A taxonomy of transcriptomic cell types across the isocortex and hippocampal formation

May combine two claims.

The GABAergic neuronal class contains 6 subclasses and 119 clusters; the glutamatergic neuronal class contains 28 subclasses and 241 clusters; the astrocyte/oligodendrocyte non-neuronal class contains 2 subclasses and 14 clusters; and the immune/vascular non-neuronal class contains 4 subclasses and 10 clusters.

“The GABAergic neuronal class contains 6 subclasses and 119 clusters; the glutamatergic neuronal class contains 28 subclasses and 241 clusters; the astrocyte/oligodendrocyte non-neuronal class contains 2 subclasses and 14 clusters; and the immune/vascular non-neuronal class contains 4 subclasses and 10 clusters”

May combine two claims.

For CGE, the Ntng1 HPF supertype does not express canonical pan-CGE marker Prox1 nor its subclass markers Vip, Sncg or Lamp5.

“For CGE, the Ntng1 HPF supertype does not express canonical pan-CGE marker Prox1 (Miyoshi et al., 2015; Rubin and Kessaris, 2013) nor its subclass markers Vip, Sncg or Lamp5”

• cites Miyoshi et al., 2015 sourceNot checked yet

  Not checked in the free run.

• cites Rubin and Kessaris, 2013Not checked yet

  Not checked in the free run.

May combine two claims.

All GABAergic neuron types in isocortex are shared with HPF, whereas HPF also contains additional GABAergic types unique to itself.

“we find that all GABAergic neuron types in isocortex are shared with HPF, whereas HPF also contains additional GABAergic types unique to itself”

May combine two claims.

Each neighborhood is further divided into 3 subclasses: Lamp5, Sncg and Vip in CGE, and Sst Chodl, Sst and Pvalb in MGE.

“Each neighborhood is further divided into 3 subclasses: Lamp5, Sncg and Vip in CGE, and Sst Chodl, Sst and Pvalb in MGE.”

May combine two claims.

The 'Other' neighborhood includes all non-neuronal subclasses, as well as two neuronal subclasses, Cajal-Retzius (CR) (glutamatergic, mostly in layer 1) and Meis2 (GABAergic, mostly in white matter).

“The ‘Other’ neighborhood, briefly mentioned here, includes all non-neuronal subclasses, as well as two neuronal subclasses, Cajal-Retzius (CR) (glutamatergic, mostly in layer 1) and Meis2 (GABAergic, mostly in white matter)”

May combine two claims.

Its clusters #5, 8 and 9 are HPF-specific and are marked by Rxfp3 which is found in CA3.

“its clusters #5, 8 and 9 are HPF-specific and are marked by Rxfp3 which is found in CA3”

May combine two claims.

Most clusters are shared by all isocortical areas and also by RHP regions.

“most clusters are shared by all isocortical areas, consistent with our previous observations (Tasic et al., 2018), and also by RHP regions”

• cites Tasic et al., 2018 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Glutamatergic neuron types from different HPF regions are highly distinct from but also homologous to those in isocortex.

“glutamatergic neuron types from different HPF regions are highly distinct from but also, surprisingly, homologous to those in isocortex”

May combine two claims.

The authors covered all of the adult mouse isocortex and HPF, analyzing more than 1.3 million cells with two different scRNA-seq platforms (10x and SMART-Seq).

“Here we cover all of the adult mouse isocortex and HPF, analyzing >1.3 million cells with two different scRNA-seq platforms (10x and SMART-Seq).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Clustering 10xv2 and SSv4 cells separately resulted in 332 10xv2 clusters and 324 SSv4 clusters.

“We first clustered 10xv2 and SSv4 cells separately, resulting in 332 10xv2 clusters and 324 SSv4 clusters.”

May combine two claims.

The greatest distinction in GABAergic interneuron type composition is between CTX and HIP itself; the RHP regions often contain both CTX and HIP clusters, with a few exceptions.

“The greatest distinction in GABAergic interneuron type composition is between CTX and HIP itself; the RHP regions often contain both CTX and HIP clusters, with a few exceptions.”

May combine two claims.

The taxonomy is also consistent with the large body of literature on cortical GABAergic interneurons and with the authors' Patch-seq study, with MET types corresponding well with the supertypes defined here.

“It is also consistent with the large body of literature on cortical GABAergic interneurons (Lim et al., 2018; Pelkey et al., 2017; Tremblay et al., 2016), and our Patch-seq study (Gouwens et al., 2020) with MET types defined in that study corresponding well with the supertypes defined here”

• cites Lim et al., 2018 sourceNot checked yet

  Not checked in the free run.

• cites Pelkey et al., 2017 sourceNot checked yet

  Not checked in the free run.

• cites Tremblay et al., 2016 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The Sst subclass has multiple HPF-enriched clusters, most of which are marked by Npffr1.

“The Sst subclass has multiple HPF-enriched clusters, most of which are marked by Npffr1.”

May combine two claims.

They developed a consensus clustering approach to combine the two datasets and derived a cell type taxonomy comprising 388 transcriptomic types, of which 364 are neuronal.

“We developed a consensus clustering approach to combine the two datasets and derived a cell type taxonomy comprising 388 transcriptomic types of which 364 are neuronal.”

May combine two claims.

They grouped the subclasses into 8 neighborhoods: 2 GABAergic (CGE and MGE), 5 glutamatergic (L2/3 IT, L4/5/6 IT Car3, PT, NP/CT/L6b, and DG/SUB/CA), and one 'Other' neighborhood.

