Claims 56
From the paper Diversification of dentate gyrus granule cell subtypes is regulated by Nrg1 nuclear back-signaling · claims file · checked 2026-10-21: ready
- fig1f-sgc-soma-superficial
In WT mice SGC somata lie closer to the GCL-MOL boundary than GC somata (SGCs in the top cell body layers, under 20 micrometres from the surface).
SGC cell body locations were significantly different from those of GCs (**** P < 0.0001, Welch’s t test [two-tailed]; t = 8.471, df = 19.16)
Figure 1F; Results, SGCs have unique morphological and electrical properties, paragraph 2How it will be judged
Soma position is the vertical distance from the cell body centre to the GCL-MOL boundary (Methods, Imaging and image analysis). Two-tailed Welch's t test of WT SGC against WT GC positions, as in the legend. Reproduced when p < 0.0001 with SGCs nearer the boundary; partially reproduced when p < 0.05 with SGCs nearer the boundary.
- fig1g-sgc-wider-splay
WT SGCs have wider dendritic splay angles than WT GCs.
The splay angle (angle between the two farthest dendrites measured at the center of the cell body) was significantly larger in SGCs compared with GCs (* P = 0.0168, Welch’s t test [two-tailed]; t = 2.754, df = 12.59)
Figure 1GHow it will be judged
Splay angle as the Methods define it (FIJI angle tool, from the most laterally displaced dendritic termination through the soma centre to the opposite one). Two-tailed Welch's t test. Reproduced when p < 0.05 with SGCs wider.
- fig1h-gc-more-branch-points
WT GCs have more dendritic branch points per primary dendrite than WT SGCs.
The number of branch points was significantly higher in GCs compared with SGCs (** P = 0.0023, Mann–Whitney test, U = 2)
Figure 1HHow it will be judged
Branch points from the Imaris reconstruction divided by the number of primary dendrites (legend). Two-tailed Mann-Whitney test (the legend does not state tails, but U = 2 with 11 and 5 cells gives the reported P only two-tailed). Reproduced when p < 0.05 with GCs higher.
- fig1i-gc-longer-dendrites
WT GCs have greater dendritic length per primary dendrite than WT SGCs.
The normalized dendritic length was significantly higher in GCs compared with SGCs (*** P = 0.0004, Welch’s t test [two-tailed]; t = 5.348, df = 9.366)
Figure 1IHow it will be judged
Total dendritic length divided by the number of primary dendrites (legend). Two-tailed Welch's t test. Reproduced when p < 0.05 with GCs longer.
- fig1j-sgc-hyperpolarized-rmp
In WT mice SGCs have a more hyperpolarized resting membrane potential than GCs.
SGCs have a significantly more hyperpolarized resting membrane potential compared with GCs (* P = 0.02, Mann–Whitney test, U = 26; N = 14 cells [GCs] and 9 cells [SGCs] from 5 mice)
Figure 1J; Results, SGCs have unique morphological and electrical properties, paragraph 2How it will be judged
RMP per cell as defined in the opening comment (code, baseline before step onset). WT SGCs against WT GCs by two-tailed Mann-Whitney test (tails are not stated, but U = 26 with 14 and 9 cells gives P = 0.02 only two-tailed). Reproduced when p < 0.05 with SGCs more negative. The cell counts are compared with 14 and 9, and U with 26.
- fig1k-sgc-narrower-spike
In WT mice SGCs have narrower action potentials than GCs.
SGCs have a significantly narrower spike width compared with GCs (* P = 0.012, Mann–Whitney test, U = 24
Figure 1K; Results, SGCs have unique morphological and electrical properties, paragraph 2How it will be judged
Spike width at half height between threshold and peak of the first spike at rheobase, threshold at dV/dt 20 mV/ms (Kim et al, the Methods' reference); the code's 10 mV/ms threshold is shown as a variant. Two-tailed Mann-Whitney test (U = 24 with 14 and 9 cells gives P = 0.012 only two-tailed). Reproduced when p < 0.05 with SGCs narrower.
- fig1l-sgc-faster-downstroke
In WT mice SGCs have a faster action potential downstroke than GCs.
SGCs have a significantly faster downstroke compared with GCs (* P = 0.046, Mann–Whitney test, U = 31
Figure 1L; Results, SGCs have unique morphological and electrical properties, paragraph 2How it will be judged
Downstroke as the minimum dV/dt on the repolarizing phase of the first spike at rheobase (Kim et al). The code's window, from 5 ms before the first peak to the end of the first 500 ms of the step, is a second reading, as the code disagrees with the Methods' reference only where the rheobase sweep holds more than one spike. Two-tailed Mann-Whitney test (U = 31 with 14 and 9 cells gives P = 0.046 only two-tailed). Reproduced when p < 0.05 with SGCs more negative under every reading; blocked as methods_underspecified when the readings disagree, with each shown.
- fig1m-phase-plots
Phase plots of the action potential at rheobase show differences in waveform kinetics between WT GCs and SGCs.
The pronounced differences in the kinetics of the SGC action potentials can be seen in the phase plots for APs at rheobase
Figure 1M; Results, SGCs have unique morphological and electrical properties, paragraph 2How it will be judged
For each cell, dV/dt against Vm minus threshold for the first spike at rheobase, dV/dt as the first difference over the time step (makephaseplot.m). Shown as a figure; the claim is judged as reproduced when fig1k-sgc-narrower-spike and fig1l-sgc-faster-downstroke are, since the paper reads the plot through those two measures.
