CoChat

ANK2 autism mutation targeting giant ankyrin-B promotes axon branching and ectopic connectivity

Open access

The DOI matches its Crossref record; 46 of the 46 references it cites were found in a registry or an index.

Abstract

Significance: Autism spectrum disorders (ASD) are heritable neurodevelopmental disorders affecting communication and behavior, and can include impaired intellectual ability. ASD is believed to result from altered synapse function and gene regulation. Among high-risk ASD genes, ANK2 is of particular interest since ANK2 mutations with an ASD diagnosis can occur with normal IQ. Here, we report that ASD ANK2 mutation in mice results in increased axon branching and ectopic connectivity associated with penetrant impairment in selected communicative and social behaviors combined with superior executive function. Thus, gain of axon branching is a candidate cellular mechanism to explain aberrant structural connectivity and dominantly inherited behavior that are compatible with normal intelligence and potentially can contribute to ASD and other forms of neurodiversity.

Overview

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

Integrity signals

Registry

Crossref · 10.1073/pnas.1904348116

Cites

46 of 46 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1. Mouse models for ANK2 ASD mutations. (A) Schematic of giant ankB and 220-kDa ankB polypeptides with de novo ANK2 ASD variants mapped. Binding sites of the pan ankB antibody (blue) or the giant ankB specific antibody (red) are marked with asterisks. (B) In-gel Western of a neuronal culture versus a glia cell culture with pan ankB antibody (panAB). (C) Schematic of gene modification and affected polypeptides in 3 mouse models of ANK2 deficiency/mutation. (D) In-gel Western of HEK293T cell lysates that were transfected with giant ankB cDNA or giant ankB-R2608fs human frameshift cDNA using the giant ankB-specific antibody (gAB). Asterisks mark a 290 kDa polypeptide. (E) In-gel Western of brain lysates from WT or ABe37fs/fs (ABfs/fs) mice using giant ankB specific antibody (gAB). Asterisks mark a 290 kDa polypeptide. (F) Quantification of ankB polypeptides in mouse models of ANK2 deficiency/mutation normalized to WT mouse (n = 3 mice per genotype for all immunoblots).
Fig. 1. Mouse models for ANK2 ASD mutations. (A) Schematic of giant ankB and 220-kDa ankB polypeptides with de novo ANK2 ASD variants mapped. Binding sites of the pan ankB antibody (blue) or the giant ankB specific antibody (red) are marked with asterisks. (B) In-gel Western of a neuronal culture versus a glia cell culture with pan ankB antibody (panAB). (C) Schematic of gene modification and affected polypeptides in 3 mouse models of ANK2 deficiency/mutation. (D) In-gel Western of HEK293T cell lysates that were transfected with giant ankB cDNA or giant ankB-R2608fs human frameshift cDNA using the giant ankB-specific antibody (gAB). Asterisks mark a 290 kDa polypeptide. (E) In-gel Western of brain lysates from WT or ABe37fs/fs (ABfs/fs) mice using giant ankB specific antibody (gAB). Asterisks mark a 290 kDa polypeptide. (F) Quantification of ankB polypeptides in mouse models of ANK2 deficiency/mutation normalized to WT mouse (n = 3 mice per genotype for all immunoblots).CC BY-NC-ND · PubMed Central
Fig. 2. Giant ankB mutant mice exhibit abnormal communication, sexual behavior responses, repetitive behavior, and increased executive function in the absence of ASD-associated comorbidities. (A) ABfs mice show a normal preference for the social stimulus in the social affiliation test. RMANOVA, n = 14 to 16. (B) Grooming bouts are increased in mutants with water misting. Unpaired t test, n = 16. (C) Motor performance on the constant speed rotarod is normal. RMANOVA, n = 8. (D) Normal short- and long-term memory in the novel object recognition memory task. RMANOVA, n = 15 to 16. (E) USVs from PND 7 mice are decreased across 3 giant ankB-deficient strains. One-way ANOVA, n = 23 to 29 (ABe37fs); unpaired t test, n = 8 to 13 (ABe37floxed), n = 6 to 7 (ABe22). (F) Decreased territory marking in response to estrous female urine in the 3 giant ankB-deficient strains as determined by the number of male urine marking made upon exposure to female urine. (F, Upper) One-way ANOVA, n = 22, 7, 18 (ABe37fs); unpaired t test, n = 4 to 7 (ABe37floxed), n = 5 to 6 (ABe22). (Lower) Representative traces from ninhydrin-exposed urine marks from ABe37fs males, female estrous urine is shown as a red circle. (G) Acquisition performance in the water T-maze is normal. RMANOVA, n = 10 to 20 (ABe37fs), n = 9 to 14 (ABe37floxed), n = 5, 21 (ABe22). (H) Superior performance in 2 strains with