“We grouped the subclasses into 8 neighborhoods, 2 GABAergic (CGE and MGE), 5 glutamatergic (L2/3 IT, L4/5/6 IT Car3, PT, NP/CT/L6b, and DG/SUB/CA), and one ‘Other’ neighborhood.”

May combine two claims.

Clusters #54-55 in supertype Vip Cbln4 HPF are marked by Qrfpr, also expressed in CA3.

“Clusters #54–55 in supertype Vip Cbln4 HPF are marked by Qrfpr, also expressed in CA3”

May combine two claims.

The current CTX-HPF GABAergic taxonomy is largely consistent with previous transcriptomic taxonomies derived from cortical areas VISp-ALM and MOp at the supertype level but exhibits notable ambiguity at the type/cluster level.

“The current CTX-HPF GABAergic taxonomy is largely consistent with previous transcriptomic taxonomies derived from cortical areas VISp-ALM (Tasic et al., 2018) and MOp (Yao et al., 2020) at the supertype level but exhibits notable ambiguity at the type/cluster level”

• cites Tasic et al., 2018 sourceNot checked yet

  Not checked in the free run.

• cites Yao et al., 2020Not checked yet

  Not checked in the free run.

May combine two claims.

In the MGE neighborhood, the Sst Chodl subclass remains as one group; the Sst and Pvalb subclasses are divided into 11 and 3 supertypes, respectively.

“In the MGE neighborhood, the Sst Chodl subclass remains as one group (representing long-range projecting Sst cells); the Sst and Pvalb subclasses are divided into 11 and 3 supertypes, respectively”

May combine two claims.

In the CGE neighborhood, the Lamp5, Sncg and Vip subclasses are divided into 4, 5, and 6 supertypes, respectively.

“In the CGE neighborhood, the Lamp5 (mostly neurogliaform cells), Sncg and Vip subclasses are divided into 4, 5 (one containing Vip cells), and 6 supertypes, respectively”

May combine two claims.

They constitute the largest proportion of cell types, with 14 subclasses that correspond well to specific layers (L2-6) and/or regions.

“They constitute the largest proportion of cell types, with 14 subclasses that correspond well to specific layers (L2–6) and/or regions.”

May combine two claims.

Many isocortical glutamatergic types are shared across multiple areas and exhibit gradient-like gene expression variations along the cortical sheet.

“Many isocortical glutamatergic types are shared across multiple areas and exhibit gradient-like gene expression variations along the cortical sheet.”

May combine two claims.

Some clusters are largely absent from HIP (e.g., Lamp5 Pax6, Sst Syndig1l and Sst Hpse supertypes), while others have CTX- or HPF-selective counterparts (e.g., Sst Myh8 and Sst Etv1 clusters in CTX versus those in HPF).

“Conversely, some clusters are largely absent from HIP (e.g. Lamp5 Pax6, Sst Syndig1l and Sst Hpse supertypes), while others have CTX- or HPF-selective counterparts (e.g. Sst Myh8 and Sst Etv1 clusters in CTX versus those in HPF).”

May combine two claims.

Sst Ctsc HPF is a highly distinct HPF-specific supertype; clusters #102-103 are marked by Rxfp3 which is expressed in CA3, while clusters #104-105 are marked by Cxcr4 expressed in the polymorphic layer of DG (DG-po, also known as the hilus).

“Sst Ctsc HPF is a highly distinct HPF-specific supertype; clusters #102–103 are marked by Rxfp3 which is expressed in CA3, while clusters #104–105 are marked by Cxcr4 expressed in the polymorphic layer of DG (DG-po, also known as the hilus)”

May combine two claims.

In CA1, supertype CA1-ve Gpc3 (#334-336) is most ventral, CA1 Lefty1 (#337-345) is next, and CA1-do Plekhg1 (#346-348) is most dorsal.

“In CA1, supertype CA1-ve Gpc3 (#334–336) is in the most ventral location, followed by CA1 Lefty1 (#337–345), and CA1-do Plekhg1 (#346–348) most dorsal.”

May combine two claims.

They performed PCA for CA1 and CA3 separately and found the top PC corresponded to a dorsal-ventral (Do-Ve) gradient.

“we performed PCA for CA1 and CA3 (excluding the Mossy Rgs12 supertype) separately and found the top PC corresponded to a dorsal-ventral (Do-Ve) gradient”

May combine two claims.

L4 IT CTX Rspo1 is an L4-specific supertype that predominantly contains cells from all sensory areas but also some from non-sensory isocortical areas.

“We identified a L4 specific supertype, L4 IT CTX Rspo1, which predominantly contains cells from all sensory areas but surprisingly, some from non-sensory isocortical areas as well”

May combine two claims.

The PT neighborhood (Bcl6+) is segregated into three region-specific subclasses, L5 PT CTX (Fam84b+), L4 RSP-ACA, and L5 PPP (the only HPF-specific cell type identified in this neighborhood).

“The PT neighborhood (Bcl6+) is segregated into three region-specific subclasses, L5 PT CTX (Fam84b+), L4 RSP-ACA, and L5 PPP (the only HPF-specific cell type identified in this neighborhood)”

May combine two claims.

In CA3, supertype CA3-do Iyd (#356-358) is dorsal and CA3-ve Fam107a (#351-355) is more ventral.

“In CA3, supertype CA3-do Iyd (#356–358) is in the dorsal location, and CA3-ve Fam107a (#351–355) more ventral.”

May combine two claims.

There is large divergence between dorsal clusters of SUB and CA1 along the proximal-distal axis, but convergence in the ventral parts of CA1 and ProS (and HATA).

“While there is a large divergence between the dorsal clusters of SUB and CA1 along the Pr-Di axis, there is a convergence in the ventral parts of CA1 and ProS (and HATA).”