- figs1a-f-properties-covary-with-position
In WT cells (GCs and SGCs together), RMP, spike width, downstroke, rheobase, sag and upstroke-downstroke ratio covary with soma position.
We found that RMP ( Fig S1A , P = 0.0129), spike width ( Fig S1B , P = 0.0543), and downstroke ( Fig S1C , P = 0.0139) significantly covaried with soma position ... We also found that rheobase ( Fig S1D , P = 0.0475), sag ( Fig S1E , P = 0.0359), and upstroke–downstroke ratio ( Fig S1F , P = 0.0131) significantly covaried with soma positions in the GCL
Figure S1A to F; Results, SGCs have unique morphological and electrical properties, paragraph 3How it will be judged
The legend reports P and R2 per property, so each is a simple linear regression of the property on soma position over the 23 WT cells (14 GC, 9 SGC), as fitted in Prism (Methods, Statistical analysis). The paper counts spike width (P = 0.0543) as significant, so a property matches when its slope p is at most 1.1 times the reported p or its R2 is within tolerance. Reproduced when all six match; partially reproduced when four or five match. Feature definitions as in the opening comment; the code's variants are shown.
- figs1-remaining-properties-no-difference
The remaining electrophysiological properties differ neither between WT GCs and SGCs nor with soma position.
The remainder of the properties did not differ significantly between GCs and SGCs or by soma position
Results, SGCs have unique morphological and electrical properties, paragraph 3How it will be judged
A claim of absence, judged on the paper's tests at p < 0.05. For each property other than the six of figs1a-f and the three of Fig 1J to L, WT GC against SGC by Welch's t test when both groups pass Shapiro-Wilk and Kolmogorov-Smirnov normality tests and by Mann-Whitney test otherwise (Methods, Statistical analysis), and a linear regression on soma position as in Fig S1. Which properties make up the remainder depends on the 20 features, which the paper does not list (opening comment); the readings are (a) the 18 features of Fig S2C less the five named (RMP, spike width, downstroke, rheobase, sag), and (b) the 21 outputs of the code less those five. Reproduced when no property in the reading finds an effect; not reproduced when any does. Every test's p and effect size is shown.
- figs1g-l-gc-only-position
Among WT GCs alone, downstroke and sag still covary with soma position, while RMP, spike width, rheobase and upstroke-downstroke ratio do not, so the GC-SGC differences in RMP and spike width are cell-type features.
we found that downstroke and sag were still significantly correlated with soma positions ( Fig S1I , downstroke, P = 0.006, and Fig S1K , sag, P = 0.03), whereas the rest of the properties did not correlate with soma positions ( Fig S1G , RMP, P = 0.22, Fig S1H , spike width, P = 0.1, Fig S1J , rheobase, P = 0.06, and Fig S1L , upstroke–downstroke ratio, P = 0.4)
Figure S1G to L; Results, SGCs have unique morphological and electrical properties, paragraph 3How it will be judged
Simple linear regressions on soma position over the 14 WT GCs, as in Fig S1A to F. Reproduced when downstroke and sag have p < 0.05 and RMP, spike width, rheobase and the ratio have p >= 0.05 (the absence half judged on this test as written); partially reproduced when the two significant ones hold and one of the four absences fails. Every slope, p and R2 is shown.
- fig2a-v321l-more-multiple-dendrite-gcs
Golgi staining labels the same total number of GCs in WT and V321L DG, but V321L DG has a higher percentage of GCs with multiple primary dendrites and a lower percentage with a single one.
There was no difference in the total number of cells labeled by the Golgi stain between genotypes ( Fig 2A , left; P = 0.68). However, V 321 L mice had a significantly higher number of Golgi-stained GCs with multiple primary dendrites compared with WT mice and a significantly lower number of GCs with single primary dendrites compared with WT mice
Figure 2A; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 1How it will be judged
Per mouse, the percentage of Golgi-stained GCs with one and with several primary dendrites. Two-way ANOVA (genotype by dendrite class) with Bonferroni post hoc comparisons (the Methods' parametric correction). The design is not named; the reported F(1,7) fits dendrite class as a repeated measure within mouse (error df = mice minus 2, so 9 mice) and not an ordinary two-way ANOVA (16 error df with 5 mice per genotype), so the repeated-measures ANOVA is the test and the ordinary one is shown as a variant. Reproduced when the interaction has p < 0.05, both genotype comparisons have p < 0.05 in the stated directions, and the total count does not differ by two-tailed Welch's t test at p < 0.05. The Results print P = 0.68 and the legend P = 0.7 for the total; both are recorded as one test.
- fig2c-v321l-sgc-position-not-different
In V321L mice SGC and GC soma positions do not differ; SGCs are found throughout the GCL.
SGC cell body locations are not significantly different to those of GCs ( P = 0.9012, Welch’s t test [two-tailed]; t = 0.1256, df = 21.78
Figure 2B, 2C and 2H; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 2How it will be judged
A claim of absence, judged on two-tailed Welch's t test of V321L SGC against V321L GC soma positions at p < 0.05. Reproduced when p >= 0.05; the difference in means is shown.
- fig2d-v321l-splay-not-different
In V321L mice the splay angle does not differ between GCs and SGCs.
The splay angle was not significantly different between GCs and SGCs
Figure 2DHow it will be judged
A claim of absence, judged on two-tailed Welch's t test at p < 0.05. Reproduced when p >= 0.05; the difference is shown.
- fig2ef-v321l-dendritic-complexity-preserved
In V321L mice GCs still have more branch points and greater dendritic length per primary dendrite than SGCs.