giant ankB mutations on the first reversal trial in the water T-maze. (Upper) Number of errors on trial one reversal. One-way ANOVA, n = 6 to 10 (ABe37fs); unpaired t test, n = 7 (ABe37floxed), n = 5 to 11 (ABe22). (Lower) Reaching the hidden platform within 1 min (yes = 1, no = 0). χ2 test: ABe37fs: n = 10 to 18 (ABe37fs), n = 9 to 14 (ABe37floxed), n = 5, 21 (ABe22). (Right) Representative traces of ABe37fs performance on reversal trial 1. T-maze in gray; black line shows the trajectory of a mouse from the start box (green cross) to the arm with the hidden platform (red rectangle). All numeric data are presented as means ± SEM; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, versus the WT or Cre control; ns, not significant. n value represents mice per genotype. See SI Appendix for full statistical description of figure.
Fig. 2. Giant ankB mutant mice exhibit abnormal communication, sexual behavior responses, repetitive behavior, and increased executive function in the absence of ASD-associated comorbidities. (A) ABfs mice show a normal preference for the social stimulus in the social affiliation test. RMANOVA, n = 14 to 16. (B) Grooming bouts are increased in mutants with water misting. Unpaired t test, n = 16. (C) Motor performance on the constant speed rotarod is normal. RMANOVA, n = 8. (D) Normal short- and long-term memory in the novel object recognition memory task. RMANOVA, n = 15 to 16. (E) USVs from PND 7 mice are decreased across 3 giant ankB-deficient strains. One-way ANOVA, n = 23 to 29 (ABe37fs); unpaired t test, n = 8 to 13 (ABe37floxed), n = 6 to 7 (ABe22). (F) Decreased territory marking in response to estrous female urine in the 3 giant ankB-deficient strains as determined by the number of male urine marking made upon exposure to female urine. (F, Upper) One-way ANOVA, n = 22, 7, 18 (ABe37fs); unpaired t test, n = 4 to 7 (ABe37floxed), n = 5 to 6 (ABe22). (Lower) Representative traces from ninhydrin-exposed urine marks from ABe37fs males, female estrous urine is shown as a red circle. (G) Acquisition performance in the water T-maze is normal. RMANOVA, n = 10 to 20 (ABe37fs), n = 9 to 14 (ABe37floxed), n = 5, 21 (ABe22). (H) Superior performance in 2 strains with giant ankB mutations on the first reversal trial in the water T-maze. (Upper) Number of errors on trial one reversal. One-way ANOVA, n = 6 to 10 (ABe37fs); unpaired t test, n = 7 (ABe37floxed), n = 5 to 11 (ABe22). (Lower) Reaching the hidden platform within 1 min (yes = 1, no = 0). χ2 test: ABe37fs: n = 10 to 18 (ABe37fs), n = 9 to 14 (ABe37floxed), n = 5, 21 (ABe22). (Right) Representative traces of ABe37fs performance on reversal trial 1. T-maze in gray; black line shows the trajectory of a mouse from the start box (green cross) to the arm with the hidden platform (red rectangle). All numeric data are presented as means ± SEM; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, versus the WT or Cre control; ns, not significant. n value represents mice per genotype. See SI Appendix for full statistical description of figure.CC BY-NC-ND · PubMed Central
Fig. 3. Synaptic function in PND 28 and PND 60 ABe37fs mice. (A and B) Quantification of mEPSCs and mIPSCs at PND 28 and PND 60 in acute slices from mice using whole-cell patch-clamp recording. Dots represent individual neurons. (n = 10 mice per genotype). (C) Quantification of paired-pulse facilitation generated by an interpulse interval (50 to 150 ms) in PND 28 and PND 60 mice (n = 8 mice per genotype). (D) Quantification of action potentials generated by direct intracellular current injections in current-clamp recordings in PND 28 and PND 60 mice. n = 8 mice per genotype; Mean ± SEM; (A and B) t test; (C and D) repeated measures 2-way ANOVA, Bonferroni post hoc test; *P < 0.05, **P < 0.01; ns, not significant.
Fig. 3. Synaptic function in PND 28 and PND 60 ABe37fs mice. (A and B) Quantification of mEPSCs and mIPSCs at PND 28 and PND 60 in acute slices from mice using whole-cell patch-clamp recording. Dots represent individual neurons. (n = 10 mice per genotype). (C) Quantification of paired-pulse facilitation generated by an interpulse interval (50 to 150 ms) in PND 28 and PND 60 mice (n = 8 mice per genotype). (D) Quantification of action potentials generated by direct intracellular current injections in current-clamp recordings in PND 28 and PND 60 mice. n = 8 mice per genotype; Mean ± SEM; (A and B) t test; (C and D) repeated measures 2-way ANOVA, Bonferroni post hoc test; *P < 0.05, **P < 0.01; ns, not significant.CC BY-NC-ND · PubMed Central
