terminus (amino acids 3661–3677) antibody was used to detect dystrophin protein, and an anti-α-SMA antibody was used as the sample loading control.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Research works printed but not linked: 38

  1. Aartsma-Rus, 2009, Theoretic applicability of antisense-mediated exon skipping for Duchenne muscular dystrophy mutations, Hum. Mutat., 30, 293, 10.1002/humu.20918 · DOI
  2. Amps, 2011, Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage, Nat. Biotechnol., 29, 1132, 10.1038/nbt.2051 · DOI
  3. Bae, 2014, Microhomology-based choice of Cas9 nuclease target sites, Nat. Methods, 11, 705, 10.1038/nmeth.3015 · DOI
  4. Bae, 2014, Cas-OFFinder: a fast and versatile algorithm that searches for potential off-target sites of Cas9 RNA-guided endonucleases, Bioinformatics, 30, 1473, 10.1093/bioinformatics/btu048 · DOI · Paper page
  5. Canver, 2014, Characterization of genomic deletion efficiency mediated by clustered regularly interspaced palindromic repeats (CRISPR)/Cas9 nuclease system in mammalian cells, J. Biol. Chem., 289, 21312, 10.1074/jbc.M114.564625 · DOI
  6. Choi, 2013, Efficient drug screening and gene correction for treating liver disease using patient-specific stem cells, Hepatology, 57, 2458, 10.1002/hep.26237 · DOI
  7. Ding, 2013, A TALEN genome-editing system for generating human stem cell-based disease models, Cell Stem Cell, 12, 238, 10.1016/j.stem.2012.11.011 · DOI
  8. Ding, 2013, Enhanced efficiency of human pluripotent stem cell genome editing through replacing TALENs with CRISPRs, Cell Stem Cell, 12, 393, 10.1016/j.stem.2013.03.006 · DOI · Paper page
  9. Filareto, 2013, An ex vivo gene therapy approach to treat muscular dystrophy using inducible pluripotent stem cells, Nat. Commun., 4, 1549, 10.1038/ncomms2550 · DOI
  10. Goudenege, 2012, Myoblasts derived from normal hESCs and dystrophic hiPSCs efficiently fuse with existing muscle fibers following transplantation, Mol. Ther., 20, 2153, 10.1038/mt.2012.188 · DOI
  11. Heigwer, 2014, E-CRISP: fast CRISPR target site identification, Nat. Methods, 11, 122, 10.1038/nmeth.2812 · DOI
  12. Hockemeyer, 2011, Genetic engineering of human pluripotent cells using TALE nucleases, Nat. Biotechnol., 29, 731, 10.1038/nbt.1927 · DOI
  13. Hotta, 2009, Isolation of human iPS cells using EOS lentiviral vectors to select for pluripotency, Nat. Methods, 6, 370, 10.1038/nmeth.1325 · DOI
  14. Koehler, 2011, The uniqueome: a mappability resource for short-tag sequencing, Bioinformatics, 27, 272, 10.1093/bioinformatics/btq640 · DOI
  15. Langmead, 2009, Searching for SNPs with cloud computing, Genome Biol., 10, R134, 10.1186/gb-2009-10-11-r134 · DOI
  16. Li, 2014, Genetic correction using engineered nucleases for gene therapy applications, Dev. Growth Differ., 56, 63, 10.1111/dgd.12107 · DOI
  17. Ma, 2013, Transcription activator-like effector nuclease (TALEN)-mediated gene correction in integration-free β-thalassemia induced pluripotent stem cells, J. Biol. Chem., 288, 34671, 10.1074/jbc.M113.496174 · DOI
  18. Maetzel, 2014, Genetic and chemical correction of cholesterol accumulation and impaired autophagy in hepatic and neural cells derived from Niemann-Pick Type C patient-specific iPS cells, Stem Cell Rep., 2, 866, 10.1016/j.stemcr.2014.03.014 · DOI · Paper page
  19. Matsui, 2014, Delivery of full-length factor VIII using a piggyBac transposon vector to correct a mouse model of hemophilia A, PLoS ONE, 9, e104957, 10.1371/journal.pone.0104957 · DOI
  20. Okada, 2013, Current challenges and future directions in recombinant AAV-mediated gene therapy of Duchenne muscular dystrophy, Pharmaceuticals (Basel), 6, 813, 10.3390/ph6070813 · DOI
  21. Okita, 2011, A more efficient method to generate integration-free human iPS cells, Nat. Methods, 8, 409, 10.1038/nmeth.1591 · DOI · Paper page