May combine two claims.

They observe two parallel continuous transitions between L6 CT and L6b types for both CTX and ENT, and a continuous transition of L6b cells between CTX and all HPF regions.

“We observe two parallel continuous transitions between L6 CT and L6b types for both CTX and ENT, and a continuous transition of L6b cells between CTX and all HPF regions”

May combine two claims.

The DG/SUB/CA neighborhood comprises cells specifically located in CA1, CA2, CA3, ProS, SUB and DG, divided into 5 region-specific subclasses.

“Lastly, the DG/SUB/CA neighborhood comprises cells that are specifically located in CA1, CA2, CA3, ProS, SUB and DG which are divided into 5 region-specific subclasses.”

May combine two claims.

L5 PT CTX contains 4 supertypes: Chrna6 (enriched in posterior sensory areas and highly distinct from the other three supertypes), Npnt, Cdh13 and C1ql2 (enriched in RSP-ACA).

“L5 PT CTX contains 4 supertypes: Chrna6 (enriched in posterior sensory areas and highly distinct from the other three supertypes), Npnt, Cdh13 and C1ql2 (enriched in RSP-ACA).”

May combine two claims.

The ENT IT group has 10 supertypes organized in a layer-selective manner, in the order of L2, L2/3, L3, L5 and L6, consistent with their correspondence with CTX L2/3-L6 IT types, and shows differential spatial distribution along the anterior-posterior axis.

“The ENT IT group (Fig. 4B–D, F–H) has 10 supertypes organized in a layer-selective manner, in the order of L2, L2/3, L3, L5 and L6, consistent with their correspondence with CTX L2/3-L6 IT types, and shows differential spatial distribution along the anterior-posterior axis”

May combine two claims.

All SUB and ProS clusters are most related to L5 PT CTX, as is cluster L5 PPP #263 (though more distantly).

“All SUB and ProS clusters are most related to L5 PT CTX, so is cluster L5 PPP #263 (though more distantly).”

May combine two claims.

As with the IT cells, the CTX L5 NP, L6 CT and L6b clusters are largely shared across cortical areas, whereas the HPF NP and CT cell types are highly distinct among SUB, ENT and PPP.

“As with the IT cells (Fig. 4), the CTX L5 NP, L6 CT and L6b clusters are largely shared across cortical areas (Fig. 5A, D, S2), whereas the HPF NP and CT cell types are highly distinct among SUB, ENT and PPP.”

May combine two claims.

In the Cdh13 supertype, clusters #251-255 (Nnat+) are mostly populated by cells from prefrontal, medial and lateral associational areas, and #251 is highly specific to PL-ILA and ACA-RSP, based on marker genes Ndnf and Dlk1.

“Of note, in the Cdh13 supertype, clusters #251–255 (Nnat+) are mostly populated by cells from prefrontal, medial and lateral associational areas, and #251 is highly specific to PL-ILA and ACA-RSP, based on marker genes Ndnf and Dlk1”

May combine two claims.

L5/6 NP CTX is closely related to NP SUB and more distantly related to NP PPP.

“L5/6 NP CTX is closely related to NP SUB and more distantly related to NP PPP.”

May combine two claims.

They systematically identified a large number of neuronal types and subclasses from different HPF regions that are highly distinct from those in CTX and from each other.

“We systematically identified a large number of neuronal types and subclasses from different HPF regions that are highly distinct from those in CTX and from each other.”

May combine two claims.

The PPP IT group contains the L2/3 IT PPP subclass and is most closely related to supertype L2 IT RSP-ACA Npnt (cluster #134), consistent with the anatomical proximity of PPP and RSP.

“The PPP IT group (Fig. 4B–C, E–F, I–J) contains the L2/3 IT PPP subclass and is most closely related to supertype L2 IT RSP-ACA Npnt (cluster #134), consistent with the anatomical proximity of PPP and RSP.”

May combine two claims.

Projection mapping showed that neurons labeled via Scnn1a-Tg3-Cre driver line in RSP have long-range projections to both intra- and extratelencephalic targets; they project to ACA, RHP regions, contralateral RSP, and anteroventral nucleus (AV) of thalamus.

“Projection mapping (http://connectivity.brain-map.org/; experiments 166269090, 166458363 and 181860879) (Oh et al., 2014) showed that neurons labeled via Scnn1a-Tg3-Cre driver line in RSP have long-range projections to both intra- and extratelencephalic targets; they project to ACA, RHP regions, contralateral RSP, and anteroventral nucleus (AV) of thalamus”

• cites Oh et al., 2014Not checked yet

  Not checked in the free run.

May combine two claims.

HPF IT cell types form two groups around the relatively continuous CTX IT types, one mainly from ENT and the other from PPP.

“HPF IT cell types form two groups around the relatively continuous CTX IT types, one mainly from ENT and the other from PPP.”

May combine two claims.

There is rarely a one-to-one correspondence between CTX glutamatergic clusters and regions, although some clusters are enriched in specific areas.

“Although enrichment in specific areas is seen for some clusters, there is rarely one-to-one correspondence between clusters and regions.”

May combine two claims.

CTX NP, CT and L6b subclasses have the greatest similarity with their counterparts in ENT, PPP and SUB, as do CTX L2/3, L4/5 and L6 IT subclasses.

“CTX NP, CT and L6b subclasses have the greatest similarity with their counterparts in ENT, PPP and SUB, as do CTX L2/3, L4/5 and L6 IT subclasses”

May combine two claims.

L6 CT CTX is closely related to CT SUB as well as supertype L6 CT ENT Rasgrf2 Rmst.