In contrast, the differences in dendritic complexity between GCs and SGCs were preserved in the V 321 L mutant DG
Figure 2E and 2FHow it will be judged
Measures as in Fig 1H and 1I. Reproduced when both tests give p < 0.05 with GCs higher (Mann-Whitney two-tailed, as U = 13 with 9 and 13 cells gives P = 0.0013 only two-tailed); partially reproduced when one does.
- fig2g-morphology-lda-separates-cell-type
Linear discriminant analysis of 15 morphological parameters separates GCs from SGCs along LD1, but not WT from V321L cells.
LDA resulted in successful capture of cell-type differences between GCs and SGCs (along the primary classifier ... LD1) but not between genotypes
Figure 2G and Figure S2A; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 2How it will be judged
Features z-scored with scale(), then LDA (MASS lda), as the Methods (LDA) describe. Whether PCA precedes this LDA, as it does for the classifier, is not stated; readings are (a) LDA on the z-scored features and (b) LDA on their top 9 PCs, and the claim is blocked as methods_underspecified when they disagree. The groups are not stated; Fig 2G shows LD1 and LD2, which needs at least three classes, so the reading is four classes (genotype by cell type). The parameter set is the morphological variables of Fig S2A. Reproduced when the LD1 scores of GCs and SGCs separate (a two-tailed Mann-Whitney test of LD1 by cell type at p < 0.05 and at least 90% of cells on their type's side of the midpoint between the type medians) while LD1 does not differ by genotype within either cell type (Mann-Whitney p >= 0.05). The 90% bound is not from the paper; the overlap is reported either way.
- figs2b-morphology-kmeans-by-cell-type
k-means clustering (k = 3) of morphological data groups cells by cell type, with all WT GCs in one cluster and all WT SGCs in another; all but one V321L GC and all but two V321L SGCs fall in the corresponding clusters.
All WT GCs and SGCs segregate into k1 and k2, respectively. All but one V 321 L GCs occupy k1, whereas the one was in k3. All but 2 V 321 L SGCs occupy k2 with one in k1 and another in k3
Figure S2BHow it will be judged
k-means with k = 3 on the morphological data, on the PCA scores shown in the panel. The paper gives neither the number of PCs, the scaling nor the seed; readings are k-means on (a) the z-scored features and (b) the first two PCs of the z-scored features, each with 100 random starts (a standard choice). Reproduced when, under every reading, the cluster that holds most WT GCs holds all WT GCs and no WT SGC and the counts of V321L cells outside their type's cluster are within one of the paper's (1 GC, 2 SGCs).
- fig2i-v321l-rmp-not-different
In V321L mice GCs and SGCs do not differ in RMP; both lie near the WT SGC median.
Unlike WT GCs and SGCs, we found no differences in electrical properties between V 321 L GCs and SGCs ( Fig 2I–K , RMP, P = 0.9
Figure 2I; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 3; Discussion, Composition of cell-type identity, paragraph 1How it will be judged
A claim of absence, judged on Welch's t test of V321L GC against SGC RMP at p < 0.05. Reproduced when p >= 0.05; the difference and both group medians, beside the WT GC and SGC medians, are shown.
- fig2j-v321l-spike-width-not-different
In V321L mice GCs and SGCs do not differ in spike width.
Spike width ( P = 0.2776, Mann–Whitney test, U = 55)
Figure 2JHow it will be judged
A claim of absence, judged on a two-tailed Mann-Whitney test at p < 0.05, with spike width as in fig1k-sgc-narrower-spike. Reproduced when p >= 0.05; the code's threshold is shown as a variant and the medians are shown.
- fig2k-v321l-downstroke-not-different
In V321L mice GCs and SGCs do not differ in downstroke.
Downstroke ( P = 0.5392, Mann–Whitney test, U = 63.5)
Figure 2KHow it will be judged
A claim of absence, judged on a two-tailed Mann-Whitney test at p < 0.05, with downstroke as in fig1l-sgc-faster-downstroke (Kim et al definition). Reproduced when p >= 0.05; the code's window is shown as a variant.
- fig2m-ephys-lda-wt-gc-apart
LDA of electrical properties separates WT GCs from all other groups (WT SGCs, V321L GCs, V321L SGCs) along LD1, driven mainly by spike width and downstroke.
LDA resulted in separation of WT GCs and all other groups (WT SGCs, V 321 L GCs, and V 321 L SGCs) along LD1 driven strongly by spike width
Figure 2M and Figure S2C; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 3How it will be judged
The 18 features of Fig S2C, z-scored with scale() (Methods, LDA), LDA (MASS lda) with four classes (genotype by cell type; Fig 2M shows LD1 and LD2). Whether PCA precedes this LDA as it does for the classifier is not stated; readings are (a) LDA on the z-scored features and (b) LDA on their top 9 PCs. Cells with a missing feature are dropped. Reproduced when, under every reading, the WT GC mean LD1 lies on one side and the means of the other three groups on the other, and spike width or downstroke has the largest absolute LD1 coefficient; blocked as methods_underspecified when the readings disagree.
- figs2d-ephys-kmeans-v321l-like-wt-sgc
k-means clustering (k = 3) of electrophysiological data places V321L GCs and SGCs in clusters in nearly the same proportions as WT SGCs, while WT GCs spread across all three clusters.