Fig. 4. ABe37fs/fs mutation impairs ankB–L1CAM interaction in axons. (A) AnkB antibody binding sites are marked with asterisks in the schematic of 220-kDa ankB and giant ankB (blue, pan ankB antibody; red, giant ankB specific antibody). (B) Immunostaining for giant ankB (gAB), L1CAM, and pan ankB (panAB) in corpus callosum of PND 28 mouse brain coronal sections (Scale bars, 50 µm). (C) PLA signal in the corpus callosal (cc) region of indicated mouse strains using L1CAM antibody combined with either pan ankB antibody (Left) or giant ankB antibody (Right). Nuclei are in blue (Scale bars, 50 µm). (D) Quantification of PLA signal. Mean ± SEM; ***P = 0.003; ****P = 0.0001, 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n = 3.
Fig. 4. ABe37fs/fs mutation impairs ankB–L1CAM interaction in axons. (A) AnkB antibody binding sites are marked with asterisks in the schematic of 220-kDa ankB and giant ankB (blue, pan ankB antibody; red, giant ankB specific antibody). (B) Immunostaining for giant ankB (gAB), L1CAM, and pan ankB (panAB) in corpus callosum of PND 28 mouse brain coronal sections (Scale bars, 50 µm). (C) PLA signal in the corpus callosal (cc) region of indicated mouse strains using L1CAM antibody combined with either pan ankB antibody (Left) or giant ankB antibody (Right). Nuclei are in blue (Scale bars, 50 µm). (D) Quantification of PLA signal. Mean ± SEM; ***P = 0.003; ****P = 0.0001, 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n = 3.CC BY-NC-ND · PubMed Central
Fig. 5. Increased axonal branching in giant ankB mutant and L1CAM-Y1229H neurons. (A) Fluorescence images from GFP-transfected 4-d-old cultured WT and Abe37fsfs hippocampal neurons stained with MAP2 antibody. Images are inverted for a better contrast (axon branches are marked by red asterisks; blue arrowheads point to the cell body of the transfected neurons) (Scale bars, 50 µm). (B and C) Quantification of axon branching from the indicated conditions. Mean ± SEM; **P < 0.01, ***P < 0.001; ns, not significant; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; from left to right: n = 19, 22, and 11 for panel 1; n = 18, 19 for panel 2; n = 10, 10 for panel 3; n = 18, 15 for panel 4.
Fig. 5. Increased axonal branching in giant ankB mutant and L1CAM-Y1229H neurons. (A) Fluorescence images from GFP-transfected 4-d-old cultured WT and Abe37fsfs hippocampal neurons stained with MAP2 antibody. Images are inverted for a better contrast (axon branches are marked by red asterisks; blue arrowheads point to the cell body of the transfected neurons) (Scale bars, 50 µm). (B and C) Quantification of axon branching from the indicated conditions. Mean ± SEM; **P < 0.01, ***P < 0.001; ns, not significant; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; from left to right: n = 19, 22, and 11 for panel 1; n = 18, 19 for panel 2; n = 10, 10 for panel 3; n = 18, 15 for panel 4.CC BY-NC-ND · PubMed Central
Fig. 6. Giant ankB represses axon branching through preventing microtubule growth into filopodia. (A) Axons from single-frame and 2-min time-lapse stacked images of EB3-tdTM–transfected neurons (Scale bar, 5 µm). Red asterisks indicate EB3 comets and blue arrowheads indicate nonaxis EB3 comets. (B) Quantification of the number of nonaxis EB3 comets (mean ± SEM; **P < 0.01; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n > 10 cells). (C, Left) Time-point images from time-lapse movie represent the progress of invasion of microtubules into newly formed filapodia in an ABe37fs/fs (ABfs/fs) neuron axon shaft (Scale bar, 5 µm). (Right) Quantification of the number of the filopodia with EB3 comets invaded (mean ± SEM; **P < 0.01; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n > 10 cells). (D) Fluorescent images of 4-d-old ABe37fs/fs (ABfs/fs) neurons transfected with GFP and stained with MAP2, treated with DMSO or taxol (Scale bar, 20 µm). The number of axons longer than 10 µm is quantified (mean ± SEM; ***P < 0.001; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n = 15, 21, 10, 10 from top to bottom). (E) Schematic of branch formation in WT (up) versus ABe37fs/fs (ABfs/fs) or L1CAM-Y1229H (down) in axon shaft, where microtubules are indicated by either blue (WT) or magenta (ABfs/fs or L1-Y1229H).