“L6 CT CTX is closely related to CT SUB as well as supertype L6 CT ENT Rasgrf2 Rmst.”

May combine two claims.

They extracted one main axis that drove CA1, ProS and SUB variation by one dimensional UMAP of all the cells in the CA1-ProS and SUB-ProS subclasses, and found that this axis corresponded to a proximal-distal (Pr-Di) gradient from CA1 to SUB, for which each stage of the transition was driven by a different set of genes.

“We first extracted one main axis that drove CA1, ProS and SUB variation by one dimensional UMAP of all the cells in the CA1-ProS and SUB-ProS subclasses, and found that this axis corresponded to a proximal-distal (Pr-Di) gradient from CA1 to SUB, for which each stage of the transition was driven by a different set of genes”

May combine two claims.

The three ALM L5 PT types previously identified correspond to cluster #248 (thalamus-projecting cells) in the Npnt supertype and clusters #249 and #252 (medulla-projecting cells) in the Cdh13 supertype, respectively.

“The three ALM L5 PT types previously identified (Economo et al., 2018; Tasic et al., 2018) correspond to cluster #248 (thalamus-projecting cells) in the Npnt supertype and clusters #249 and #252 (medulla-projecting cells) in the Cdh13 supertype, respectively”

• cites Economo et al., 2018 sourceNot checked yet

  Not checked in the free run.

• cites Tasic et al., 2018 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

In the DG/SUB/CA neighborhood, DG, CA2 and CA3 subclasses are highly distinct, whereas SUB-ProS and CA1-ProS subclasses are more closely related.

“In the DG/SUB/CA neighborhood, DG, CA2 and CA3 subclasses are highly distinct, whereas SUB-ProS and CA1-ProS subclasses are more closely related”

May combine two claims.

The SUB-ProS and CA1-ProS subclasses both contain clusters from ProS.

“The SUB-ProS and CA1-ProS subclasses both contain clusters from ProS”

May combine two claims.

It expresses PT marker gene Bcl6 but not Fam84b; it also expresses a pan-IT marker Slc30a3, as well as L4 IT-specific markers Rspo1 and Scnn1a but not Rorb.

“It expresses PT marker gene Bcl6 but not Fam84b; it also expresses a pan-IT marker Slc30a3, as well as L4 IT-specific markers Rspo1 and Scnn1a but not Rorb”

May combine two claims.

The L6b CTX subclass also contains cells from ENT, PPP and SUB; these cells are mostly in supertype L6b RHP Nxph4 Cobll1.

“The L6b CTX subclass also contains cells from ENT, PPP and SUB; these cells are mostly in supertype L6b RHP Nxph4 Cobll1.”

May combine two claims.

The supertypes within L2/3 IT PPP follow a rostral dorsal to caudal ventral transition, starting with the Pdlim1 supertype in L2 of RSPv, POST and PRE, then the Kit supertype in L2/3 of POST-PRE, followed by the Wfs1 Prlr supertype in L2/3 of PAR, and the Cfap58 supertype (#145) in caudal PAR.

“The supertypes within L2/3 IT PPP follow a rostral dorsal to caudal ventral transition (Fig. 4E, I–J), starting with the Pdlim1 supertype in L2 of RSPv, POST and PRE, then the Kit supertype in L2/3 of POST-PRE, followed by the Wfs1 Prlr supertype in L2/3 of PAR, and the Cfap58 supertype (#145) in caudal PAR.”

May combine two claims.

In almost all CTX glutamatergic subclasses, medial (RSP/ACA) and lateral (TPE) regions are more distinct, while the anterior-to-posterior transition is more continuous.

“In almost all cases, the medial (RSP/ACA) and lateral (TEa-PERI-ECT, collectively TPE) regions are more distinct while the anterior-to-posterior transition is more continuous.”

May combine two claims.

Glutamatergic neuron types are more diverse than GABAergic neuron types, both molecularly and spatially.

“Glutamatergic neuron types are more diverse than GABAergic neuron types, both molecularly and spatially.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Sorting more cells into one tube diluted the ACSF in the collection buffer, causing cell death.

“We found that sorting more cells into one tube diluted the ACSF in the collection buffer, causing cell death.”

May combine two claims.

The homologous relationships between HPF and isocortical cell types are supported by similar expression patterns of a large set of marker genes, including canonical isocortical subclass-specific transcription factors, and by similar laminar localization in ISH and Visium data for corresponding IT/NP/CT/L6b pairs.

“These homologous relationships are supported by similar expression patterns of a large set of marker genes including canonical isocortical subclass-specific transcription factors (Fig. S3), and similar laminar localization as demonstrated by ISH and Visium data, between corresponding pairs of IT/NP/CT/L6b cell types.”

May combine two claims.

Across all isocortical areas, most glutamatergic cell types are shared among multiple areas, contrary to previous findings from VISp and ALM.

“Across all isocortical areas, most glutamatergic cell types are shared among multiple areas (Fig. 3B, S2), contrary to what we found in our previous study of two distantly located cortical areas, VISp and ALM”

• cites Tasic et al., 2018 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Pairwise correlation heatmaps are consistent with the area dendrogram and show TPE and RSP as the most distinct areas.

“The pairwise correlation heatmaps of all areas for each subclass reveal consistent patterns with the tree, and again showing TPE and RSP as the most distinct areas”

Probably not a substantial claim: may not pass S3 (load-bearing).

Decreased cell viability was observed for longer sorts.

“We also observed decreased cell viability for longer sorts.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

SSv4 libraries were sequenced to a target read depth of 0.5M reads per cell, with range 100,275-12,329,698 and median 1,003,867.