WT GCs are evenly distributed across all clusters, whereas SGCs largely occupy k3 with a minority in k1. All V 321 L GC and SGC cluster occupancy was virtually identical to WT SGCs
Figure S2D; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 3How it will be judged
k-means with k = 3 on the 18 features of Fig S2C. The scaling, number of PCs and seed are not given; readings are (a) z-scored features and (b) the first two PCs of the z-scored features, 100 random starts each. Reproduced when, under every reading, WT GCs occupy all three clusters with none holding more than half of them, and WT SGCs, V321L GCs and V321L SGCs each have most cells in one same cluster.
- v321l-multiple-regression-wt-only
Controlling for soma position, spike amplitude and upstroke covary with cell type in WT mice, while no electrophysiological property shows a cell-type effect in V321L mice.
In WT mice, we found that two electrophysiological properties, spike amplitude ( P = 0.042, β1 = −46.2) and upstroke ( P = 0.038, β1 = −117.88), significantly covaried with cell type when controlling for soma position. However, this model did not detect significant cell-type effects for any electrophysiological properties in V 321 L mice
Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 4How it will be judged
Per property and genotype, lm(property ~ CellType + SomaPosition + CellType x SomaPosition) in R (Methods, Statistical analysis), soma position in micrometres from the GCL-MOL boundary. The coding of cell type is not stated; readings are SGC = 1 and GC = 1 (the sign of beta1 flips between them, so a beta1 matches when its magnitude is within tolerance and its sign fits one coding used for both properties). Reproduced when in WT the beta1 p-values for spike amplitude and upstroke are below 0.05 with beta1 within tolerance, and in V321L no property's beta1 has p < 0.05 (the absence half judged on this test). The property set for the V321L half is open (opening comment); readings are the 18 features of Fig S2C and the 21 outputs of the code. The WT properties other than these two are shown.
- figs2e-pca-separation-lost
PCA of the electrophysiological features separates WT GCs from WT SGCs on the first two components, while V321L GCs and SGCs overlap.
Principal component analysis (PCA) revealed a clear separation between GCs and SGCs in WT mice along the first two principal components, whereas cells in V 321 L mice showed substantial overlap
Figure S2E; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 5How it will be judged
PCA (prcomp) of the 18 z-scored features of Fig S2C over all cells, with 95% ellipses per group as in the legend. The ellipse type is not stated; readings are (a) the 95% normal data ellipse (ggplot2 stat_ellipse default) and (b) the 95% confidence ellipse of the group mean. Reproduced when, under every reading, the WT GC and WT SGC centroids each lie outside the other group's ellipse and the V321L GC and SGC centroids each lie inside the other's; blocked as methods_underspecified when the readings disagree. The centroid distances are shown.
- figs2f-centroid-separation-reduced
In the WT-trained LDA, the distance between GC and SGC centroids along LD1 is much smaller in V321L than in WT, with both V321L centroids on the WT SGC side of the boundary.
Centroid analysis demonstrated a striking reduction in the separation between GC and SGC centroids in V 321 L mice along the first linear discriminant (LD1) axis
Figure S2F; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 5How it will be judged
The classifier of figs2g-lda-classifier-accuracy; LD1 scores of WT cells and of V321L cells (projected through the WT PCA and LDA). Reproduced when the V321L GC-SGC centroid distance on LD1 is less than half the WT distance and both V321L centroids lie on the WT SGC side of the LD1 decision boundary. The bound of one half is read from Fig S2F (about 1.3 against 3.3 units) and is not stated in the text.
- figs2g-lda-classifier-accuracy
An LDA classifier trained on WT electrophysiology classifies held-out WT cells with 78% accuracy (Cohen's kappa 0.57) but V321L cells with only 52% accuracy (kappa 0.08).
The LDA classifier, trained on 80% of the WT data, achieved 78% accuracy in cell-type classification of the remainder 20% WT cells (Cohen’s κ = 0.57) ( Fig S2G , top). Using this classifier on the recordings obtained from V 321 L mice resulted in 52% classification accuracy (Cohen’s κ = 0.08)
Figure S2G; Results, V321L mutation in Nrg1 results in altered composition of the DG granule cell population, paragraph 5How it will be judged
As the Methods (LDA) describe. The 18 features of Fig S2C, z-scored with scale(); PCA with prcomp on WT; the top 9 PCs; caret train with method lda and trainControl(method cv, number 5); the WT PCA and LDA applied to the V321L data. Kappa tolerances are 0.1, not 10%, because with about 21 to 23 cells one reclassified cell moves kappa by about 0.08. Open details, with readings run. (1) Whether V321L data are scaled with their own mean and SD or the WT ones ("applied to the scaled V321L data"). (2) Whether 78% is caret's mean accuracy over the five folds or the pooled out-of-fold accuracy; the WT confusion matrix in Fig S2G counts 21 cells (16 correct, 76%), so it pools all folds. (3) The fold seed, not given; each value is the median over 100 random fold assignments (a standard choice), with the range shown. Reproduced when under one reading all four values are within tolerance; partially reproduced when the two accuracies are, or when WT accuracy exceeds V321L accuracy by at least 15 points. The V321L matrix in Fig S2G (8, 5, 7, 3; 11 of 23 correct, 48%, kappa near -0.09) does not match the text's 52% and 0.08; both are compared and the mismatch is recorded.
- fig3b-penk-position-altered
In WT mice most Penk+ GCs lie in the top two cell body layers of the suprapyramidal GCL; this bias toward the GCL-MOL boundary is lost in V321L mice.