  22. Osborn, 2013, TALEN-based gene correction for epidermolysis bullosa, Mol. Ther., 21, 1151, 10.1038/mt.2013.56 · DOI · Paper page
  23. Ousterout, 2013, Reading frame correction by targeted genome editing restores dystrophin expression in cells from Duchenne muscular dystrophy patients, Mol. Ther., 21, 1718, 10.1038/mt.2013.111 · DOI
  24. Pichavant, 2011, Current status of pharmaceutical and genetic therapeutic approaches to treat DMD, Mol. Ther., 19, 830, 10.1038/mt.2011.59 · DOI
  25. Popplewell, 2013, Gene correction of a duchenne muscular dystrophy mutation by meganuclease-enhanced exon knock-in, Hum. Gene Ther., 24, 692, 10.1089/hum.2013.081 · DOI
  26. Rousseau, 2011, Endonucleases: tools to correct the dystrophin gene, J. Gene Med., 13, 522, 10.1002/jgm.1611 · DOI
  27. Sakuma, 2013, Efficient TALEN construction and evaluation methods for human cell and animal applications, Genes Cells, 18, 315, 10.1111/gtc.12037 · DOI
  28. Sakuma, 2013, Repeating pattern of non-RVD variations in DNA-binding modules enhances TALEN activity, Sci. Rep., 3, 3379, 10.1038/srep03379 · DOI
  29. Sebastiano, 2011, In situ genetic correction of the sickle cell anemia mutation in human induced pluripotent stem cells using engineered zinc finger nucleases, Stem Cells, 29, 1717, 10.1002/stem.718 · DOI
  30. Soldner, 2011, Generation of isogenic pluripotent stem cells differing exclusively at two early onset Parkinson point mutations, Cell, 146, 318, 10.1016/j.cell.2011.06.019 · DOI · Paper page
  31. Sugiura, 2014, Induced pluripotent stem cell generation-associated point mutations arise during the initial stages of the conversion of these cells, Stem Cell Rep., 2, 52, 10.1016/j.stemcr.2013.11.006 · DOI
  32. Suzuki, 2014, Targeted gene correction minimally impacts whole-genome mutational load in human-disease-specific induced pluripotent stem cell clones, Cell Stem Cell, 15, 31, 10.1016/j.stem.2014.06.016 · DOI · Paper page
  33. Takahashi, 2007, Induction of pluripotent stem cells from adult human fibroblasts by defined factors, Cell, 131, 861, 10.1016/j.cell.2007.11.019 · DOI · Paper page
  34. Tanaka, 2013, Efficient and reproducible myogenic differentiation from human iPS cells: prospects for modeling Miyoshi Myopathy in vitro, PLoS ONE, 8, e61540, 10.1371/journal.pone.0061540 · DOI
  35. Tuffery-Giraud, 2009, Genotype-phenotype analysis in 2,405 patients with a dystrophinopathy using the UMD-DMD database: a model of nationwide knowledgebase, Hum. Mutat., 30, 934, 10.1002/humu.20976 · DOI
  36. Ye, 2014, Seamless modification of wild-type induced pluripotent stem cells to the natural CCR5Δ32 mutation confers resistance to HIV infection, Proc. Natl. Acad. Sci. USA, 111, 9591, 10.1073/pnas.1407473111 · DOI · Paper page
  37. Zou, 2011, Site-specific gene correction of a point mutation in human iPS cells derived from an adult patient with sickle cell disease, Blood, 118, 4599, 10.1182/blood-2011-02-335554 · DOI
  38. Zou, 2011, Oxidase-deficient neutrophils from X-linked chronic granulomatous disease iPS cells: functional correction by zinc finger nuclease-mediated safe harbor targeting, Blood, 117, 5561, 10.1182/blood-2010-12-328161 · DOI

Cited by

Sorted by citations

See all 471

Metadata

Authors
Li, H. L., Fujimoto, N., Sasakawa, N., Shirai, S., Ohkame, T., Sakuma, T., Tanaka, M., Amano, N., Watanabe, A., Sakurai, H., Yamamoto, T., Yamanaka, S., & Hotta, A.
Venue
Stem Cell Reports
Publisher
Elsevier BV
Published
2015-01
Volume
4
Issue
1
Pages
143-154
ISSN
2213-6711
DOI
10.1016/j.stemcr.2014.10.013
PMC
PMC4297888
Cited by
471
References
11
Language
en
Licence
Funders
  • JSPS KAKENHI
  • JST PRESTO
  • JST Yamanaka iPS Cell Special Project
  • JST Research Center Network for Realization of Regenerative Medicine
  • JSPS

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review