“Briefly, libraries were sequenced on an Illumina HiSeq2500 platform (paired-end with read lengths of 50 bp) to a target read depth of 0.5M reads per cell (range 100,275–12,329,698, median 1,003,867).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

There were three small, sex specific clusters, #158, #262, #335, all female specific and almost all the cells in these clusters are from highly specific regions in AI, RSP and HIP respectively.

“There were three small, sex specific clusters, #158, #262, #335, all female specific and almost all the cells in these clusters are from highly specific regions in AI, RSP and HIP respectively.”

May combine two claims.

In both glutamatergic and GABAergic classes, some cell types are highly specific to a region, layer, or location, while others are widely distributed and shared among multiple regions.

“In both classes, some cell types are highly specific to a region, layer or location, while others are widely distributed and shared among multiple regions.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

For 10xv2, the study profiled 12 joint regions for CTX and 3 joint regions for HPF with 30,000-300,000 cells per region; for SSv4, it profiled 17 regions for CTX and 5 regions for HPF with 1,100-14,000 cells per region.

“In total, for 10xv2, we profiled 12 joint regions for CTX and 3 joint regions for HPF with 30,000–300,000 cells per region; for SSv4, we profiled 17 regions for CTX and 5 regions for HPF with 1,100–14,000 cells per region.”

May combine two claims.

The study defines 28 subclasses, 56 supertypes, and 241 types/clusters in the glutamatergic class and 6 subclasses, 30 supertypes, and 119 types/clusters in the GABAergic class.

“We define 28 subclasses, 56 supertypes and 241 types/clusters in the glutamatergic class and 6 subclasses, 30 supertypes and 119 types/clusters in the GABAergic class.”

May combine two claims.

Regional variations are not uniform among subclasses; L5 PT and L4/5 IT types appear more distinct across areas than types in other subclasses, and different subclasses contain types enriched in different areas.

“Regional variations are not uniform among different subclasses, for example, L5 PT and L4/5 IT types appear more distinct across areas than types in other subclasses, and different subclasses contain types that are enriched in different areas”

May combine two claims.

Medial, prefrontal, and lateral associational areas are more distinct from sensorimotor areas and from each other; sensorimotor areas are more continuous but several domains (posterior sensory, lateral sensory, frontal motor) can be discerned.

“Overall, we find that medial, prefrontal and lateral associational areas are more distinct from sensorimotor areas and from each other; sensorimotor areas are more continuous but several domains (posterior sensory, lateral sensory and frontal motor) can be discerned.”

May combine two claims.

For isocortical glutamatergic types, a small number are specific to one or two regions, all in associational areas, while most types are shared among several or multiple areas with graded variations.

“For glutamatergic types in isocortex, we find a small number of types that are specific to one or two regions, all in associational areas; but most types are shared among several or multiple areas with graded variations”

Probably not a substantial claim: may not pass S3 (load-bearing).

The actual median sequencing depth achieved was 59,728 reads per cell across 175 libraries, against a target of 60,000 reads per cell.

“We targeted sequencing depth of 60,000 reads per cell; the actual median achieved was 59,728 reads per cell across 175 libraries.”

May combine two claims.

There were three main categories of noise clusters: clusters located outside areas of interest due to inaccurate dissection, clusters with significantly lower gene detection due to extensive drop out, and clusters due to doublets or contamination.

“There were three main categories of noise clusters: clusters located outside areas of interest due to inaccurate dissection, clusters with significantly lower gene detection due to extensive drop out, and clusters due to doublets or contamination.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The study used 530 male and female animals to collect 76,381 cells for SSv4 and 54 male and female animals to collect 1,561,952 cells for 10xv2.

“We used 530 male and female animals to collect 76,381 cells for SSv4 and 54 male and female animals to collect 1,561,952 cells for 10xv2.”

May combine two claims.

Glutamatergic types in different HPF regions are highly distinct from each other and from those in isocortex, consistent with a more ancient emergence of HPF and its subregions.

“Glutamatergic types in different HPF regions are highly distinct from each other and from those in isocortex, consistent with the notion of a more ancient emergence of HPF and its subregions.”

May combine two claims.

There is very little confusion between distant cortical regions, but considerable confusion between neighboring regions, particularly sensory areas.

“There is very little confusion between regions that are distant, whereas considerable confusion exists between neighboring regions, particularly for sensory areas.”

May combine two claims.

The current CTX-HPF GABAergic taxonomy corresponds well with previous transcriptomic and multimodal studies from individual regions, with the most robust correspondence at supertype level.

“Our current CTX-HPF GABAergic taxonomy corresponds well with previous transcriptomic and multimodal studies from individual regions; the correspondence is the most robust at supertype level”

May combine two claims.

Individual clusters within each CTX glutamatergic subclass occupy specific domains on the gradient map, with more similar clusters located closer to each other.

“Individual clusters within each subclass occupy specific domains on the gradient map, with more similar clusters located closer to each other”

May combine two claims.

The number of animals contributing to each cluster varies between 2 and 213, with an average of 56 and median of 44.

“The number of animals contributing to each cluster varies between 2 and 213, with an average of 56 and median of 44.”

May combine two claims.

L5/6 NP and GABAergic cells appear much less activated and show little regional difference.

“L5/6 NP and GABAergic cells appear much less activated and show little regional difference.”

May combine two claims.

Cell types are shared between ventral RSP and POST-PRE, and between lateral associational cortical areas and ENT.

“We also identify cell types that are shared between ventral RSP and post- and presubiculum (POST-PRE), and between lateral associational cortical areas and entorhinal cortex (ENT)”

May combine two claims.