In WT mice, most of the Penk + GCs were in the top 2 cell body layers of the suprapyramidal GCL, whereas this bias toward the GCL-MOL boundary was lost in the V 321 L DG
Figure 3B; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 1How it will be judged
Distances of Penk+ somata from the GCL-MOL boundary in FISH images (cells with at least 5 puncta or large clusters; Methods), pooled per genotype. Two-sample Kolmogorov-Smirnov test. Reproduced when p < 0.05 with V321L cells farther from the boundary.
- fig3d-more-penk-gcs
V321L mice have a higher percentage of Penk-expressing GCs in snRNA-Seq.
the total percentage of Penk + GCs was significantly higher in V 321 L mice
Figure 3D; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 2How it will be judged
Per mouse, the percentage of GCs (clusters 0, 1, 5 and 18) with a raw Penk count above 0 (snMultiome code). Two-tailed Mann-Whitney test. Reproduced when p < 0.05 with V321L higher (U = 0 gives the smallest p available with 4 and 4, 0.029).
- figs3b-prox1-count-not-different
The total number of Prox1+ GCs does not differ between WT and V321L mice.
We found no differences in the total number of Prox1 + neurons (GCs) ( Fig S3B ; Mann–Whitney test, P = 0.69, U = 6)
Figure S3B; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 2How it will be judged
A claim of absence. The Results and the legend name different tests; both are the paper's, so both are run at p < 0.05 on per-mouse Prox1+ GC counts. Reproduced when neither finds a difference.
- figs3c-cluster18-penk-enriched
Among the four GC clusters (0, 1, 5, 18), cluster 18 has the highest percentage of Penk+ cells in both genotypes.
cluster 18 contained the highest percentage of Penk + cells indicating enrichment for SGCs in cluster 18 in both genotypes
Figure S3C; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 2How it will be judged
Per mouse and cluster, percentage of Penk+ GCs. Two-way repeated-measures ANOVA (cluster by genotype, cluster within mouse) with Bonferroni comparisons of each cluster against 18. Reproduced when the cluster effect and all three comparisons have p < 0.05 with cluster 18 highest.
- figs3d-sorcs3-marks-cluster18
Sorcs3 is the distinguishing marker of GC cluster 18, and Sorcs3+ cells are biased toward the GCL-MOL boundary in FISH.
However, cluster 18 was uniquely marked by the expression of Sortilin Related VPS10 Domain Containing Receptor 3 ( Sorcs3 )
Figure S3D; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 2How it will be judged
Seurat FindConservedMarkers across WT and V321L for each GC cluster (Methods). Reproduced when Sorcs3 is among the conserved markers of cluster 18 and not of clusters 0, 1 or 5. The FISH half is not tested here (no statistic is given).
- figs5a-penk-sorcs3-subsets
V321L mice have more Penk+Sorcs3+ and Penk+Sorcs3- GCs than WT, while Sorcs3+Penk- GCs are unchanged.
V 321 L mice also showed an increase in the percentage of Penk + Sorcs3 + GCs ( Fig S5A ; P = 0.009). We found that there were no significant differences in the proportion of Penk-Sorcs3+ cells between WT and V 321 L mice
Figure S5A; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 3How it will be judged
Per mouse, the percentage of GCs in each marker class (raw count above 0 for each gene). Two-way ANOVA (marker class by genotype) with Bonferroni comparisons of genotype within class. The design is not named; the reported F(2, 12) fits marker class as a repeated measure within mouse (error df (3 - 1) x (8 - 2) = 12) and not an ordinary two-way ANOVA (18 error df), so the repeated-measures ANOVA is the test and the ordinary one is shown as a variant. Reproduced when the two Penk+ classes differ at p < 0.05 with V321L higher and Sorcs3+Penk- does not (p >= 0.05).
- fig4a-more-penk-sorcs3-cells-fish
In FISH, V321L mice have more Penk+Sorcs3+ cells in the suprapyramidal blade than WT mice.
We found that V 321 L mice had significantly higher numbers of cells co-expressing Penk and Sorcs3
Figure 4A; Results, Discovery of novel marker genes for SGCs, paragraph 1How it will be judged
Per mouse, the count of Penk+Sorcs3+ cells in the suprapyramidal blade (averaged over its sections, each dot one mouse). Two-tailed Welch's t test. Reproduced when p < 0.05 with V321L higher.
- fig4a-prime-penk-sorcs3-position
Penk+Sorcs3+ cells in V321L DG are not restricted to the layers at the GCL-MOL boundary.
the V 321 L mice showing more Penk+Sorcs3+ cells present in the lower layers of the GCL, away from the GCL-MOL boundary
Figure 4a'How it will be judged
Distances of Penk+Sorcs3+ cells from the GCL-MOL boundary pooled per genotype; two-sample Kolmogorov-Smirnov test. Reproduced when p < 0.05 with V321L cells farther from the boundary.
- fig4b-penk-enriched-genes
Differential expression finds 102 genes significantly enriched in Penk+ GCs over Penk- GCs (fold change of at least 2, adjusted P at most 0.05).
We assessed differential gene expression in Penk + GCs compared with Penk -negative GCs yielding 102 genes whose expression was significantly enriched in Penk + GCs (fold change of ≥2 and adj. P ≤ 0.05)
Figure 4B; Results, Discovery of novel marker genes for SGCs, paragraph 2How it will be judged
Seurat FindMarkers of Prox1+Penk+ against Prox1+Penk- GCs (snMultiome code, workbook 58), log2 fold change at least 1 and adjusted p at most 0.05. Reproduced when the count is within tolerance.