Fig. 6. Giant ankB represses axon branching through preventing microtubule growth into filopodia. (A) Axons from single-frame and 2-min time-lapse stacked images of EB3-tdTM–transfected neurons (Scale bar, 5 µm). Red asterisks indicate EB3 comets and blue arrowheads indicate nonaxis EB3 comets. (B) Quantification of the number of nonaxis EB3 comets (mean ± SEM; **P < 0.01; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n > 10 cells). (C, Left) Time-point images from time-lapse movie represent the progress of invasion of microtubules into newly formed filapodia in an ABe37fs/fs (ABfs/fs) neuron axon shaft (Scale bar, 5 µm). (Right) Quantification of the number of the filopodia with EB3 comets invaded (mean ± SEM; **P < 0.01; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n > 10 cells). (D) Fluorescent images of 4-d-old ABe37fs/fs (ABfs/fs) neurons transfected with GFP and stained with MAP2, treated with DMSO or taxol (Scale bar, 20 µm). The number of axons longer than 10 µm is quantified (mean ± SEM; ***P < 0.001; 1-way ANOVA followed by Dunnett’s multiple-comparisons test; n = 15, 21, 10, 10 from top to bottom). (E) Schematic of branch formation in WT (up) versus ABe37fs/fs (ABfs/fs) or L1CAM-Y1229H (down) in axon shaft, where microtubules are indicated by either blue (WT) or magenta (ABfs/fs or L1-Y1229H).CC BY-NC-ND · PubMed Central
Fig. 7. Giant ankB colocalizes with L1CAM in periodic plasma membrane domains and coordinates cortical microtubules. (A) Schematic and transmission electron microscopy images of axon cross-sections at the corpus callosum. Regular spaced black dots are microtubules (Scale bars, 100 nm). (B) Histogram of distance of individual microtubules to the closest axon membrane (n = 219, 184, 208 in WT, ABfs/fs, L1-Y1229H mice accordingly with 2 mice per genotype, P < 0.0001 in both WT vs. ABfs/fs and WT vs. L1-Y1229H; unpaired t test). (C). STED images of axons in cultured neurons that were either phalloidin-labeled for actin (green) or double-labeled for giant ankB (green) or L1CAM (red) (Scale bars, 5 µm). (D) The staining intensity of actin (Upper Left) or giant ankB (Upper Right; Bottom) along the membrane is displayed. (E) Histogram of F-actin or giant ankB periodicities of indicated genotypes. For actin periodicity, P > 0.05; 1-way ANOVA. For giant ankB periodicity, P = 0.01 in WT vs. ABfs/fs, P = 0.0003 in WT vs. L1-Y1229H; 1-way ANOVA. (F) Schematic of giant ankB/L1CAM membrane complexes coordinating cortical microtubules in an axon.
Fig. 7. Giant ankB colocalizes with L1CAM in periodic plasma membrane domains and coordinates cortical microtubules. (A) Schematic and transmission electron microscopy images of axon cross-sections at the corpus callosum. Regular spaced black dots are microtubules (Scale bars, 100 nm). (B) Histogram of distance of individual microtubules to the closest axon membrane (n = 219, 184, 208 in WT, ABfs/fs, L1-Y1229H mice accordingly with 2 mice per genotype, P < 0.0001 in both WT vs. ABfs/fs and WT vs. L1-Y1229H; unpaired t test). (C). STED images of axons in cultured neurons that were either phalloidin-labeled for actin (green) or double-labeled for giant ankB (green) or L1CAM (red) (Scale bars, 5 µm). (D) The staining intensity of actin (Upper Left) or giant ankB (Upper Right; Bottom) along the membrane is displayed. (E) Histogram of F-actin or giant ankB periodicities of indicated genotypes. For actin periodicity, P > 0.05; 1-way ANOVA. For giant ankB periodicity, P = 0.01 in WT vs. ABfs/fs, P = 0.0003 in WT vs. L1-Y1229H; 1-way ANOVA. (F) Schematic of giant ankB/L1CAM membrane complexes coordinating cortical microtubules in an axon.CC BY-NC-ND · PubMed Central
Fig. 8. ABe37fs mice have a gain of stochastic connectivity by DTI. (A) Brains of WT (AB+/+) and ABe37fs/fs (ABfs/fs) male PND 28 mice imaged based on fractional anisotropy in black and white in coronal sections (Left) and in pseudocolor in 3D (Right). Positions of the corpus callosum (arrowheads), fimbria (green asterisks), and anterior commissure (black box) are indicated. (B) Histogram of pair-wise connectivity from WT and ABe37fs/fs brains. (C) Quantification of interhemispheric voxel asymmetry analysis from tracks greater than 7 mm in length originating in the whole brain or whole cortex. n = 5 mice per genotype; mean ± SEM; Kolmogorov–Smirnov test, 2-way, unpaired t test for B; 2-way, unpaired t test for C; *P < 0.05; ns, not significant.
Fig. 8. ABe37fs mice have a gain of stochastic connectivity by DTI. (A) Brains of WT (AB+/+) and ABe37fs/fs (ABfs/fs) male PND 28 mice imaged based on fractional anisotropy in black and white in coronal sections (Left) and in pseudocolor in 3D (Right). Positions of the corpus callosum (arrowheads), fimbria (green asterisks), and anterior commissure (black box) are indicated. (B) Histogram of pair-wise connectivity from WT and ABe37fs/fs brains. (C) Quantification of interhemispheric voxel asymmetry analysis from tracks greater than 7 mm in length originating in the whole brain or whole cortex. n = 5 mice per genotype; mean ± SEM; Kolmogorov–Smirnov test, 2-way, unpaired t test for B; 2-way, unpaired t test for C; *P < 0.05; ns, not significant.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