At the transcriptomic level, cell types can be organized hierarchically over a complex landscape with both discrete and continuous variations, and these molecular relationships correlate strongly with spatial arrangement (location and layer).

“At transcriptomic level, cell types can be organized in a hierarchical manner over a complex landscape with both discrete and continuous variations. Such molecular relationships correlate strongly with the spatial arrangement (both location and layer) of the cell types.”

May combine two claims.

Fmo1, a previously known CA3 proximal gene, is expressed in dorsal clusters #356 and #358 but not #357.

“a previously known CA3 proximal gene, Fmo1 (Thompson et al., 2008), that is expressed in dorsal clusters #356 and #358, but not #357.”

• cites Thompson et al., 2008 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Many deep layer markers such as Lpl have ubiquitous expression in dorsal distal CA1, positioning these cells closer to the deep end of the pseudo Su-De axis.

“while many deep layer markers such as Lpl have ubiquitous expression in the dorsal distal CA1, positioning these cells closer to the deep end of the pseudo Su-De axis.”

• AssumptionRests on an assumption: The top principal component computed from a selected gene set faithfully represents the biological axis it is named after (e.g., Pr-Di, Do-Ve, Su-De, activity).

May combine two claims.

Besides the genes in the core program, each region also has a specific set of genes that contribute to this axis, e.g., ventrally Coch (CA3) and Gpc3 (CA1/ProS, likely in HATA), dorsally Rxfp1 and Elfn1 (SUB/ProS), Wfs1 (CA1), and Rph3a (CA3).

“Besides the genes in the core program, each region also has a specific set of genes that contribute to this axis (Fig. S5C), e.g., ventrally Coch (CA3) and Gpc3 (CA1/ProS, likely in HATA), dorsally Rxfp1 and Elfn1 (SUB/ProS), Wfs1 (CA1), and Rph3a (CA3).”

May combine two claims.

We selected a core set of genes with absolute correlation greater than 0.4 in both CA1 and CA3, including 84 dorsal-specific genes and 116 ventral-specific genes shared by CA1 and CA3.

“We selected a core set of genes with absolute correlation greater than 0.4 in both cases. These include 84 dorsal-specific genes and 116 ventral-specific genes that are shared by CA1 and CA3.”

May combine two claims.

Most regional markers identified show distinct layer specificity, and very few genes show the same regional preference across all layers.

“Interestingly, most regional markers we identified show distinct layer specificity, and very few genes show the same regional preference across all layers.”

May combine two claims.

While differences in the number of genes detected at the subclass level were observed, cells within a subclass had relatively homogeneous distribution.

“While we observed differences in the number of genes detected at the subclass level, cells within a subclass had relatively homogeneous distribution.”

Classified as probably not a substantial claim. May combine two claims.

We computed the top 50 DE genes in each direction for all pairs of glutamatergic clusters in CTX (2,100 genes) and HPF (2,467 genes) separately, with 1,633 genes in common.

“To study their correspondence systematically, we first computed the top 50 DE genes in each direction for all pairs of glutamatergic clusters in CTX (2,100 genes) and HPF (2,467 genes) separately, with 1,633 genes in common.”

May combine two claims.

Using the same gene set in SUB/ProS and DG regions, we again observed segregation of clusters along this gradient, which also corresponded to the dorsal-ventral axis in SUB/ProS/DG based on ISH of key marker genes.

“We used the same gene set to compute the top PC in SUB/ProS and DG regions and again observed segregation of clusters along this gradient, which also corresponded to dorsal-ventral axis in SUB/ProS/DG based on ISH of key marker genes.”

May combine two claims.

In CA3, St18, Hopx, Sgcd, and Prss23 are expressed in a very thin deep layer, while Nos1, Kctd4, and Kcnq5 have complementary expression and label most CA3 cells.

“Similarly, in CA3 we identified a set of markers, including St18, Hopx, Sgcd and Prss23, that are expressed in a very thin deep layer, while Nos1, Kctd4, Kcnq5 have complementary expression pattern and label most of the CA3 cells”

May combine two claims.

Glutamatergic cell types are highly distinct between CTX and HPF regions, but they also have intricate relationships according to the taxonomy tree, UMAP projections, and constellation plots.

“Glutamatergic cell types are highly distinct between CTX and HPF regions, but they also have intricate relationships according to the taxonomy tree, UMAP projections, and constellation plots.”

May combine two claims.

In both CA1 and CA3, the top PC corresponded to a dorsal-ventral gradient, validated by ISH images of key genes driving the axis.

“In each case, we observed that the top PC corresponded to a dorsal-ventral (Do-Ve) gradient (Fig. 6F–I), validated by ISH images of the key genes that drive the axis.”

May combine two claims.

Tshz2 is highly enriched in RSP/ACA across all layers except for cells in the L5/6 NP subclass, which have strong expression of Tshz2 in all regions.

“One exception is Tshz2, which are highly enriched in RSP/ACA across all layers except for cells in L5/6 NP subclass, which have strong expression of Tshz2 in all regions.”

May combine two claims.

The initial random forest classifier produced a very strong result, but the results were partially driven by donor-specific genes.

“While our initial attempt using random forest classifier produced a very strong result (data not shown), we realized that the results were partially driven by donor-specific genes.”

• AssumptionRests on an assumption: Donor-specific transcriptional signatures are subtle and do not affect the global clustering results.

Probably not a substantial claim: may not pass S3 (load-bearing).

We selected the top 50 genes in each direction and pooled such genes from all pairwise comparisons with a total of 5,981 gene markers.