- fig4b2-marker-penk-cells-increased
For each selected SGC marker gene except Grp, V321L mice have a higher proportion of GCs co-expressing Penk and that marker.
Except for gastrin-releasing peptide ( Grp ) ... all other marker gene Penk-co-expressing cells formed a significantly higher proportion of all GCs in the V 321 L mice
Figure 4b''; Results, Discovery of novel marker genes for SGCs, paragraph 2How it will be judged
Per mouse, cells co-expressing Penk and each marker divided by all cells in clusters 0, 1, 5 and 18 (legend). Mann-Whitney test per gene, FDR by the two-stage Benjamini-Krieger-Yekutieli method. The nine genes are those of Fig 4b' and 4b''. With 4 and 4 mice the smallest two-tailed p is 0.029, so the threshold is the paper's FDR adjusted p of 0.03 as printed. The Results and the legend disagree on what 0.03 is (the Results give P = 0.03 with q = 0.002, and for Grp P = 0.1 with q = 0.01; the legend gives FDR adjusted P = 0.03, and 0.14 for Grp), so the judgement rests on U and the q values are shown. Reproduced when every gene except Grp has U = 0 with V321L higher and Grp does not.
- fig4d-penk-marker-overlap
In WT FISH, 84% of Penk+ GCs express Nptx2, 67% express Sorcs3, and 64% express both.
We found that on average 84% of the Penk + GCs also expressed Nptx2 , and 67% of the Penk + GCs also expressed Sorcs3 . The co-expression of Nptx2 and Sorcs3 was detected in 64% of the Penk + GCs
Figure 4D and 4d'; Results, Discovery of novel marker genes for SGCs, paragraph 2How it will be judged
Per mouse, percentages over Penk+ cells in the suprapyramidal blade, then the mean over the three mice; tolerances are the SDs the legend gives. Reproduced when all three are within tolerance.
- figs6-increase-in-nrg1-expressing
The increase in marker+Penk+ GCs in V321L mice is seen in Nrg1-expressing Penk+ GCs and not in Nrg1-non-expressing ones; about 94% of Penk+ GCs co-express Nrg1 in FISH.
Intriguingly, this increase in numbers of marker + Penk + cells was only observed in Penk and Nrg1 co-expressing GCs ( Fig S6A ) and not in Nrg1 -non-expressing Penk + GCs ( Fig S6B ). In FISH experiments for Penk and Nrg1 , we found that most of the Penk + GCs co-expressed Nrg1 (∼94%)
Figure S6A and S6B; Results, Discovery of novel marker genes for SGCs, paragraph 2How it will be judged
As fig4b2-marker-penk-cells-increased, split by Nrg1 raw count above 0. The paper's threshold here is FDR adjusted p < 0.1 (legend). Reproduced when in Penk+Nrg1+ GCs every marker except Grp reaches adjusted p < 0.1 with V321L higher, and in Penk+Nrg1- GCs fewer markers reach it. The FISH percentage is checked only if the FISH counts are available.
- fig5b-pseudotime-sorcs3-top
In WT GCs the top gene explaining pseudotime is Sorcs3, and late pseudotime falls at the end of cluster 1 and in cluster 18; in V321L GCs trajectories are altered and the latest pseudotime is in cluster 5.
The top gene explaining pseudotime in WT DG was found to be Sorcs3 ... Trajectories in the V 321 L DG were highly altered, and cells with late pseudotime were present in cluster 5
Figure 5B; Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 1How it will be judged
Monocle3 on the GC clusters per genotype, root in cluster 0 (highest Camk4), graph_test with neighbor_graph principal_graph (Methods; snMultiome code lines 2948 to 3228). Reproduced when Sorcs3 has the highest Moran's I among genes with q < 0.05 in WT, the WT cells in the top decile of pseudotime are mostly in clusters 1 and 18, and the V321L top decile is mostly in cluster 5. The decile is not from the paper.
- trajectory-geometry-directionality
WT (original and downsampled) and V321L GC trajectories have significantly lower spherical distances than randomized ones, so they have definite directionality; WT spherical distances exceed V321L ones.
WT original and downsampled trajectories resulted in significantly lower spherical distances compared with randomized ones indicating the presence of a definite directionality within the trajectories ... the spherical distances obtained for the WT trajectories were greater than those for the V 321 L trajectory
Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 1How it will be judged
TrajectoryGeometry analyseSingleCellTrajectory on the top 50 PCs rescaled to 0 to 100 with Monocle3 pseudotime, nSamples 1000, N 1000, randomization byPermutation and permuteWithinColumns, statistic mean (Methods). The downsampling seed is not given; the downsampled WT result is the median over 20 draws. Reproduced when all three tests give p < 0.05 with lower distances than random and the mean WT distance exceeds the V321L one.
- fig5d-atac-cluster-composition
In ATAC space, the Penk+ GC cluster 9 is populated evenly by WT and V321L nuclei, while cluster 4 (Camk4+ GCs) is about 75% WT.
Cluster 9 was comprised of the Penk + GCs, and it was uniformly populated by WT and V 321 L GCs; however, cluster 4 that was comprised of the Camk4 + GCs (cluster 0 in the “RNA space”) was disproportionately populated by WT GCs, which made up ∼75% of all cells in that cluster
Figure 5d'; Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 3How it will be judged
Clustering of GC nuclei in ATAC space at resolution 0.5 (Methods; ATAC code). Reproduced when the WT share of cluster 4 is within tolerance and that of cluster 9 is between 40 and 60% (a reading of "uniformly", not stated by the paper).