Show all 46 references

Cited by

See all 102

Metadata

Authors
Yang, R., Walder-Christensen, K. K., Kim, N., Wu, D., Lorenzo, D. N., Badea, A., Jiang, Y.-H., Yin, H. H., Wetsel, W. C., & Bennett, V.
Affiliations
  • Duke University (US)
  • University of North Carolina at Chapel Hill (US)
Corresponding author
  • Walder-Christensen, Kathryn K.
  • Bennett, Vann
Venue
Proceedings of the National Academy of Sciences
Published
2019-07-08
Volume
116
Issue
30
Pages
15262-15271
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
OpenAlex
W2961212325
Cited by
99
References
46
Citations by year
2026: 6 · 2025: 14 · 2024: 15 · 2023: 15 · 2022: 15 · 2021: 13 · 2020: 19 · 2019: 2
FWCI
7.12
Type
journal article
Language
en
Open access
Locations
  • Proceedings of the National Academy of Sciences (publishedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • Carolina Digital Repository (University of North Carolina at Chapel Hill) (submittedVersion)
  • UNC Libraries
Funders
  • Howard Hughes Medical Institute (NA)
  • George Barth Geller Endowed Professorship (NA)
  • NIH (R21MH115155, F31NS096848, K01 AG041211, P41 EB015897)
  • Khan Family Grant (NA)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Autism Spectrum Disorder Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetics and Neurodevelopmental Disorders
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Neuroscience and Neuropharmacology Research
Keywords
ANK2, autism spectrum disorder, axon branching, synaptic function, mouse model, social behavior, executive function, neurodiversity, normal intelligence
MeSH terms
Autism Spectrum Disorder, Neuronal Outgrowth, Animals, Behavior, Animal, Cell Membrane, Disease Models, Animal, Humans, Male, Mice, Transgenic, Microtubules, Mutation, Neurons, Social Behavior, Synapses, Gene Expression, Alternative Splicing, Ankyrins, Neural Cell Adhesion Molecule L1, Mice, Gene Knock-In Techniques, Executive Function, Primary Cell Culture, Connectome

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: ANK2 autism mutation targeting giant ankyrin-B promotes axon branching and ectopic connectivity. https://app.cochat.ai/papers/10.1073/pnas.1904348116
The paper
Yang, R., Walder-Christensen, K. K., Kim, N., Wu, D., Lorenzo, D. N., Badea, A., Jiang, Y.-H., Yin, H. H., Wetsel, W. C., & Bennett, V. (2019). ANK2 autism mutation targeting giant ankyrin-B promotes axon branching and ectopic connectivity. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.1904348116

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.