“We selected the top 50 genes in each direction and pooled such genes from all pairwise comparisons with a total of 5,981 gene markers (Table S4).”

May combine two claims.

In CA1, Nptx2, Sulf2, and Lpl are expressed in the deep layer, while Pde11a and Anln are expressed in the superficial layer.

“Overall, in CA1, Nptx2, Sulf2, Lpl are expressed in the deep layer, while Pde11a, Anln are in the superficial layer”

Classified as probably not a substantial claim. May combine two claims.

This gene set was used to compute the top PC within each subclass and scaled in range [0,1], which corresponds to the Ve-Do axis in all these regions.

“This gene set was used to compute the top PC within each subclass and scaled in range of [0,1] as shown in Fig. S5C, which corresponds to the Ve-Do axis in all these regions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Glutamatergic neuronal types are organized by layers and their long-range projection patterns.

“Glutamatergic neuronal types are organized by layers and their long-range projection patterns”

• cites Harris and Shepherd, 2015 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Archicortex and paleocortex are considered evolutionarily older structures, whereas iso/neocortex emerged later and substantially expanded in vertebrate evolution.

“Archicortex and paleocortex are considered evolutionarily older structures, whereas iso/neocortex emerged later and substantially expanded in vertebrate evolution”

• cites Rakic, 2009 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The diversity in interareal axon-projection patterns of glutamatergic neurons forms the structural basis of the hierarchical network.

“their tremendous diversity in interareal axon-projection patterns forms the structural basis of the hierarchical network”

• cites D’Souza et al., 2016 sourceNot checked yet

  Not checked in the free run.

• cites Harris et al., 2019 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Area specialization is likely a major contributing factor to the diversity of functions supported by cortex.

“This area specialization is likely a major contributing factor to the diversity of functions supported by cortex”

• cites Cadwell et al., 2019 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The cortex is partitioned into multiple areas with specific input and output connections with many subcortical and other cortical regions.

“The cortex is partitioned into multiple areas with specific input and output connections with many subcortical and other cortical regions”

• cites Rakic, 2009 sourceNot checked yet

  Not checked in the free run.

• cites Van Essen and Glasser, 2018Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Developmentally, cortex originates from pallium, a main part of the telencephalon, which can be divided into several parts.

“Developmentally, cortex originates from pallium, a main part of the telencephalon, which can be divided into several parts”

• cites Pessoa et al., 2019 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Many molecular, anatomical, and physiological studies have revealed a broad spectrum of neuronal cell types in different cortical and hippocampal regions.

“Many molecular, anatomical, and physiological studies have revealed a broad spectrum of neuronal cell types in different cortical and hippocampal regions”

• cites Zeng and Sanes, 2017 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Neocortical areas form a hierarchical neural network with functionally distinct modules and feedforward and feedback pathways within and between modules.

“these neocortical areas together form a hierarchical neural network with functionally distinct modules and feedforward and feedback pathways both within and between modules”

• cites Coogan and Burkhalter, 1993Not checked yet

  Not checked in the free run.

• cites Felleman and Van Essen, 1991 sourceNot checked yet

  Not checked in the free run.

• cites Harris et al., 2019 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

There is a prominent functional transition along the dorsal-ventral axis of HPF, with the dorsal network mainly mediating spatial navigation and the ventral network mainly mediating emotional behaviors.

“there is a prominent functional transition along the dorsal-ventral axis of HPF, with the dorsal network mainly mediating spatial navigation and the ventral network mainly mediating emotional behaviors”

• cites Cembrowski and Spruston, 2019Not checked yet

  Not checked in the free run.

May combine two claims.

Since isocortical Sst Myh8 and Etv1 cells are L5 Martinotti cells, the HPF Myh8 and Etv1 cells are likely the Martinotti-like oriens lacunosum-moleculare (OLM) cells.

“Since isocortical Sst Myh8 and Etv1 cells are L5 Martinotti cells (Gouwens et al., 2020) (Fig. S1B), the HPF Myh8 and Etv1 cells are likely the Martinotti-like oriens lacunosum-moleculare (OLM) cells (Leao et al., 2012).”

• cites Gouwens et al., 2020 sourceNot checked yet

  Not checked in the free run.

• cites Leao et al., 2012 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Previous analysis of single-cell transcriptomes defined 133 cell types in two isocortical areas, primary visual cortex (VISp) and anterolateral motor cortex (ALM), in mice, and found that they have shared GABAergic interneuron types but distinct glutamatergic neuron types.

“Previously, we analyzed single-cell transcriptomes to define 133 cell types in two isocortical areas, primary visual cortex (VISp) and anterolateral motor cortex (ALM), in mice (Tasic et al., 2018) and found that they have shared GABAergic interneuron types but distinct glutamatergic neuron types.”

• cites Tasic et al., 2018 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Because cell isolation aimed toward enrichment for neurons, the study had limited sampling of non-neuronal cells (15,241 10xv2 cells and 1,828 SSv4 cells after QC) and focused on neuronal cell types.

“Since the cell isolation in this study aimed toward enrichment for neurons, we had limited sampling of non-neuronal cells (15,241 10xv2 cells and 1,828 SSv4 cells after QC) and this study is focused on neuronal cell types.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some CGE cell types (e.g., some neurogliaform cells) may be developmentally derived from the nearby preoptic region (PO).

“Note that some of the CGE cell types (e.g., some neurogliaform cells) may, in fact, be developmentally derived from the nearby preoptic region (PO) (Niquille et al., 2018).”

• cites Niquille et al., 2018 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Meis2 neurons were identified as related to olfactory bulb interneurons.