- fig5f-sgc-genes-broadly-accessible
SGC-enriched genes are accessible in cluster 9 and in nearly all other ATAC clusters of GCs in both genotypes, though their expression is largely restricted to cluster 18 in WT.
However, we also found SGC-enriched genes to be accessible in cells in nearly all other clusters for both genotypes
Figure 5E and 5F; Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 3How it will be judged
Signac GeneActivity per gene (TSS and 1 kb), nine genes of Fig 5E and 5F. Reproduced when for each gene the fraction of nuclei with nonzero activity is above zero in at least 8 of the 10 ATAC clusters in each genotype (a reading of "nearly all"), while in WT RNA the expressing fraction is highest in cluster 18.
- figs8b-penk-accessibility-not-different
The percentage of GCs with an accessible Penk locus does not differ between WT and V321L mice.
the total number of GCs having the Penk locus accessible was not significantly different between the genotypes
Figure S8b''; Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 3How it will be judged
A claim of absence, judged on a Mann-Whitney test at p < 0.05 of per-mouse percentages of GCs with nonzero Penk gene activity. Reproduced when p >= 0.05.
- tables4-suz12-enrichment
Genes enriched in Penk+ GCs and correlated with maximal pseudotime are strongly enriched for regulation by Suz12, a core PRC2 component.
We found strong enrichment for regulation of these genes by Suz12, a core component of the polycomb repressive complex 2 (PRC2)
Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 4; Table S4How it will be judged
The 60 genes (fig s10 intersection) submitted to ChEA and ENCODE. The paper names the databases, not the tool or library versions; the Methods use the enrichR package for GO terms, so the readings are the enrichR libraries ENCODE_and_ChEA_Consensus_TFs_from_ChIP-X, ChEA_2022 and ENCODE_TF_ChIP-seq_2015, and the claim is blocked as methods_underspecified when they disagree. Within a reading that gives ChEA and ENCODE separately, reproduced when SUZ12 is the top term by adjusted p in at least one of them; within the consensus library, when it is the top term.
- fig6a-penk-increases-with-age
The number of Penk+ cells in the suprapyramidal GCL of WT mice increases with age, from between 2 and 3 months.
We confirmed that WT mice indeed show an increase in Penk expression starting between 2 and 3 mo of age
Figure 6a'; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 1How it will be judged
Per mouse, the total count of Penk+ cells in the suprapyramidal GCL at 1, 2, 3 and 9 to 12 months. Brown-Forsythe ANOVA with Dunnett's T3 comparisons. The Results print P = 0.0032 and the legend P = 0.002. Reproduced when the age effect and the 2 against 3 month comparison have p < 0.05 with the count higher at 3 months.
- fig6b-sgc-count-increases-with-age
The number of Penk+Nptx2+Sorcs3+ cells (SGCs) in the suprapyramidal GCL of WT mice increases with age.
We found that the numbers of Penk+Nptx2+Sorcs3+ GCs also increased with age
Figure 6b'; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 1How it will be judged
Per mouse counts at 1, 2, 3 and 9 to 12 months; Kruskal-Wallis test with Dunn's comparisons. The Results print P = 0.0012 and the legend P = 0.001. Reproduced when both have p < 0.05 with 12 months higher than 1 month.
- figs9a-age-dynamics-lost-v321l
The age-dependent increase in SGC number seen in WT mice is lost in V321L mice.
These dynamics of SGC abundance over adolescence were lost in the V 321 L DG
Figure S9A; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 1How it will be judged
Two-way ANOVA (genotype by age) of per-mouse Penk+Nptx2+Sorcs3+ counts with Bonferroni comparisons between ages within genotype (legend). Reproduced when the WT 1 against 12 month comparison has p < 0.05 and no V321L comparison does (the absence half judged on these comparisons).
- figs9b-marker-fraction-stable
The percentage of Penk+ cells that are Nptx2+Sorcs3+ does not change with age in either genotype.
There were no significant differences in the overall percentage of Nptx2 - or Sorcs3 -expressing Penk + GCs at any of the ages
Figure S9B; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 1How it will be judged
A claim of absence, judged on Brown-Forsythe ANOVA per genotype at p < 0.05. Reproduced when neither genotype shows an age effect.
- figs9d-penk-not-immature
At 3 months, Penk+ cells in the GCL of WT mice do not express the immature neuron marker Dcx.
we did not find any Penk + cells that expressed the immature neuron marker doublecortin ( Dcx ) at 3 mo of age
Figure S9D; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 2How it will be judged
Count of Penk+Dcx+ cells in the GCL in FISH (cells with at least 5 puncta of each). Reproduced when the count is zero.
- figs10d-go-resting-membrane-potential
The 60 genes both enriched in Penk+ GCs and in the cluster 18 end-trajectory modules are enriched for gene ontology terms whose top term by odds ratio is regulation of resting membrane potential, driven by Kcnk1 and Kcnj2.
we identified 60 genes that met this criterion ... the feature of highest odds ratio being regulation of RMP, in line with our results showing a more hyperpolarized RMP in morphologically identified SGCs ( Fig 1J ). This enrichment was driven by two potassium channel genes ... Kcnk1 and Kcnj2
Figure S10C and S10D; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 2How it will be judged
Monocle3 find_gene_modules on WT GCs, modules correlated with the cluster 18 end trajectories, intersected with the Penk+ enriched genes; enrichR GO_Biological_Process_2023, terms with adjusted p < 0.05 ranked by odds ratio (legend and Methods). Reproduced when the count is within tolerance and the top term is regulation of resting membrane potential with Kcnk1 and Kcnj2 as its genes.
- fig7a-type3-nrg1-declines
Type III Nrg1 expression in the suprapyramidal DG of WT mice declines from 1 to 12 months of age; the decline from 1 to 3 months is not significant.
We found that Type III Nrg1 expression declined from 1 to 12 mo of age in WT mice; however, the reduction between 1 and 3 mo of age was not significant
Figure 7A; Results, Natural decline in capacity for Nrg1 nuclear back-signaling corresponds to the SGC fate, paragraph 1How it will be judged
Per mouse, Type III Nrg1 FISH fluorescence intensity density normalized to the area of the suprapyramidal blade (Methods). Kruskal-Wallis with Dunn's comparisons. Reproduced when the age effect and 1 against 12 months have p < 0.05 with lower values at 12 months, and 1 against 3 months has p >= 0.05.
- fig7bc-bace1-low-in-sgcs
Bace1 expression peaks at intermediate pseudotime and drops at maximal pseudotime while Nrg1 rises along pseudotime; fewer than 10 WT cells co-express Penk, Nrg1 and Bace1, so most Penk+Nrg1+ GCs lack detectable Bace1.
Bace1 expression was found to be higher in cells at intermediate pseudotime compared with the beginning or end of the trajectories ... examination of cells expressing Penk, Nrg1 , and Bace1 in WT DG resulted in <10 cells showing co-expression of all three markers
Figure 7B and 7C; Results, Natural decline in capacity for Nrg1 nuclear back-signaling corresponds to the SGC fate, paragraph 2How it will be judged
WT GC pseudotime from fig5b-pseudotime-sorcs3-top, split into thirds. Reproduced when mean Bace1 expression is highest in the middle third and lower in the last third than the middle, mean Nrg1 is highest in the last third, and fewer than 10 WT GCs have raw counts above 0 for all of Penk, Nrg1 and Bace1. The thirds are not from the paper.
- figs5bc-penk-sorcs3-neg-outside-cluster18
Penk+Sorcs3- GCs are excluded from cluster 18 in both WT and V321L mice.
we found that Penk + Sorcs3 - cells were excluded from cluster 18 in both WT and V 321 L mice
Figure S5B and S5C; Results, V321L mice have a higher number and altered positioning of Penk+ GCs compared with WT mice, paragraph 3How it will be judged
GCs with a raw Penk count above 0 and a raw Sorcs3 count of 0 (as in figs5a-penk-sorcs3-subsets), counted per cluster and genotype. Reproduced when cluster 18 holds none of them in either genotype, as the legend of Fig S5C says ("Note the absence of cells in cluster 18 in both genotypes").
- figs7-downsampled-pseudotime-robust
With WT GCs downsampled to the V321L number, the pseudotime ordering is preserved and Sorcs3 still marks the cells with maximal pseudotime.
We found that the pseudotime ordering of GCs and associated gene expression states were robust to differences in sampling frequency, with Sorcs3 still representing cells with the maximal pseudotime
Figure S7A and S7B; Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 1How it will be judged
WT GCs downsampled at random to the number of V321L GCs, then Monocle3 and graph_test as in fig5b-pseudotime-sorcs3-top. The seed is not given; 20 draws, as in trajectory-geometry-directionality. Reproduced when in most draws Sorcs3 is among the top ten genes by Moran's I with q < 0.05 (the legend says "some of the top pseudotime-associated genes") and the Spearman correlation of each downsampled cell's pseudotime with its pseudotime in the full WT analysis is positive. The top-ten bound and the correlation are readings, not the paper's.
- figs8a-atac-pseudotime-cluster9
In ATAC space the Penk+ GCs (cluster 9) have the maximal pseudotime; WT GC trajectories have three end points and V321L GC trajectories one, ending at cluster 9.
Once again, we found that the Penk + GCs (now located in ATAC cluster 9) had the maximum pseudotime values based on chromatin accessibility ... However, in the V 321 L GCs, there was only one end point terminating at cluster 9
Figure S8A; Results, Non-SGCs harbor the potential to express an SGC-like transcriptomic state, paragraph 3How it will be judged
Monocle3 on the ATAC GC subset, run separately on WT (LL) and V321L (VV) cells (Methods, Single-nucleus ATAC-Seq analysis), root chosen from Camk4 and the other markers the Methods list. Reproduced when cluster 9 has the highest median pseudotime in each genotype and the learned principal graph has three leaf end points (other than the root) in WT and one in V321L. Monocle3 graph learning depends on its settings, which the Methods do not give beyond the workflow; the snMultiome code's settings are used.
- figs9c-embryonic-sgc-genes
At E18.5, cells expressing SGC-enriched genes (Penk, Scg2, Nrg1, Cit) lie at the dorsal border of the developing GCL, so marker expression is unlikely to reflect cellular age alone.
We found cells expressing SGC-enriched genes such as Penk , Scg2 , Nrg1 , and Cit in the dorsal border of the developing GCL at E18.5
Figure S9C; Results, SGC numbers in the DG increase during the adolescence-adulthood transition in WT mice, paragraph 1How it will be judged
A qualitative reading of public ISH images, with no statistic. Reproduced when the Allen E18.5 ISH images for each of the four genes show labelled cells at the dorsal border of the developing GCL. The image series are not given; any E18.5 series for the gene is a reading.
Pipeline
- Paper
- done (1 of 1)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none