“Meis2 neurons were identified as related to olfactory bulb interneurons (Frazer et al., 2017).”

• cites Frazer et al., 2017 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The distinct subclass of Car3 includes neurons from L6 of many lateral cortical areas, and previous study showed that these Car3+ L6 neurons, like cortical IT neurons, have extensive intracortical axon projections.

“The distinct subclass of Car3, which includes neurons from L6 of many lateral cortical areas, is included here as our previous study showed that these Car3+ L6 neurons, like cortical IT neurons, have extensive intracortical axon projections (Peng et al., 2020).”

• cites Peng et al., 2020Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Ntng1+ cells in HIP may be the trilaminar cells or radiatum-retrohippocampal neurons projecting to RSP.

“may be the trilaminar cells or radiatum-retrohippocampal neurons projecting to RSP”

May combine two claims.

While previous studies focused on one or two individual regions, the current taxonomy provides an overview of cell type variation across regions.

“While these studies focused on one or two individual regions, our current taxonomy provides an overview of cell type variation across regions.”

Classified as probably not a substantial claim.

Single-cell RNA-sequencing studies have systematically characterized and classified cell types in individual regions of isocortex and hippocampus.

“Single-cell RNA-sequencing (scRNA-seq) studies have systematically characterized and classified cell types in individual regions of isocortex and hippocampus”

• cites Harris et al., 2018 sourceNot checked yet

  Not checked in the free run.

• cites Hodge et al., 2019 sourceNot checked yet

  Not checked in the free run.

• cites Tasic et al., 2016 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S1 (asserted).

They hypothesize that this core set is the core program for dorsal-ventral gradient specification.

“which we hypothesize is the core program for dorsal-ventral gradient specification”

May combine two claims.

L2/3 IT CTX Baz1a likely corresponds to a subset of cortical L2/3 neurons that express IEG c-fos under basal conditions and are preferentially interconnected within the L2/3 network.

“likely corresponds to a subset of cortical L2/3 neurons that express IEG c-fos under basal conditions and are preferentially interconnected within the L2/3 network (Yassin et al., 2010)”

• cites Yassin et al., 2010 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The difference between broad sharing and the previous VISp/ALM finding is reconciled by the finding that transcriptomic cell types are distributed continuously and gradedly across the cortical sheet, often along anterior-posterior (more continuous) and medial-lateral (more discrete) dimensions.

“This difference is reconciled by our finding that these transcriptomic cell types are distributed in a continuous and graded manner across cortical sheet, often along anterior-posterior (more continuous) and medial-lateral (more discrete) dimensions”

May combine two claims.

The effect of transgenic line variation on cell type identities is minimal because nearly all clusters are dominated by 10xv2 cells from Snap25-IRES2-Cre mice, and SSv4 cells mainly came from pan-glutamatergic and pan-GABAergic lines with other specialized Cre lines contributing to only a small fraction of cells.

“Since nearly all clusters are dominated by 10xv2 cells from Snap25-IRES2-Cre mice, and SSv4 cells mainly came from pan-glutamatergic and pan-GABAergic lines with other specialized Cre lines contributing to only a small fraction of cells, the effect of transgenic line variation is minimal.”

May combine two claims.

Both isocortex and HPF in mammals evolved from the simpler three-layered cortex in reptiles into two parallel "six-layered" circuit organizations, while isocortex further underwent accelerated evolution resulting in multiplication of areas each as an independent circuit unit.

“we suggest that both isocortex and HPF in mammals evolved from the simpler three-layered cortex in reptiles into two parallel “six-layered” circuit organizations, while isocortex further went through accelerated evolution resulting in a multiplication of areas each as an independent circuit unit.”

May combine two claims.

This predicted isocortex-like circuit organization is highly consistent with currently known interareal connections of HPF, providing validation to the transcriptomic cell type framework.

“This prediction is highly consistent with the currently known interareal connections of HPF, providing a level of validation to the transcriptomic cell type framework”

• cites Bienkowski et al., 2018 sourceNot checked yet

  Not checked in the free run.

• cites Ding et al., 2020 sourceNot checked yet

  Not checked in the free run.

• cites Gergues et al., 2020 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The study used mice of both sexes, except that MOp 10xv2 data had only male cells.

“We used mice of both sexes, except that MOp 10xv2 data had only male cells (Methods S1).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The package performs consensus clustering by repeating the iterative clustering step on 80% subsampled set of cells 100 times, and derives the final clustering result based on cell-cell co-clustering probability matrix.

“The package also performs consensus clustering by repeating the iterative clustering step on 80% subsampled set of cells 100 times, and then derives the final clustering result based on cell-cell co-clustering probability matrix.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Manual microdissections are imperfect, but the dissections still contain substantial enrichment of cells for the targeted regions despite the presence of cells from neighboring regions.

“Note that manual microdissections are imperfect, but despite the presence of cells from neighboring regions, the dissections still contain substantial enrichment of cells for the targeted regions.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The origin of hippocampal GABAergic neurons from MGE/CGE and their following the same migration streams as isocortical ones may explain the finding that all isocortical GABAergic types are also present in all retrohippocampal regions and most are found in HIP.

“This may explain our finding that all isocortical GABAergic types are also present in all retrohippocampal regions and most of them are found in HIP as well.”

Classified as probably not a substantial claim. May combine two claims.

The study covered all regions of the cerebral cortex and hippocampal formation, using mid-ontology level sampling with judicious joining of neighboring regions.

“We covered all regions of CTX and HPF and chose sampling at mid-ontology level with judicious joining of neighboring regions (Fig. 1A, Methods S1, Table S1).”